Chemical Analysis of the Fornax Dwarf Galaxy Letarte, Bruno

University of Groningen

Chemical analysis of the Fornax dwarf galaxy Letarte, Bruno

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record

Publication date: 2007

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA): Letarte, B. (2007). Chemical analysis of the Fornax dwarf galaxy. [s.n.].

Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license. More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne- amendment.

Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

Download date: 08-10-2021 Rĳksuniversiteit Groningen

Chemical Analysis of the Fornax Dwarf Galaxy

Proefschrift

ter verkrĳging van het doctoraat in de Wiskunde en Natuurwetenschappen aan de Rĳksuniversiteit Groningen op gezag van de Rector Magniﬁcus, dr. F. Zwarts, in het openbaar te verdedigen op vrĳdag 30 maart 2007 om 14.45 uur

door

Bruno Letarte

geboren op 12 juni 1976 te Québec, Canada Promotor: Prof. dr. E. Tolstoy Copromotor: Dr. V. Hill

Beoordelingscommissie: Prof. dr. M. Spite Prof. dr. P. C. van der Kruit Prof. dr. J. W. Pel

ISBN 90-367-2927-0 ISBN 90-367-2928-9 (electronic version) In the beginning the Universe was created. This has made a lot of people very angry and has been widely regarded as a bad move.

–Douglas Adams Cover page – Fornax Dwarves, by Jesse Giroux

Contact information:

Bruno Letarte [email protected] [email protected]

This thesis has been funded by:

With support from: LKBF Leids Kerkhoven-Bosscha Fonds Contents

1 Introduction 9 1.1 The Cosmological Importance of Dwarf Galaxies ...... 9 1.2 The Formation of the Elements ...... 11 1.3 Abundances in Galaxies ...... 12 1.3.1 The Milky Way ...... 13 1.3.2 The Magellanic Clouds & Dwarf Galaxies ...... 14 1.4 The DART project ...... 15 1.4.1 Photometry ...... 15 1.4.2 Spectroscopy ...... 15 1.4.3 This Thesis ...... 17

2 Fornax and the Local Group 19 2.1 Dwarf galaxies in the Local Group ...... 19 2.2 Fornax dSph ...... 21 2.3 Globular Clusters in Fornax ...... 24

3 Using stellar atmospheric models ... chemical abundances 27 3.1 Describing the stellar atmosphere ...... 27 3.1.1 The flux ...... 28 3.1.2 The absorption coefficient ...... 30 3.1.3 Stellar atmospheric models ...... 35 3.2 Determining Stellar Atmospheric parameters ...... 35 3.2.1 Effective Temperature (Teff )...... 35 3.2.2 Surface Gravity (log g)...... 36 3.2.3 Metallicity ...... 39 3.2.4 Microturbulence velocity ...... 39 3.3 The abundance determination ...... 39 3.3.1 Measuring the equivalent widths ...... 40 3.3.2 The Stellar Models used ...... 40 3.3.3 Computing the abundances ...... 42 3.4 The line list ...... 43 3.4.1 Building a line list ...... 43 vi CONTENTS

3.4.2 The line by line selection ...... 45

4 Abundances with the FLAMES multi-fibre instrument 47 4.1 UVES vs FLAMES ...... 48 4.1.1 UVES ...... 49 4.1.2 FLAMES ...... 49 4.2 The FLAMES Spectra ...... 50 4.2.1 Extracting, calibrating ...... 50 4.2.2 Combining ...... 50 4.2.3 Determining the radial velocities (Vrad) ...... 52 4.2.4 Measuring the Equivalent Widths ...... 52 4.2.5 Cleaning up the spectra ...... 56 4.3 Selecting our stellar parameters ...... 57 4.3.1 Photometric gravity ...... 57 4.3.2 Photometric Teff ...... 57 4.3.3 Iterating on the parameters ...... 62 4.3.4 Precision and error estimates ...... 65 4.4 Systematics and corrections ...... 68 4.4.1 Systematics ...... 68 4.4.2 Hyperfine splitting correction ...... 71 Appendix 4.A Large tables ...... 72

5 HR spectroscopy in Fornax Globular Clusters 77 5.1 Introduction ...... 78 5.2 Observations ...... 79 5.3 Data Reduction and Analysis ...... 81 5.4 Interpretation ...... 85 5.4.1 The Iron abundance ...... 85 5.4.2 The Alpha elements ...... 86 5.4.3 Deep mixing pattern ...... 89 5.4.4 Iron-peak elements ...... 90 5.4.5 Heavy elements ...... 92 5.5 Conclusions ...... 95 Appendix 5.A Large tables ...... 97

6 HR spectroscopic study of Fornax Field Stars 105 6.1 Sample selection ...... 106 6.2 Results ...... 107 6.2.1 Iron abundance ...... 107 6.2.2 Alpha Elements ...... 108 6.2.3 Iron peak elements ...... 113 6.2.4 Deep-mixing pattern ...... 114 6.2.5 The Na-Ni relationship ...... 115 6.2.6 Heavy elements ...... 116 6.3 Discussion ...... 120 6.3.1 Comparison of Fornax and Sculptor ...... 121 6.3.2 Age and [Fe/H] ...... 122 6.4 Conclusions ...... 123 CONTENTS vii

Appendix 6.A Large tables ...... 124

7 Conclusions 141 7.1 New Data Reduction and Analysis Techniques ...... 141 7.2 The Fornax Globular Clusters ...... 142 7.3 Fornax Field stars ...... 142

Bibliography 145

Nederlandse samenvatting 151

Résumé français 155

Acknowledgements 159

Chapter 1 Introduction

warf galaxies are in principle the most simple and straightforward type of galaxy D and their study can be used to test numerous theories of the formation and evolution of stars and galaxies in a range of environments. This thesis concentrates on the detailed study of the chemical elements in individual stars in the nearby dwarf spheroidal galaxy, Fornax. A dwarf spheroidal galaxies are small roughly spherical galaxies that are typically found in the vicinity of larger galaxies, such as the Milky Way. They typically do not have any ongoing star formation, nor to they appear to have any gas associated to them. The abundance ratios of diﬀerent elements in individual stars with a range of ages provide a detailed insight into the various chemical enrichment processes (e.g., supernovae, stellar winds) which in turn improves our understanding of the global processes of formation and evolution of a galaxy as a whole.

1.1 The Cosmological Importance of Dwarf Galaxies

The most straightforward model of galaxy formation is that all galaxies form in the early Universe in a rapid collapse scenario (so called monolithic collapse, Eggen, Lynden-Bell, & Sandage 1962). These galaxies then evolve solely by changing their gas mass into a stellar mass with time. This model assumes that the majority of the mass of all galaxies was in place at their formation. However this basic picture was updated (e.g., Searle & Zinn 1978) to a model which assumes that galaxies are not formed in a single collapse, but that they are built up in time from smaller fragments. This theory came in parallel with the very successful “cold dark matter” (CDM) vision of structure formation in the Universe which assumes that the dark matter content of a galaxy is built up through the continuous accretion of small clumps, to build up the galaxies and clusters of galaxies we see today (e.g., White & Rees 1978; Navarro, Frenk, & White 1995).

If we take the CDM model of structure formation and assume that the ratio of baryonic to dark matter is roughly constant and known then this naturally results in the concept of numerous “building blocks”, or small galaxies, which are continuously being accreted onto larger galaxies over the history of the Universe. These small galaxies, with 10 chapter 1: Introduction a similar mass to the dwarf galaxies we see today, might act as stellar nurseries, creating the stars we see in the Milky Way (MW) today. Stars within the Galactic halo are some of the oldest objects ever observed and they should be representative of the earliest star formation in the Local Group (LG). These stars either formed in the proto-Milky Way or they may have formed in smaller satellite galaxies that were accreted to the Milky Way at a later time. CDM based models thus suggest that a considerable fraction of the stars in the Milky Way today should have formed in smaller building blocks. For example, the Sagittarius dwarf galaxy behaves exactly like a CDM building block, showing signs of being tidally disrupted and merging in its entirety into the Milky Way (Ibata et al. 1994).

As required by the CDM view of the Universe small galaxies do appear to be dark matter dominated (e.g., Mateo 1998). Observations of dwarf spheroidal galaxies in the Local Group, such as Fornax dSph, suggest that dwarf galaxies must be considerably 9 10 more massive than the visible mass would suggest (e.g., ∼ 10 − 10 M , as compared 7 8 to visible masses of ∼ 10 − 10 M ), (Mateo et al. 1991; Walker et al. 2006; Battaglia et al. 2006). However there are inconsistencies in the predicted properties of the DM proﬁles of the observed dwarfs and the predictions of CDM (e.g., Wilkinson et al. 2006). It also appears that the properties of the stellar populations, the dark to baryonic matter ratio, and the kinematic properties of dwarf galaxies we see today are inconsistent with the requirements of building blocks of the Milky Way, i.e., adding together all the small galaxies we see today, or at any time in the past, will not result in a galaxy like the Milky Way (e.g., Shetrone et al. 2003; Tolstoy et al. 2003; Venn et al. 2006; Helmi et al. 2006).

CDM also appears to over-predict the number of small satellite galaxies around larger galaxies such as our own, an inconsistency that is known as the “missing dwarf problem” (e.g., Moore et al. 1999). However, recent discoveries of several faint satellites around the Milky Way in the last couple of years are changing our view about the LG (e.g., Belokurov et al. 2006a,b; Willman et al. 2005a,b; Zucker et al. 2006a,b). These studies suggest that the dwarf spheroidal galaxies we have studied to date are only the tip of the iceberg; they are the most massive satellites of a larger population of fainter, lower mass satellites (e.g., Stoehr et al. 2002), which could bring our Milky Way environment back into consistency with the CDM predictions for the amount of sub-structure. However, our knowledge about these new faint galaxies, especially their dark matter content, is still quite limited as they have been discovered relatively recently.

Thus dwarf galaxies are useful probes of our understanding of galaxy formation and evolution on the smallest scales and potentially also as building blocks of the largest galaxies. By studying the nearest examples we can obtain the kind of detailed com- parisons between theory and observation that are required to test current theories and provide a solid observational basis for future models. More speciﬁcally, studying the abundance patterns of stars of a range of age allows us to understand in detail the evolutionary processes that shape galaxies everywhere. Thus looking at individual stars in dwarf galaxies in the Local Group is an important component in understanding the big picture of galaxy formation and evolution throughout the Universe. 1.2: The Formation of the Elements 11 1.2 The Formation of the Elements

It is believed that the Universe started as an explosion, known as the Big Bang, where hydrogen, deuterium, helium and lithium were created. These are thus considered to be primordial elements and all other elements are formed subsequently by nucleosynthesis in stars. Stellar nucleosynthesis is thus responsible for almost all of what we see around us on the Earth today. It was ﬁrst explained in the 1950s in work done by Fowler and Hoyle, culminating in the B2FH (Burbidge, Burbidge, Fowler, & Hoyle 1957) paper.

The ﬁrst most fundamental process of converting hydrogen into heavier elements is hydrogen burning, which is the conversion of hydrogen nuclei into helium, via the proton- proton chain in low mass stars with low core temperatures, and via proton captures by carbon, nitrogen, and oxygen atoms (in the CNO cycles) in more massive stars with higher temperatures. The CNO cycle traces the origin of most of the observed nitrogen today, while most of the helium produced is consumed in the next stage: helium burning. As helium builds up in the core of the star the core contracts until the temperature and density increase enough to allow for another reaction in which helium is the fuel. This thermonuclear phase is the triple-α process in which three 4He nuclei fuse to form a carbon nucleus. The next stage is shell burning: carbon burning, oxygen burning, silicon burning. This can produce elements as heavy as 56Fe which is the most massive element that can be formed by fusion in the core of a star.

The most significant group of heavier elements are the so called alpha elements, with nuclei that are multiples of He, e.g., O, Mg, Ca, Si and Ti. They are predominantly synthesised by alpha capture during the various burning phases in massive stars, and expelled into the ISM by SN II explosions. Another significant and important group of elements is the iron-peak, including Fe itself. It is predominantly produced and expelled into the ISM by SN Ia, supernovae thought to be due to the explosion of a white dwarf in an evolved binary with a less massive progenitor star. Those typically occur ∼1 Gyr after the first episode of star formation, contrary to SN II which have short-lived massive star progenitors (as short as ∼10 Myrs). As a consequence, elemental ratios of the type [α/Fe] inform us of the relative contribution from the two types of supernovae at a given time, indicative of star formation timescale. Figure 1.1 sketches how the [α/Fe] ratio can be viewed as a kind of chronometer (starting to decrease after 1Gyr) while the [Fe/H] metallicity index provides the efficiency with which star formation has occurred. When the star formation rate (SFR) is high, then the gas will reach higher [Fe/H] before the first SN Ia occur and α-elements start to decrease (the “knee”). The formation efficiency and time scale of a stellar system can be estimated by the position of this “knee”. And, because more massive stars are more efficient in producing α-elements, the level of [α/Fe] at low metallicity (before the “knee”) is an indication of the mass of the stars that con- tributed to enrich the ISM and therefore provides a indirect measure of the IMF.

Heavier elements beyond the iron peak are created by neutron capture, where the two most important processes (in the astrophysical context) are the s- and r- processes. The s-process (or slow-process) occurs when the neutron ﬂux is not very high, so that the intervals between neutron captures are long compared to the beta decay characteristic timescale of an unstable nucleus. These conditions are found in the envelopes of ther- 12 chapter 1: Introduction

Figure 1.1: Simple view of how α-elements can be used to trace the IMF and SFH of a galaxy (taken from McWilliam 1997).

mally pulsating AGB stars, and are most efficient in 3-5 M stars. Because of the slow evolution of intermediate-mass stars, s- process will only enter the chemical enrichment of a galaxy several 100 Myrs after the first episode of star formation. In addition, it requires pre-existing iron-peak elements seeds in the AGB envelope, and is therefore in- efficient at very low metallicity. The s- process is unlikely to be significant in the earliest stages of star formation in a galaxy.

The r-process (or rapid-process) occurs when there is sufficient neutron flux which allows rapid captures of neutrons. This is believed to occur predominantly in environments like those produced by SNe II. With such rapid successive captures, neutrons can accumulate on an unstable nuclei before it has time to either beta or alpha decay. The stars responsible for these explosions are massive, therefore have a short lifetime and are believed to be the first objects that will contribute heavy elements to the ISM. Observing the relative abundances of s- and r- process nuclei can therefore constrain the impact of AGB stars on chemical evolution and probe star formation timescales.

1.3 Abundances in Galaxies

Because elemental abundances are preserved∗ at the stellar surface during the whole stellar lifetime, and can be (relatively) easily measured from absorption lines in high- resolution stellar spectra, they have become a very important tool to understand the genesis of a stellar population. Abundances of various elements can be measured in stars of different ages and, thanks to their different nucleosynthetic origin, allow us to infer what enrichment processes have been dominant at different epochs of galaxy formation. Not surprisingly our earliest studies have concentrated on the Milky Way, and it is only relatively recently that similarly detailed studies have been made of other galaxies, such as the Magellanic Clouds and most recently the nearby dwarf spheroidal galaxies.

∗ Except for a few light elements which may be aﬀected by internal mixing: Li, C, N. 1.3: Abundances in Galaxies 13

1.3.1 The Milky Way The Milky Way contains several stellar components which are distinguished by different spatial distribution, kinematics and stellar populations, namely the halo, the thick disk, the thin disk, and the bulge. Each component has clearly had a different formation history and their stars show marked differences in their age distribution, metallicity distribution and most importantly here, abundance ratios. Ever since the discovery by Chamberlain & Aller (1951) that two stars with high radial velocities (halo stars) had their iron and calcium abundances an order of magnitude lower than that of the Sun, it gradually became clear that the various stellar populations that comprise the Milky Way have both kinematics and chemical signatures associated to each of them. and that combining the two properties was necessary to better understand galaxy evolution (Wyse & Gilmore 1995).

A review by McWilliam (1997), covering the Galactic disk, halo and bulge suggest that the environment plays an important role in chemical evolution and that supernovae come in many ﬂavors, with a range of element yields. Below are a some recent examples of detailed abundance studies of the Milky Way:

The detailed abundance studies of extremely metal poor stars in the halo of our Galaxy have given us a clearer picture of its earliest enrichment history. The high [Zn/Fe] observed and absence of very strong depletion of odd-numbered elements have ruled out pair instability SN (from 130-300 M progenitors) as a dominant source of enrichment (Cayrel et al. 2004). The dispersion in heavy neutron-capture element abundances of the most metal poor stars suggests incomplete mixing of the ejecta from individual supernovae into the galactic interstellar medium (McWilliam 1997).

Studies of large samples (∼200) nearby disk stars (F and G dwarf) provide observational constraints by linking chemical abundance of up to 30 chemical elements to precise kinematics and photometric ages (e.g., Edvardsson et al. 1993; Chen et al. 2000; Reddy et al. 2003). This has allowed to understand that the thin disk formed stars at a steady rate over the last 4-8 Gyrs, allowing a full evolution of the abundance ratios from almost pure SN II ejecta to a full mix of SN II, stellar winds and SN Ia. Although the mean metallicity increases with time, the age-metallicity relation is neither well deﬁned nor tight in the galactic disk, ruling out the “instantaneous mixing” assumption of simple models of galaxy chemical evolution.

Recent precision work has shown that the [α/Fe] ratio for thick-disc stars shows a clear enhancement compared to thin-disc members of the same metallicity, which is a sign that star formation was more eﬃcient and restricted to a shorter period of time in the thick disk (e.g. Bensby et al. 2003, 2005; Reddy et al. 2006). Several hypotheses have been proposed for the origin of the thick disk: the debris of a merger, a merger that heated a preexisting thin disk into a thick disk, etc. The ﬁrst indications of a population that could be ascribed to debris from the satellite whose merger caused the thick disk was presented in Gilmore et al. (2002): thick disk stars should then bear the chemical signature of the star formation history of the merging (dwarf ?) galaxy. Galactic stars seen along lines of sight to some dSph galaxies seem to have the expected properties of “satellite debris” in the thick disk-halo interface, which is interpreted as remnants of the 14 chapter 1: Introduction merger that heated a preexisting thin disk to form the thick disk (Wyse et al. 2006). Thick disk stars would then have the chemical signature of the former thin disk.

In studies of Galactic bulge stars, two α-element ratios, [O/Fe] and [Mg/Fe] have been found to be higher than in thick disk stars, which are known to be more oxygen rich than thin disk stars (e.g., Zoccali et al. 2006; Fulbright et al. 2006; Lecureur et al. 2006). This supports a scenario in which the bulge formed before and more rapidly than both the thin and thick disks, and therefore the MW bulge can be regarded as a prototypical old spheroid, with a formation history similar to that of early-type (elliptical) galaxies.

1.3.2 The Magellanic Clouds & Dwarf Galaxies Other galaxies are in principle simpler to interpret than the Milky Way as we have an external view of the entire system and distance differences are unimportant. Stars in the Magellanic Clouds (at ∼50 kpc distance) were the first extragalactic stars targeted for detailed abundance studies and the results of these studies gave us the first insights into a more metal poor star forming environment than is available in the disk of our galaxy. At the end of the 80’s and in the 90’s, 4m-class telescopes were used to study detailed abundances of supergiant stars in both the Large and Small Magellanic Clouds, reflecting the current interstellar medium within these galaxies (Russell & Bessell 1989; Hill et al. 1995; Hill 1997; Venn 1999). Probing the chemical composition of stars as a function of age and therefore chemical evolution per se had to wait until 8-10m class telescopes gave access to high-resolution spectra of RGB stars in the Large Magellanic Cloud, initially in small numbers, (Hill et al. 2000; Smith et al. 2002), followed by the first abundance study of a large sample (Pompeia et al. 2006).

Similarly, the Sagittarius dSph has also been targeted in high resolution studies of some tens of RBG stars (Bonifacio et al. 2000; Monaco et al. 2005). These studies re- vealed distinctive evolutionary paths for the Large Magellanic Cloud and the Sagittarius dSph, showing a diﬀerent chemical enrichment process from the Milky Way and other dwarf galaxies (Bonifacio et al. 2000).

However the Magellanic Clouds and Sagittarius are clearly in the process of inter- acting strongly with our galaxy and so the lessons they have to teach about galaxy formation and evolution are not so straightforward to interpret. Dwarf galaxies, on the other hand, especially the nearby dSph are arguably simpler and more clearly preserved environments. These are however twice as distant as the Magellanic Clouds, and thus detailed abundances require 8-10m class telescopes.

Using Keck to look at individual stars in the Draco, Sextans and Ursa Minor dSph (Shetrone et al. 1998, 2001), and soon after the VLT for four southern dSph (e.g., Shetrone et al. 2003; Tolstoy et al. 2003), studies of LG dSph were initially based on very small samples of stars and yet they provided fundamental insights into galaxy formation and evolution. From these studies it became evident that, whereas the metallicity of dSph stars seemed to lie between the bulk of Galactic disk and halo stars, α-elements were typically under abundant when compared to MW stars of similar metallicity (hence lower than in the halo), while r- and s- process elements in dSph stars were typically halo-like. 1.4: The DART project 15

This suggests that the satellite galaxies we see today cannot be signiﬁcant recent contributors to the stellar population of our Galaxy, with the possible exception of the outer halo. However, the lack of statistically signiﬁcant samples of objects (2−5 stars per dSph) undermined the strength of this conclusion.

More importantly still, although dSph are simpler systems when compared to the MW, with most of them having typically much lower star formation rates, each of them has a unique and diﬀerent star formation history. Abundance ratios were yet to be studied in large enough samples in several diﬀerent dSph to understand the internal evolution of these systems.

1.4 The DART project

DART is an acronym for Dwarf Abundance and Radial-velocity Team (Tolstoy et al. 2004, 2006). It involves more than 16 persons, from 10 institutes in 10 diﬀerent countries. The main goal of the project is to obtain detailed chemical abundances (requiring high resolution) and radial velocities (low resolution) for a large sample of stars in four nearby dSph galaxies, Sculptor, Fornax, Sextans and Carina (for which we obtained high resolution spectroscopy only). The project is primarily based on two observing proposals, the ESO Large Programme 171.B-0588 (PI: Tolstoy) entitled: “Dwarf galaxies: remnants of galaxy formation and corner stones for understanding galaxy evolution” and the Meudon GTO Programme 71.B-0641 (PI: Hill) entitled: “Star formation history of the Sculptor dwarf spheroidal galaxy” which began obtaining data in August 2003.

1.4.1 Photometry Wide-field accurate photometry was needed both to select targets for our spectroscopic survey, and to allow a colour-magnitude diagram analysis of the global properties (mean ages and metallicities) of the stellar populations in the galaxy and the underlying star formation history. Precise astrometry (better than 0.300) of the targets selected for spectroscopic follow-up is also required to insure a proper placement of FLAMES fibres (1.200 fibre entrance on the sky).

The instrument we used for our photometric survey is the wide field imager WFI, (Baade et al. 1999) on the 2.2-m MPG/ESO telescope on La Silla. The large field of view (330× 340) of this instrument allowed us to efficiently map out dwarf galaxies to beyond their tidal radius. Our photometric survey was conducted in the visible band V and I. Figure 1.2 shows the spatial distribution of our imaging for Fornax. We have also plotted the low-resolution spectroscopic survey, with bigger black points representing the FLAMES LR targets (Battaglia et al. 2006).

1.4.2 Spectroscopy We used VLT/FLAMES, described in Pasquini et al. (2002) as well as in chapter 4 of this thesis, to carry out our spectroscopic survey. For each of the four galaxies, we obtained one FLAMES pointing in high resolution mode, each consisting of ∼100 target stars for which we will obtain chemical abundances (and radial velocities). To obtain suﬃcient 16 chapter 1: Introduction

Figure 1.2: Spatial distribution of the Fornax DART imaging sample. The coordinates ξ and η are de-projected rectan- gular coordinates, using the centre of Fornax derived in (Battaglia et al. 2006). The ellipses are drawn at 1, 2, 3 and 4 core radius, Rcore, the last one corresponding to the tidal radius, Rtidal. wavelength coverage for an accurate analysis of the abundances and to include a variety of chemical elements, we used several diﬀerent setups that cover diﬀerent wavelength ranges. Three setups were obtained in order to perform an abundance analysis, totalling almost 30h of observation per galaxy (see chapter 4).

On the other hand, a low resolution pointing can be obtained in about an hour, allowing for a greater number of stars for which we get a basic metallicity tracer (Ca II triplet, or CaT) and a radial velocity. The Fornax study in LR consisted of 11 pointings and 1063 targets, as illustrated in Figure 1.2 (see Battaglia et al. 2006).

The DART studies to date, as well those of other groups, (e.g., Koch et al. 2006, 2007) have shown that neither the kinematics nor the metallicities nor the spatial distributions of dSph are easy to explain in a straightforward manner even for these smallest galaxies. Dwarf galaxies show complex and highly speciﬁc evolutionary and metal-enrichment processes. Details of these results coming from low resolution CaT spectroscopy are presented in Tolstoy et al. (2004) (for Sculptor) and in Battaglia et al. (2006) for Fornax. Speciﬁcally, in Fornax, we have shown that the galaxy contains at least two morpholog- ically (concentration), chemically (metallicity) and kinematically (velocity dispersion) distinct intermediate to old components. The centre of Fornax is dominated by the more metal-rich and kinematically cooler (and younger) component. This is the population from which our high-resolution sample was drawn. 1.4: The DART project 17

Figure 1.3: Digital Sky Survey (DSS) image of the central region of Fornax (850 x 620 or 3.4 x 2.5 kpc) with the central 250 field identified by a big circle and the five globular clusters with smaller circles.

1.4.3 This Thesis The main emphasis of this thesis is to determine detailed chemical abundances of individual stars in the nearby Fornax dwarf spheroidal galaxy, based on high resolution observations with VLT/FLAMES. We have targeted stars in the central 250 diameter region of Fornax, as well as in three of its globular clusters. An image of Fornax is shown in Figure 1.3, where the central FLAMES field we observed in HR is identified, as well as the location of the five globular clusters of Fornax. The goal was to make a consistent study of the chemical properties of a representative sample of the stellar population of Fornax, and to make a comparison between the properties of stars in its old globular clusters (GCs) and predominantly intermediate age field stars. Detailed abundance analysis from HR spectroscopy is necessary for the full understanding of a complicated star formation history, where classic colour-magnitude diagram (CMD) analysis is not sufficient to provide a definitive answer.

Although earlier studies have provided hints of the evolutionary processes in dwarf galaxies the unparallelled multi-tasking capability of VLT/FLAMES allows us to map out the large scale processes which are important on the scale of a dSph and also to distinguish “the weather from the climate” in these galaxies – with regard to the chemical evolution with time.

Chapter 2 Fornax and the Local Group

he cold dark matter (CDM) paradigm states that small galaxies are the building T blocks of larger galaxies. This formation scenario is quite successful at modelling large scale structures but it has a problem for objects the size of current day dwarf galaxies: many more dwarfs are predicted than are actually observed in the Local Group. The surviving dwarf galaxies give us the opportunity to learn more about them and their relation to larger galaxies such as the Milky Way and M 31. By studying photometric properties, kinematics and the detailed chemistry of individual stars in diﬀerent systems, both large and small, we can hope to better understand galaxy formation and evolution.

In my thesis I carry out a detailed high resolution spectroscopic study of individual stars in a nearby dwarf galaxy: the Fornax dwarf spheroidal galaxy. In this chapter I provide an introduction to what is currently known about Fornax and the environment in which it is evolving.

2.1 Dwarf galaxies in the Local Group

The Local Group contains ∼40 dwarf galaxies, mostly clustered around two big spiral galaxies, the Milky Way (MW) and M 31 (van den Bergh 2000), see Figure 2.1 for a schematic overview. The majority of dwarf galaxies generally fall into two categories, Dwarf Spheroidals (dSphs) and Dwarf Irregulars (dIrrs). The dSphs are generally found close to a host galaxy and they typically don’t have current star formation or H i gas associated with them and the dIrrs are typically more distant and generally have at least some current star formation and gas (Mateo 1998).

The Local Group is a useful laboratory to study galaxies in detail because, as opposed to high redshift surveys, we can resolve individual stars. This allows us a deeper insight into the evolutionary path galaxies have followed since the earliest times. This can be achieved using a range of techniques, including Colour-Magnitude Diagram analysis and spectroscopic abundances. By observing large and small nearby galaxies in detail we can hope to see evidence of galaxy building processes. Recent signs of this are bursts of 20 chapter 2: Fornax and the Local Group

Figure 2.1: Schematic representation of the Local Group (Grebel 1998).

star formation, tidal debris and on-going mergers, but to ﬁnd evidence of similar events in the distant past we need to uncover more deeply hidden information. One method is to look for unique chemical signatures in stellar abundance patterns which reveal the detailed evolutionary history of star formation in galaxies and can be used to determine how much diﬀerent galaxies have in common with each other throughout the history of the Universe and thus if the assumptions of hierarchical galaxy formation are valid.

Dwarf galaxies offer us the opportunity to study the star formation history and chemical evolution of complete systems that are quite different to the MW and likely to be more similar to (the metal poor and small) galaxies found in the early universe. Their small size also means that to a first approximation, they can be considered as chemically homogeneous “single cell organisms”, creating stars as more of a single unit than a larger galaxy such as the Milky Way; largely unaffected by complexities such as spiral arms, and distinct components such as disk, halo and bulge. There is however the complication that it is relatively easy for a small galaxy to loose metals during a supernovae explosions 2.2: Fornax dSph 21

Figure 2.2: From Coleman & Da Costa (2005), the distribution of Fornax RGB stars, where each star has been convolved with a Gaus- sian of width 100. The outer shell of Fornax is clearly visible, located 1.3◦ north-west of the centre. The ﬁrst shell is too close to the centre to be visible.

(e.g. Mac Low & Ferrara 1999). Due to their proximity to the gravitational potential well of bigger galaxies (like the Milky Way or M 31 for Local Group Galaxies), dwarf galaxies are also more likely to loose gas than to attract it and this may explain the different characteristics of dSphs and dIrrs (Einasto et al. 1974). When a dwarf galaxy falls within the gravitational inﬂuence of a larger galaxy it may loose gas, stars and maybe even globular clusters to the larger host galaxy. It is also plausible that tidal forces will drive the star formation events (e.g. Mayer et al. 2001). There is clearly a range of dwarf galaxy properties in the Local Group, and most recently there is evidence that we have overlooked a large number of extremely faint dwarf galaxies around the Milky Way (e.g. Belokurov et al. 2006a,b; Willman et al. 2005a,b; Zucker et al. 2006a,b).

2.2 Fornax dSph

The Fornax dSph galaxy is a relatively isolated, dark matter dominated dwarf galaxy ∗ 8 9 with a total mass of 10 − 10 M (Walker et al. 2006, Battaglia et al. in prep.), at a distance of roughly 135 kpc (Bersier 2000). It is well resolved into individual stars, and colour-magnitude diagram (CMD) analyses have been made going down to the oldest main sequence turn-oﬀs (e.g. Stetson et al. 1998; Buonanno et al. 1998; Saviane et al. 2000; Gallart et al. 2005). In common with most other dSph, Fornax has no obvious H i associated to it at present, down to a density limit of 4 × 1018cm−2 in the centre and 1019cm−2 at the tidal radius (Young 1999). Unusually for dwarf galaxies, Fornax contains ﬁve globular clusters (see section 2.3).

∗ luminous mass ' 7 × 107, Mateo et al. (1991) 22 chapter 2: Fornax and the Local Group

Figure 2.3: From Dinescu et al. (2004) proper motion study, the projection of the orbit (gray line) of Fornax (F). The black lines represents the orbital paths of Fornax and the LMC over the last Gyr. The dashed line represents the Galactic plane. The Magellanic Stream is represented with H i column density contours (from Putman et al. 2003), down to a column density of 1019 cm−2. The Sculptor dSph (S) and Phoenix dwarf (P) are also marked on this plot.

Traditionally, dSphs are considered to be simple, uniform spherical systems. How- ever, in a wide ﬁeld photometric survey of Fornax, a small overdensity of stars was found located approximately 170 (or 670 pc) south-east from the centre of Fornax apparently dominated by a relatively young stellar population with an age of ∼2 Gyr (Coleman et al. 2004) It is possible that this might be a shell structure, something previously unseen in a dwarf galaxy, which may be the remnant of a merger with a small, gas-rich system that occurred approximately 2 Gyr ago, although a detailed study by Olszewski et al. (2006) suggests that the metallicity of this stellar population is the same as Fornax. A second large shell-like structure has been discovered 1.3◦ north-west from the centre of Fornax, outside the nominal tidal radius (Coleman & Da Costa 2005, see Figure 2.2).

Clearly Fornax exists in a complex environment. Proper motion studies of Fornax, using a combination of photographic plate material and HST Wide Field Planetary Camera 2 data suggest that Fornax crossed the Magellanic plane ∼190 Myr ago (Dinescu et al. 2004, see Figure 2.3). This crossing appears to roughly coincide with the termination of all star formation in Fornax (Stetson et al. 1998). It is possible that ram pressure stripping of the ISM may have caused the end of star formation in Fornax. There are H i clouds found all along the proposed orbit of Fornax consistent with stripped material from Fornax as it crossed the orbit of the Magellanic Clouds (Dinescu et al. 2004). However, there remains a distance discrepancy between Fornax (135 kpc) and the LMC (50 kpc) which makes it diﬃcult to be sure if there was any interaction at all. 2.2: Fornax dSph 23

Figure 2.4: (top): Relative SFR for Fornax, taken from Tolstoy et al. (2001). (bottom): SFR of Fornax inner ﬁeld, taken from Gallart et al. (2005).

Using CMDs from several sources, Tolstoy et al. (2001) constructed a schematic star formation history for Fornax dSph. This is presented in the top panel of Figure 2.4. More recently, a new star formation history has been published by Gallart et al. (2005), for the inner field of Fornax, which we reproduced in the bottom panel of Figure 2.4. The two plots do not agree very well, although both show a peak in star formation at ∼4 Gyr. The relative number of young and ancient stars differs significantly. This might be due to studies covering different regions of Fornax which we now know has quite a lot of spatial variation in stellar population (Battaglia et al. 2006). These differences require further study.

Low resolution spectroscopic studies of individual stars have been made to determine Ca II triplet metallicities for samples of ∼30 stars (Tolstoy et al. 2001), ∼100 stars (Pont et al. 2004) and most recently ∼600 stars (Battaglia et al. 2006). These studies have shown that Fornax contains a relatively metal-rich stellar population and has a complex star formation history where the majority of stars have been created at intermediate ages 2 − 6 Gyr ago with a peak at 5.4±1.7 Gyr ago. (Saviane et al. 2000). Fornax also has a young stellar population (<1 Gyr) as well as an ancient one (10-12 Gyr) and stars with a range in metallicity going from −2.8 dex to solar.

Detailed abundance analyses based on UVES high resolution spectroscopy have also been carried out (Shetrone et al. 2003; Tolstoy et al. 2003) but these were limited to three stars. In chapter 5, we present our abundance results for nine stars belonging to three of the Fornax globular clusters, and in chapter 6, the abundances of an additional 81 stars belonging to the central (250) ﬁeld of Fornax. 24 chapter 2: Fornax and the Local Group

Figure 2.5: The HST Colour Magnitude Diagrams of three of the Fornax Globular Clusters, from Buonanno et al. (1998). The RGB stars for which we obtain HR spectra (see chapter 5) are marked.

2.3 Globular Clusters in Fornax

Globular clusters are rarely found associated to low mass dwarf galaxies. It is not clear if this is because low mass galaxies loose their globular cluster populations, or if they typically don’t form stars actively enough to warrant a globular cluster population. Fornax and Saggitarius are the closest dwarf galaxies with globular clusters. Contrary to the Sagittarius dSph which is obscured by dust, and confused by merging with our Galaxy, the Fornax dSph is high above the Galactic plane and oﬀers a uniquely useful target for investigation. The next example of a dwarf galaxy with globular clusters is WLM (Wolf- Lundmark-Melotte), which is nearly 1 Mpc away, with only one globular cluster. By studying the Fornax globular cluster population we can determine if all globular clusters share the same properties regardless of the environment in which they were created or if there are diﬀerences related to the size, type or location of their host galaxy. At the very least globular clusters are a part of the overall picture of the star formation history of a galaxy.

Unlike most other dSphs, Fornax contains ﬁve globular clusters (Shapley 1938; Hodge 1961), .1 kpc from its centre. This highly unusual speciﬁc frequency (∼70) is an order of magnitude higher than the expected value for a galaxy of its size (. 5, Harris 1991). According to Goerdt et al. (2006), in a cuspy cold dark matter halo, Fornax GCs should sink to the centre within a few Gyr, raising the question of how these old GCs could have survived to the present epoch. As Goerdt et al. (2006) show, a solution to this timing problem is to adopt a cored dark matter halo. Under these conditions, it will take the GCs many Hubble times to sink to the centre, as they will stall at the dark matter core radius.

Globular clusters are typically associated with the oldest stellar population component of a galaxy. We do not know for sure under which conditions they form and survive but they are generally assumed to be a ubiquitous old population associated with the epoch of galaxy formation. Every large galaxy (spiral or elliptical) appears to have a population of very old globular clusters (Harris 1991). It is commonly believed that globular clusters are formed during periods of exceptional star formation, such as during 2.3: Globular Clusters in Fornax 25 the initial formation of the galaxy (Searle & Zinn 1978) or during a major merger (which might be the same thing). However, this view breaks down in the Magellanic Clouds where there are (at least) two distinct populations of globular clusters (one young and relatively metal-rich, and the other old and metal-poor) and there is no similar signature in the ﬁeld star stellar population (van den Bergh 1981).

The Fornax dSph GCs, similarly to Galactic GCs, are single-age stellar populations. Their globular-cluster-like ages have been determined by isochrones fitting on HST CMDs (see Figure 2.5) going down to oldest main sequence turn-offs and are the same to within ± 1 Gyr, with the possible exception of cluster 4, which is buried in the centre. Some studies have found it to be younger by about 3 Gyr (Buonanno et al. 1998). The metallicities of the clusters vary, as summarised in Strader et al. (2003), but are more metal-poor than the average for the field stellar population (by a factor of more than ∼1 dex), with a bluer RGB, well populated blue horizontal branches (HB) and a range of HB morphology (Buonanno et al. 1998, 1999). Thus the Fornax globular clusters represent quite a different stellar population to the Fornax field stars (Stetson et al. 1998; Buonanno et al. 1999; Saviane et al. 2000).

Chapter 3 Using stellar atmospheric models to determine chemical abundances

n order to quantify the different chemical elements present in a stellar atmosphere, we I need to describe the star in a physical way. The light that we receive from a star comes through it’s atmosphere, more specifically the photosphere. Photons emerge from these transparent layers of gas, releasing the energy produced by the thermonuclear re- actions in the star’s opaque centre. The temperature, pressure and chemical composition of the atmosphere will determine the features of the star’s spectrum. Absorption lines are created when a particle (atom or molecule) absorbs a photon from the emerging flux at a specific wavelength. Each different chemical element will absorb photons at specific wavelengths and by measuring the relative depth of these absorption lines we can determine the abundance of that particular element.

The following paragraphs are not meant as a summary of the physics of radiative transport in stellar atmospheres. The theory of stellar atmospheres is a well developed part of astrophysics and the detailed physical processes are well described in standard textbooks such as Gray (1992, chapters 5–14) or Carroll & Ostlie (1996, chapters 9– 10). On the other hand, some general background will help to understand the analysis techniques used in this thesis. Sections 3.1 – 3.4 therefore give an overview of the most important concepts and terminology that will be frequently used in the subsequent chapters. In order to facilitate reading, Table 3.1 lists the physical constants used in this chapter.

3.1 Describing the stellar atmosphere

Stellar atmospheres are low density gas, so the “ideal gas law” may be used to relate the pressure, density and temperature. Here I summarise the physical description of the ﬂux emerging from an ideal atmosphere and the diﬀerent absorption sources present in such an atmosphere. 28 chapter 3: Using stellar atmospheric models ... chemical abundances

Figure 3.1: Volume element illustrating ﬂux and intensity in a stellar atmosphere. Adapted from Gray (1992), ﬁgure 5.1 and 5.3.

Table 3.1: Constants used in this chapter Name symbol value units Speed of light c 2.99792458 × 108 m s−1 Planck’s constant h 6.63 × 10−34 J s Boltzmann’s constant k 1.38 × 10−23 JK−1 8.62 × 10−5 eV K−1 Stefan-Boltzmann’s constant σ 5.6705 × 108 W/m2 K4 Gravitational constant G 6.672 × 10−11 m3 kg−1 s−2 −31 Mass of the electron me 9.11 × 10 kg

3.1.1 The ﬂux

Lets consider a cylindrical volume element of surface dA and thickness dx, as shown in Figure 3.1, radiating at a frequency ν and intensity Iν . The radiation is emitted in a direction θ with respect to the cylindrical axis, per unit area, unit solid angle (dω), unit time and unit frequency. The basic equation that describes radiative transfer in a case like this is the following:

dIν = −Iν + Sν (3.1) dτν where Sν is the source function (Sν = jν /κν ), jν and κν are the emission and absorption R coefficients, τν is the line of sight optical depth (τν = κν ρ dx, with ρ representing the density of matter in the unit volume.) So the flux Fν that is crossing the volume element per unit time and frequency is defined by: 3.1: Describing the stellar atmosphere 29

I Fν = Iν cos θ dω (3.2)

Although the basic radiative transfer relation (3.1) looks extremely simple, this simplicity is very delusive, mainly because the quantity κν involves a large amount of complex physics. In order to solve the transfer equation and arrive at the atmosphere structure and the photospheric spectrum of the star a number of simplifying assumptions are needed:

Hydrostatic Equilibrium: This is the case when pressure forces balance gravity. There is no expansion and no signiﬁcative mass loss.

Thin atmosphere: The thickness of the photosphere is small compared to the radius of the star. Thus, we need only consider the atmosphere as a superposition of parallel planes or “onion shells” (layers) with a single (1D, radial) dimension describing the structure. We may thus assume that the variation of gravity over the thickness of the photosphere is negligible and we can approximate the gravity as a constant.

Local Thermodynamic Equilibrium (LTE): We assume that LTE is a valid approximation for each volume element in the atmosphere. Every layer has a unique temperature (T = T (τν )) and the source function is the Planck function:

2hν3 1 S = B (T ) = (3.3) ν ν c2 exp(hν/kT ) − 1 where c is the speed of light, h is the Planck constant and k is the Boltzmann constant. The LTE approximation allows us to use the following two laws:

• Boltzmann’s Law: To know whether a particular line may occur, you have to know the relative populations of the excited states of the particles in the gas. The relative population of excited states in a gas in thermodynamic equilibrium is given by the Boltzmann Excitation Distribution. The number of atoms of energy level n per unit volume Nn is proportional to the total number of atoms (N) of the same species:

N g χ n = n exp − n (3.4) N Un(T ) kT

th where gn is the statistical weight of the n level, χn is the excitation potential of the th n level and Un(T ) is the partition function of the particle in a gas of temperature T and is deﬁned as: Un(T ) = Σgi exp(−χi/kT ). It is often the case that χn is expressed in eV and the term (1/kT ) is often expressed as θ = log e/kT = 5040/T which lead to: 30 chapter 3: Using stellar atmospheric models ... chemical abundances

Nn gn = 10−θχn (3.5) N Un(T )

• Saha’s Law: In order to describe an absorption line, we need to know what fraction of the atoms of a particular element are in the ionization state corresponding to the line. Saha’s law describes the distribution of particles of the same species in diﬀerent ionization states. The ratio of atoms in ionization state i and i + 1 is related to the electronic pressure (Pe) and temperature T of the gas :

3/2 5/2 Ni+1 (2πme) (kT ) Ui+1(T ) χi Pe = 3 exp − (3.6) Ni h Ui(T ) kT

where me is the mass of the electron and χi is the ionization potential of the ion in the state i. The Pe term in that equation explains why stellar spectra are sensitive to pressure. The assumption of LTE is a very important simpliﬁcation of the general problem, as it allows us to calculate the source function, the population of the atomic energy levels and the ionization equilibria from only a small number of free physical parameters. In very thin extended atmospheres, or in the case of strong absorption lines which are formed in high atmospheric layers, the LTE assumption breaks down. The calculation of the excitation and ionization equilibria then becomes enormously more complicated because all interactions between matter and radiation have to be considered in detail.

Radiative Equilibrium In the top layers of any stellar atmosphere, all the energy is carried by radiation. Conser- vation of energy tells us that the energy absorbed by one layer in the atmosphere must be re-emitted to the next, or in other words, the ﬂux must be constant (F(x) = F0) throughout the atmosphere. In the case of a 1D model, we have:

d F(x) = 0 (3.7) dx where F(x) is the total ﬂux (in W/m2). When all the energy is carried via radiation, we have:

Z ∞ 4 Fν dν = F0 = constant = σTeﬀ (3.8) 0 where σ is the Stefan-Boltzmann constant and Teﬀ is the black body temperature of the stellar atmosphere.

3.1.2 The absorption coefficient Any process that captures or prevents photons from being emitted by the atmosphere will contribute to the absorption coefficient (or opacity). This includes scattering as well as absorption of photons by atomic electrons making level transitions. The absorption coefficient (κν ) of a gas is obviously going to be frequency dependent. 3.1: Describing the stellar atmosphere 31

Continuous absorption This is the sum of the absorption resulting from many physical processes. The wavelength dependence of the continuous absorption coeﬃcient shapes the continuous spectrum emitted by a star. Photoionization, when a photon has enough energy to ionise an atom (bound-free absorption) is a source of continuous opacity. Also free-free absorption (when a free electron in the vicinity of an ion absorb a photon) contributes to the continuous opacity of the star. Electron scattering (Thompson, Compton, Rayleigh) can also divert photons from an incident light source, so they also contribute to the continuous absorption. Hydrogen, being the most abundant element, is also the main contributor to the absorption coeﬃcient. In cool stars like those of our sample (∼ 4000 K) most of the continuous absorption in the visible and infrared part of the spectrum is due to the negative hydrogen ions H− (hydrogen atoms with one very loosely bound extra electron), while “metals” start to dominate the UV part of the spectrum.

Specific absorption Absorption specific to the line, (bound-bound transitions) occurs when an electron in an atom or an ion makes a transition (by absorbing a photon) from one orbital to another. It is, by definition, very wavelength specific, corresponding to the energy of the photon that was absorbed. The depth and width of this absorption line is related to the transition probability, the population of the lower energy level, and the abundance of the element that absorbed the photon, but also to some intrinsic effects not related to the abundance. Natural broadening is caused by Heisenberg’s uncertainty principle, where the orbital energy cannot have a precise value, allowing for photons of slightly different wavelength to be absorbed. This results in a non-discrete (fuzzy) energy level. Thermal (or Doppler) broadening is caused by the fact that atoms are in thermal motion, producing a range of line of sight velocities. This motion will change the observed fre- quencies (Doppler shifting) of the absorbed photons, making the line broader. There is also pressure broadening, caused by the electric field of a large number of (close by) ions and collisional broadening, when the orbitals of an atom are perturbed due to collision with a neutral atom.

Macro and micro turbulence are two broadening mechanisms that act on scales that are large (macro) or small (micro) compared to the mean free path of the photons. Microturbulence can be considered as an additional thermal velocity. When the line of sight goes through many cells of motion (turbulence cells) in the photosphere, the velocity of the cells will modify the line profile in the same way as the particle distribution. It is approximated to be isotropic (Gaussian) and can be included directly into the line absorption coefficient with a convolution, as detailed in chapter 18 (p.405) of Gray (1992). There is macroturbulence when the turbulence cells in the photosphere are large enough so that a photon will stay in the same cell from the time it is created to the time it leaves the star. Each of these cells will have the same Doppler shift, corresponding to the velocity of the cell, therefore acting in a way similar to rotation, which can be applied as a convolution of the emergent spectrum by an appropriate function (Gaussian or other). To summarise, micro turbulence acts on the absorption line profile, like a thermal component desaturating strong lines while macro turbulence acts on both strong and weak lines in the same way by smearing them out over a frequency range. 32 chapter 3: Using stellar atmospheric models ... chemical abundances

Figure 3.2: Electronic pressure (top), gas pressure (middle) and temperature (bottom) as a function of optical depth. This figure present models with constant log g and [Fe/H] in order to illustrate the Teff dependence of the models. Teff start at 3800 K (solid line) and increase by 100 K each time to reach 4200 K (dotted line). The models used are those of MARCS 2005 and Plez 2005, presented in section 3.3.2. 3.1: Describing the stellar atmosphere 33

Figure 3.3: Electronic pressure (top), gas pressure (middle) and temperature (bottom) as a function of optical depth. This ﬁgure present models with constant Teﬀ and log g in order to illustrate the [Fe/H] dependence of the models. [Fe/H] start at -2.5 dex (solid line) and increase by 0.5 dex each time to reach -0.5 dex (dotted line). The models used are those of MARCS 2005, presented in section 3.3.2. 34 chapter 3: Using stellar atmospheric models ... chemical abundances

Figure 3.4: Electronic pressure (top), gas pressure (middle) and temperature (bottom) as a function of optical depth. This ﬁgure present models with constant Teﬀ and [Fe/H] in order to illustrate the log g dependence of the models. The log g start at 0.0 dex (solid line) and increase by 0.3 dex each time to reach 1.2 dex (dotted line). The models used are those of MARCS 2005, presented in section 3.3.2. 3.2: Determining Stellar Atmospheric parameters 35

3.1.3 Stellar atmospheric models Stellar atmospheric models are a tabulation of physical parameters used to represent the conditions inside an atmosphere. Models are typically given as the electronic pressure (Pe), the gas pressure (Pg), the temperature (T ) and the optical depth for photons with λ=5000Å (τ5000) for several layers (∼50) of a stellar atmosphere. This is shown in Figures 3.2, 3.3 and 3.4 where we plot Pe (top), Pg (middle) and T (bottom) as a function of τ5000 for five different set of parameters, varying Teff , [Fe/H] and log g respectively, sampling the full range of stellar parameters we used in chapter 6. It is customary in stellar atmosphere work to use log g as equivalent for the pressure in the atmosphere (which can be done if hydrostatic equilibrium is valid). In these three figures, where we can see the similarity in the shape of the curves when only one parameter changes, allowing us to interpolate between two curves to get the exact parameter needed for our model. As can be seen from the plots of T versus τ5000, the parameter that will most influence the line formation is Teff . Especially in the region where most of the lines are forming, (−1 < τ5000 < 1), a change of 100 K will change the T (τ) relation much more than a change in log g and/or [Fe/H]. Therefore it is critical to have stellar models made for the Teff corresponding to the star observed in order to produce accurate abundances. These models are needed as an input for the line formation code used to derive the abundance, as describe in section 3.3.3.

3.2 Determining Stellar Atmospheric parameters

In the previous section, we have shown that we can simplify our stellar atmosphere model so that it can be described using only a few parameters. We will describe them (and how to derive them) in this section.

3.2.1 Eﬀective Temperature (Teﬀ )

The effective temperature, Teff , is the temperature of a black body radiating as the star, 4 F = σ Teff . There is more than one way to determine the Teff of a star, and I will describe the two methods we used.

Photometric colour

A powerful method to obtain the Teff of a star, is the InfraRed Flux Method, (IRFM) described by Blackwell & Lynas-Gray (1998). It provides an accurate procedure to derive stellar angular diameters and effective temperatures by measuring the monochromatic flux at an infrared frequency and the bolometric flux. It then uses theoretical atmospheric models to estimate the monochromatic flux at the star’s surface (the infrared flux has a small dependency on the Teff ). The IRFM is an iterative procedure: from a first guess of the Teff , the angular diameter is deduced and used to derive an improved Teff . Practically, the IRFM is not easy to use since measuring a stellar angular diameter is not always possible. An empirical method, calibrated on the IRFM, has been developed to give a relation between photometric colours (like V − I, V − K) and Teff . The general method and correction polynomials are described in the series of papers by Ramírez & Meléndez (2005), Alonso et al. (1999a,b, 2001), and references therein, and we will use these calibrations in the following chapters. 36 chapter 3: Using stellar atmospheric models ... chemical abundances

Excitation equilibrium

We define Teff such that the abundance of an element is independent of the excitation potential (χex) of the individual lines. Obviously, in principle, all the lines of an element should give the same abundance for a given star. In practice, there is a (small) scatter around an average value. A Teff which is incorrect will affect the weak excitation potentials more than the strong potentials. It will also change the gradient of the temperature and the optical depth (T and τ5000) and since lines with different χex are not all forming at the same depth, the resulting abundance will be different. If a correlation between abundance and excitation potentials occurs, it is a sure sign that Teff has been incorrectly determined for the star. In order to use this method, we need many lines of a single element sampling a range of χex. Figure 3.5 (middle panel) illustrates this for a sample star of our FLAMES dataset, with the proper Teff , where the slope is effectively zero. Contrasting with this figure, we show in Figure 3.6 the same figure but with a Teff that is 400 K higher than our ”correct“ one. The precision with which Teff can be determined depends upon the resolution, the choice and number of lines and signal to noise of each spectrum used. Usually, we use ≈50 Fe i lines to determine the Teff of a star.

3.2.2 Surface Gravity (log g) 2 The surface gravity of a star of mass M? and radius R? is deﬁne as g? = GM?/R?, where 2 G is the gravitational constant. In solar values, we get g? = g (M?/M /)/(R?/R ) . There exists several methods (isochrones, pressure broadening in the wings of strong lines) to estimate the gravity of a star (which we often use in log scale, log g) and in this section we describe the two diﬀerent methods we used to calculate it.

Photometric We can use photometry to estimate the surface gravity of a star if we know the mass, the Bolometric magnitude (distance modulus + a bolometric correction) and the Teﬀ (two or more photometric colours). We do so using the relation linking the luminosity, radius and temperature of a star (L = 4πR2σT 4) and Bolometric magnitude deﬁnition (MBol? − MBol = −2.5 log(L?/L )) we get:

M? Teﬀ? log g? = log g + log + 4 × log + 0.4 × (MBol? − MBol ) (3.9) M Teﬀ

Spectroscopic (Ionization Equilibrium) For stellar types F, G or K, we can easily measure elements in two ionization states, like Fe i and Fe ii or Ti i and Ti ii. For a given star and a given element, there should be a single value for the abundance, no matter if the abundance is determined from the neutral or the ionized state. We can then iterate on the gravity of our model until the abundance of Fe i and Fe ii are the same, constraining the stellar gravity. By deﬁnition, 2/3 gravity is related to the gas pressure (Pg ∝ g ) and therefore to the electronic pressure 1/3 2 (Pe ∝ g ) since Pg ∝ Pe . From the Saha’s equation (3.6) and the cool stars case (our case), where the number of atoms of Fe i Fe ii, we can state that Fe i, the dominant species, will depend on 1/Pe and Fe ii, the minority species, with the majority of atoms 2 in the state i−1 = 1, will depend on 1/Pe . This is what makes the ionization equilibrium 3.2: Determining Stellar Atmospheric parameters 37

Figure 3.5: Fitting a model (MARCS 2005) to a star in our FLAMES sample (BL239), showing the [Fe i/H] (full symbols) and [Fe ii/H] abundances (empty symbols) as a function of λ,(top) χex (middle) and EW (bottom). The text above the top plot gives the parameters of the model used: star name, Teff (T), log g (g), vt (v), metallicity (z), the average [Fe i/H] (also plotted with a dashed line) and the number of Fe i lines used. The thick line in the middle and bottom plots are linear regressions, with their respective coefficients and associated errors on top of each plot. The different symbols for the Fe i lines are related to their equivalent width, as shown in the bottom panel. 38 chapter 3: Using stellar atmospheric models ... chemical abundances

Figure 3.6: Same as Figure 3.5 but with a Teff that is 400 K higher than the ”correct“ one. The slope in the middle plot, although only significative at the 1.8σ level, shows that there is a tendency for weak χex to produce higher abundances. This extreme change in Teff changes the derived [Fe i/H] by ∼0.2 dex, and shift the [Fe ii/H] abundances (empty squares) from relatively comparable to the [Fe i/H] to way below them, another hint that this temperature is not appropriate for this star. 3.3: The abundance determination 39 a good tool to constrain gravity. This method presumes that non-LTE effects are not modifying the ionization equilibrium. This is an assumption that is not always correct, especially at low surface gravities or metallicities, where Fe is known to be overionised in non-LTE (Asplund 2005). This translates into an underpopulation of Fe i levels with respect to what was predicted by LTE and therefore an [Fe i/H] abundance lower than [Fe ii/H]. Although it is claimed by many authors (including Asplund (2005)) that Fe ii is immune to departures from LTE in late-type stars, the typical number of Fe i lines observed versus Fe ii lines (factor ∼10) makes Fe i a more reliable measure of Fe than Fe ii.

3.2.3 Metallicity Each stellar atmosphere model is computed with a parameter representing the chemical composition of the star. It is often referred to as [Fe/H], although it does not only represent the contribution of Fe atoms. It represents the electronic pressure (Pe) in the atmosphere of a star with the same chemical element ratios as the Sun, scaled to a given [Fe/H]. It is the abundance of the elements that contribute to the continuous absorption properties of the atmosphere. A higher metallicity will increase Pe in the atmosphere by contributing extra electrons. Some models have diﬀerent [α/Fe] ratios to represent stars that are systematically diﬀerent from the sun.

3.2.4 Microturbulence velocity

The microturbulence velocity, vt affects the lines by broadening and hence desaturating them. It is caused by small cells of motions in the photosphere and is treated like an additional thermal velocity in the line absorption coefficient. The desaturation effect depends on the strength of the line, where only strong lines are affected. Weak lines will not be affected by desaturation since increasing the vt will broaden the line and make it shallower, conserving the equivalent width. In this regime, the abundance is proportional to the EW . But for a saturated line, increasing the vt will widen the wavelength range covered by the absorption, thus desaturating the line: the equivalent width is not conserved anymore. This is detailed in in chapter 18 of Gray (1992). Typical values for the vt are 1-2 km/s for low-mass giants. We can determine vt for a star by making sure that for a single element, the abundance is independent of the EW of the line, as illustrated in Figure 3.5 (bottom panel) for a sample star of our FLAMES dataset, for which the vt has been chosen correctly for the observed spectra (negligible slope). Again, Fe i, having many observed lines, is the most suitable element for this determination.

3.3 The abundance determination

After describing the physics of a stellar atmosphere and parameterising it, we need to transform the absorption lines into chemical abundances. We will ﬁrst need to measure the equivalent widths of the absorption lines, and then use a stellar atmosphere model with the right parameters for each individual star before being able to derive an abundance. 40 chapter 3: Using stellar atmospheric models ... chemical abundances

Figure 3.7: The equivalent width of an absorption line is deﬁned as the width of a rectangle that has an area equal to the line, as illustrated in grey. Adapted from Figure 9.18 of Carroll & Ostlie (1996)

3.3.1 Measuring the equivalent widths The ﬁrst step in determining the abundance is the actual measurement of the strength of each absorption line. We refer to this value as the equivalent width (EW , in text or W , in equations), it corresponds to the total absorption coming from a line, and it is deﬁned in the following way:

Z +∞ Fc − Fν EW = W = dFν (3.10) −∞ Fc Where EW is the width of a rectangle of depth 100% (going from 0 to 1) in a normalised spectrum that covers the same area as the real line. This is illustrated in Figure 3.7, where Fc is the ﬂux level of the continuum (normalised at 1), Fλ = Fν is the ﬂux at the frequency ν = c/λ, with c = the speed of light.

3.3.2 The Stellar Models used In 2005, a major improvement in the models available occurred with the release of new MARCS spherical stellar models∗ which are described in in Gustafsson et al. (2003). For chapter 5, (which was made in prior to 2005) we used models from Plez (2000, 2002). For chapter 6, we used the new MARCS 2005 models extended by Plez (2005) to cover the range of stellar parameters of our sample. Here is a summary of the models used:

Plez 2000-2002 • Geometry: Plane-parallel approximation

• Temperature: 3800 ≤ Teﬀ ≤ 5200 K in steps of 200 K • Gravity: 0.5 ≤ log g ≤ 4.5 dex in steps of 0.5 dex

∗ http://marcs.astro.uu.se/ 3.3: The abundance determination 41

• Metallicity: −4.0 ≤ [Fe/H] ≤ −1.00 dex in steps of 0.25 dex • Alpha: Enhanced, [α/Fe] = 0.4 MARCS 2005 • Geometry: Spherical

• Temperature: 4000 ≤ Teﬀ ≤ 5500 K in steps of 250 K • Gravity: 0.0 ≤ log g ≤ 3.5 dex in steps of 0.5 dex • Metallicity: −1.5 ≤ [Fe/H] ≤ +1.00 dex in steps of 0.25 dex • Alpha: Standard, [α/Fe] = 0 at [Fe/H] = 0, +0.1 for each -0.25 dex until it reaches +0.4 at [Fe/H] ≤ -1.0. Plez 2005 • Geometry: Spherical

• Temperature: 3600 ≤ Teff ≤ 4000 K in steps of 200 K • Gravity: same as MARCS 2005 • Metallicity: −3.0 ≤ [Fe/H] ≤ −1.50 dex in steps of 0.5 dex • Alpha: Poor, [α/Fe] = 0.00 for all models. Models are interpolated for all parameters, including [α/Fe] when mixing standard models with α-poor ones, in order to create the correct model for individual stars. We used models with two types of geometry: plane-parallel and spherical, but for a given sample, we used either one or the other for the entire analysis. The geometry affects the abundance in two fundamental ways, namely the line formation (discussed in section 3.3.3) and the model atmosphere structure. Spherical geometry in the model structure is a better representation of reality but as they are relatively new models they have not been used intensively in the literature, since prior to the MARCS 2005 models, the most commonly used models for abundance analysis in giants stars were those of Gustafsson et al. (1975), in plane-parallel. This makes a direct comparison with previous work more complex but since these models became available, we decided to start to use them. An overview of the difference between models with spherical geometry and plane-parallel approximation is presented in Heiter & Eriksson (2006), where they compare the effect of using the plane-parallel approximation in the model atmosphere structure and/or the line formation code (p_p and s_p) using fully consistant spherical geometry (s_s) in the abundance analysis. They show that lines with different χex will not behave in the same way (introducing a bias on the determined Teff ), lines with different EW will also show a different behaviour (affecting the vt) and that lines of different ionization state (Fe i versus Fe ii) will also react differently to a change in geometry (bias in log g). They give the maximum combined systematic error caused by different geometry in the s_p case with respect to the s_s case to be of the order of -0.1 dex, while for the p_p case, the differences are up to +0.35 dex. Thus using spherical models can make a big difference in the abundance determination, much more than the way we treat the line formation (which in our case is plane-parallel). 42 chapter 3: Using stellar atmospheric models ... chemical abundances

3.3.3 Computing the abundances As explained in detail in chapter 14 of Gray (1992), for weak lines (dominated by Doppler broadening) we can show that:

2 Wλ πe Ni/N 5040 log = log 2 NH + log A + log(gf λ) − χ − log(κν ) (3.11) λ mec U(T ) T where Wλ is the equivalent width of the line, e is the charge of the electron, e = −1.60 × −19 10 C; Ni/N is the ratio of the number of atoms of a particular element in the ionization state i with respect to the total number of atoms of that element, NH is the number of hydrogen atoms per unit volume, A = N/NH is the abundance of the specific element relative to hydrogen, Un(T ) is the partition function, defined in equation 3.4, κν is the continuous absorption coefficient and gf is the transition probability∗. Note that the first term on right hand side of the equation is constant for a given star and a given ion. This equation gives us some general information about the abundance of an element in a star:

• for weak lines, the equivalent width (W ) varies in a linear way with abundance.

• the abundance of an element (A) varies with the inverse of the temperature (5040/T ).

• the abundance depends linearly on the gf -values.

The line formation code To calculate the abundances we use CALRAI, developed by Spite (1967) with many improvements over the years, which uses the plane-parallel approximation. Once we have determined the appropriate stellar parameters for a given star, we can interpolate the model from our grid and use the measured EW s to determine the abundance of the diﬀerent elements. This is an iterative process, where the software will vary the abundance of a given element until it is consistent with the observed EW s. Once this has been done for all the lines of a given element, we obtain a distribution of abundances for each element. The more lines of a single element we have the greater the reliability of the abundance derived. The diﬀerent lines do not necessarily behave in the same way with respect to abundance. Weak and moderately strong lines will vary in an almost linear way with respect to the abundance. When lines become saturated (without prominant wings), a strong variation in abundance will be almost insensitive to the EW . A curve of growth can be used to illustrated this dependence.

The Curve of Growth

For a given abundance (α?) of a speciﬁc element, the EW s (or Wλ) of a given line will vary as a function of the gf -value (the transition probability, eq. 3.11). When we have a weak line, log (Wλ/λ) will vary linearly with log (α?gf). We can deﬁne Γ? for weak lines when this equation is true:

∗ where g is the statistical weight (2J+1, J is the inner quantum number) of the lower level, and f is the oscillator strength 3.4: The line list 43

W log λ = log (α gf) + log Γ (3.12) λ ? ?

We can calculate Γ? for each line as a function of the model used, the element and its ionization state and its χex. A curve of growth is a plot of log (Wλ/λ) versus log (α?gf) + log Γ?. Figure 5.3 of chapter 5 illustrates this for one star in our sample. Since we don’t know a priori the value of α?, we use the solar value, α and the equation becomes log (α gf) + log Γ?. If we consider Fe for example, then all the lines of Fe i and Fe ii will be aligned on a curved shifted by log (α?) − log (α ) with respect to the theoretical curve of growth. The value of this shift is, by definition, [Fe/H]. When we compare the observed curves of growth for Fe i and Fe ii (one for each state) we can verify if the chosen gravity is representative of the star. The gravity of the model will be representative of the star if the two different (ionization) states fall on their respective curve. By plotting different symbols for lines of strong and weak χex, we can confirm our choice of Teff . Lines of different χex should be randomly distributed higher and lower than the curve. If it’s not the case, it is a sign that the Teff chosen for the model is not a good match to the star (so we have to change it). Also, if the vt is not the right one for the star, the non-linear part of the curve of growth will not fit the observed lines.

3.4 The line list

High resolution spectroscopy can provide accurate measurements of numerous absorption lines for many different chemical elements and hence is an accurate method of determining detailed abundance patterns in a star. The line list is a critical part of the analysis, and building a proper line list is a complex task. Lines needs to be chosen carefully, making sure that they have a reliable gf -values and are sufficiently isolated from their neighbours at the resolution of the observations to be accurately measured. A line that is isolated for a star of a given metallicity (or Teff ) can be too blended to be useful at another metallicity. Also, a line that is isolated in a high resolution instrument (like UVES for example) can be blended in a lower resolution one (like GIRAFFE). In addition, because of different wavelength coverage, some lines are available for one instrument but not for another. So a line list needs to be adapted for the data set it will be used on.

3.4.1 Building a line list One way to build a line list is to start from basic reference line list and add/substract lines from it. Such a basic line list can be found in recently published work made on similar objects. Then, extra lines can be added from other work and/or web-based atomic libraries like NIST∗, VALD† or Kurucz‡. It is always important to check that there is compatibility in the average values of the abundances derived from each list. A small but noticeable oﬀset in the average value of [Fe i/H] between two lists will add an ar- tiﬁcial scatter to the abundances values. The consistency (precision and uniformity) of the atomic data used for each list – especially the gf -values – has to be checked before

∗ http://physics.nist.gov/PhysRefData/ASD/lines_form.html † http://ams.astro.univie.ac.at/vald/ ‡ http://cfa-www.harvard.edu/amdata/ampdata/kurucz23/sekur.html 44 chapter 3: Using stellar atmospheric models ... chemical abundances

Figure 3.8: ([Fe i/H]) abundance determination for Arcturus using four different line lists, represented by different symbols. Also plotted are the average abundance for each of the sets of lines. Lines tagged Shet2003 (filled circles) are from Shetrone et al. (2003), Hill2000 (empty squares) from Hill et al. (2000), Grat2003 (gray diamonds) from Gratton et al. (2003) and Zoca2004 (empty triangles) from Zoccali et al. (2004).

we make a big list out of smaller ones. For some elements, the gf -values are not known precisely and the uncertainties are much bigger than what we would hope to achieve with our measurements. The gf -values aﬀect the abundance in a direct way, an error of 0.1 on the log gf will aﬀect the abundance by 0.1.

Our strategy was to start with the reference line list from Shetrone et al. (2003), which was optimised for use with UVES on metal poor stars (−3.0 ≤ [Fe/H] ≤ −1.5). For our work on the globular clusters of Fornax, (chapter 5), we used it without any modifications, as it was already appropriate for our metallicity range and instrumental resolution. But for the Fornax field stars (chapter 6), because of the higher metallicity of the stars and the different (smaller) wavelength coverage, we had to significantly adapt the line list. We selected lines from Gratton et al. (2003), Hill et al. (2000) and Zoccali et al. (2004) as potential candidates to be included in our master line list. We used a high resolution, (R ≈ 120 000) high signal-to-noise spectrum of Arcturus, downloaded ∗ from the ESO UVES public archive . Arcturus has Teff 4250 K and [Fe/H] ∼ −0.5 and

∗ http://archive.eso.org/wdb/wdb/eso/uves/form 3.4: The line list 45 is thus a good “template” for our most metal rich stars. On average, as you increase in metallicity, the strength of a line will increase, making weak lines stronger and saturating already strong lines. Our reference line list was optimised (i.e. having both weak and strong lines) for metal poor stars. In order to optimise our line list also for the richest part of our sample, we need to add lines that are weak in metal rich stars.

From the Arcturus spectrum we calculated abundances using the four line lists for all the elements and compared the results before accepting new lines. We illustrate this for the Fe i lines in Figure 3.8. From this we can determine if the inclusion of a line list will improve the abundance determination or if it will just add more scatter. If there is not a big difference in the average value between two lists and/or if a list has a scatter comparable to other lists, we can safely add the new lines to our master list. From Figure 3.8, we concluded that the lines used by Hill et al. (2000), apart from being too strong, would only add scatter to our Fe i distribution. Lines from Zoccali et al. (2004) have an average Fe i abundance difference of ≈ 0.2 dex compared to the Shetrone list, a sign that the gf -values of the two lists are not compatible. It is possible to shift the gf -values so that the average abundance comes out the same as the Shetrone one but since the Zoccali list has almost no weak lines, we didn’t do it and didn’t use their lines. Only the lines from Gratton et al. (2003), because of their relative weakness (30-70 mÅ) and small scatter, were added to our master list. Since the difference in average log gf is small (0.04 dex), we decided to ignore this difference and use the gf -values without modification.

3.4.2 The line by line selection To facilitate the addition of “clean” lines to our line list, we used two synthetic spectra that have been made with models that have the stellar parameters of Arcturus, Teﬀ =4250, [Fe/H]=0.5, log g= 1.5, one containing most of the known atomic and molecular lines convolved to our instrumental resolution and the other an exact duplicate but without the lines that we are interested in, as shown in the top panel of Figure 3.9. This allows us to detect lines that are contaminated by lines of other elements (superposed or blended) at our resolution. Lines don’t need to be fully isolated, as it’s possible to deblend lines when we measure the EW s, but we just need to check that a weak line is not blended and eﬀectively lost in the wing of a strong and saturated line. By making an identical synthetic spectrum but without the lines we are interested in, we can check what the spectrum look like without those lines. If it goes to zero at the position of those lines, then there is nothing known that directly contaminates the line. If there is a large residual, then we have to reject the line. Of course there is always the possibility that there is something unknown contaminating our lines, and having many lines to determine the abundance of an element can protect us against this unknown. By looking at Arcturus spectrum, (bottom panel of Figure 3.9) we can judge if our synthetic spectra are realistic and that there are not many unknown features.

Now that we have the means to determine accurate abundances, we will, in chapter 4, explain in detail how to apply this theory into practice with our FLAMES data set. 46 chapter 3: Using stellar atmospheric models ... chemical abundances

Figure 3.9: (top) Synthetic spectra of a star with stellar parameters similar to Arcturus (Teff = 4250, [Fe/H] = 0.5, log g = 1.5). The thin line contains all known atomic and molecular lines. The thick one has the same but without the lines present in our line list. The resolution of this spectrum is R ≈ 20 000. (bottom) UVES archive spectrum of Arcturus at a resolution of R = 120 000. The Fe i line in the centre and Si i line on the left were kept for the analysis, despite the Fe i line not being fully isolated at our resolution. The Fe ii line right next to it was rejected since it is a residual flux directly underneath it. The strong V i on the right was also considered usable, but flagged as “unsure”, due to the proximity of another potentially strong line. Chapter 4 Abundances with the FLAMES multi-fibre instrument

s described in chapter 3, the spectroscopic determination of chemical abundances A in individual stars using atomic absorption lines can be obtained by using stellar atmosphere models to predict the abundance of an element based on the relative depths of absorption lines. The most accurate way to determine the iron abundance of a star is to directly measure the strength of as many iron lines as possible. This is typically done on a spectrum of high resolution (HR), with a resolving power R ∼ 40 000 (∼ 0.125 Å at λ = 5000 Å). For precision, it is necessary to have as many lines as possible, therefore a large wavelength coverage (∼ 2000 Å) is required. These are the optimal conditions for a detailed abundance analysis of individual stars. In this chapter, I will summarise the added complexity of making such an analysis with an instrument (FLAMES) that has only half the resolution (R ∼ 20 000), a smaller wavelength coverage but a high degree of multiplexing that requires the automation of many tasks usually carried out star by star. Our FLAMES observations include the three Fornax stars observed by Shetrone et al. (2003) with UVES. In this chapter and in chapter 6, we will use them as a reference point for comparison and will refer to them as the (three) Shetrone stars. Presented in Table 4.1 are the corresponding ID from our work & their work.

Table 4.1: Name and coordinates of the 3 Shetrone stars. Our ID Shetrone RA(J2000) DEC(J2000) BL239 Fnx-M25 02 39 47.09 -34 31 49.8 BL266 Fnx-M12 02 40 10.00 -34 29 58.8 BL278 Fnx-M21 02 40 04.38 -34 27 11.3 48 chapter 4: Abundances with the FLAMES multi-fibre instrument

Figure 4.1: (top): Comparison of two spectra of the same star (BL239) at diﬀerent resolutions. The spectra are normalised to have their continuum ﬂux equal to 1, as indicated by the horizontal line. The GIRAFFE (thick line) spectrum has a lower resolution compared to UVES (thin line). The dashed line is placed on top of an Fe i line of medium strength (EW ≈ 100 mÅ.) (bottom): Same as top panel but the UVES spectrum is convolved to the GIRAFFE resolution.

4.1 UVES vs FLAMES

In this thesis, two HR instruments have been used: UVES (chapter 5) and FLAMES (chapter 6). UVES is the type of instrument that is classically used to make high resolution abundance analysis of individual stars; it has a high resolution over a large wavelength range. UVES typically looks at one star per exposure, whereas FLAMES /GIRAFFE∗, allow us to observe 100+ stars per exposure but with a factor two lower in resolution, as illustrated in Figure 4.1 (top panel) where we show two spectra of the same star. The GIRAFFE spectrum is from our sample and the UVES spectrum is from the work of Shetrone et al. (2003), convolved to the GIRAFFE resolution (bottom panel) for visual comparison.

∗ GIRAFFE is the dedicated intermediate resolution spectrograph on FLAMES 4.1: UVES vs FLAMES 49

4.1.1 UVES

The Ultraviolet and Visual Echelle Spectrograph (UVES) has two arms, the blue (UV to blue) and the red (visual to red). For our observations of individual stars in the Fornax globular clusters (chapter 5) we used the red arm of UVES, covering wavelengths between 4800 − 6800 Å. Shetrone et al. (2003) used exactly the same setup for the three comparison stars. We used a slit of 1-arcsec, giving us a resolving power of about 40 000. The diﬀerent orders of the echelle spectra are dispersed onto two CCDs. More information on UVES can be found in Dekker et al. (2000). Nothing more will be said about UVES in this chapter, (more details are in chapter 5) but it will be used as a point of reference for the FLAMES analysis, to compare our abundance results to a “classical” reference.

4.1.2 FLAMES

The Fibre Large Array Multi Element Spectrograph (FLAMES) is a fibre-fed, multi- object instrument connected to GIRAFFE, a spectrograph which has a medium-high resolution (R = 7500 − 30 000). GIRAFFE is not an acronym, the name comes from the original design concept, where it was standing vertically on a platform. It can cover the entire visible range 3700 − 9000 Å, but not at the same time. The higher the resolution the shorter the wavelength coverage. It has two gratings, one low (LR) and one high- resolution (HR). On a given exposure, only a single order can be observed, referred to as a setup. In total, there are 24 HR setups and 8 LR setups. Different observing modes are available, ARGUS, Integral Field Unit (IFU) and MEDUSA. We only used the MEDUSA mode, which consists of up to 132 individual fibres that can be “placed” on the sky. Each individual fibre has a footprint (aperture) of 1.2 arcsec on the sky. The resulting spectra are projected onto a CCD with a spatial scale of 0.3 arcsec/pixel. More information about FLAMES is available in Pasquini et al. (2002). For our observations of the Fornax field stars, (chapter 6) we used FLAMES/GIRAFFE in MEDUSA mode with setups HR10, HR13 and HR14. In Table 4.2 we give an overview of the different setups used, with the old/new definition for HR14 referring for observations taken before/after October 10th 2003. On that date, HR14 grating was changed for a higher-efficiency version.

Table 4.2: Wavelength coverage and resolution of FLAMES/GIRAFFE in MEDUSA mode for setups HR10, HR13 and HR14. (Eﬀective coverage may be slightly smaller).

setup λmin(Å) λmax(Å) R HR10 5339 5619 19800 HR13 6120 6406 22500 HR14 (old) 6383 6626 28800 HR14 (new) 6308 6701 17740 50 chapter 4: Abundances with the FLAMES multi-fibre instrument 4.2 The FLAMES Spectra

4.2.1 Extracting, calibrating We used a pipeline made by the Geneva Observatory, girBLDRS∗ (GIRAFFE Base-Line Data Reduction Software) to extract and calibrate our spectra, except for the subtraction of the sky emission lines and continuum, for which we used a software written by Mike Irwin. A master sky spectrum is made out of the 10-20 sky spectra taken during the observations, is split into line and continuum components that are scaled and aligned before being subtracted to the object spectrum (more details in Battaglia et al. in prep). The spectral information needs to be extracted taking into account that the light does not necessarily follow a straight line on a CCD row or column and the pixel/wavelength ratio varies with wavelength. There is more than one spectrum on the CCD, corresponding either to another star, an “empty sky” or a calibration lamp. There are five fibres tar- getted on a calibration lamp (Thorium-Argon) for simultaneous wavelength calibration, which are used for the zero-point correction with respect to the daytime wavelength and to adjust the transverse PSF used in the “optimal” extraction method. The flat fielding is done on the extracted spectra (NFF, narrow flat fielding). After the extraction, the spectra can be rebinned into a constant wavelength increment per pixel.

4.2.2 Combining Our data consist of multiple exposures of the same stars in three setups, as can be seen in Table 4.3. To reach the signal to noise required for our analysis, all the spectra of each star need to be stacked together. Before co-adding, the spectra need to be on the same rest frame and since our observations were made during two distinct periods, Septem- ber 2003 and January 2004, the observed radial velocities (Vrad) were diﬀerent. Our spectra were corrected for the heliocentric motion before they were co-added. Details of how we determined the Vrad and the heliocentric correction of our stars are available in section 4.2.3. Within each observation run, the heliocentric corrections varied very little, by amounts typically smaller or of the same order as the uncertainty on the derived radial velocities (∼0.5 km/s), and inducing a completely negligible shift compared to the typical line width in GIRAFFE (FWHM ∼15 km/s).

The combining of the heliocentric corrected spectra was carried out with the IRAF† task scombine, using a flux weighted average with median sigma clipping (cosmic ray removal) for each period, in order to create two master spectra, one for September and one for January. These two stacked spectra were then combined in a flux weighted average. For the two HR14 set of spectra with different resolution and coverage, we used only the common overlapping section, (∼6400-6600 A) convolved the higher resolution HR14 (old) to the resolution of HR14 (new) and treated them like our other spectra. Figure 4.2 show the extracted spectra of the three HR setups used in our observations, HR10, HR13 and HR14. By looking at the top panel of Figure 3.5, we can verify that the individual Fe i abundance of this star are scattered around the same central value, meaning that the three continuum levels are compatible.

∗ http://girbldrs.sourceforge.net/ † http://iraf.noao.edu/ 4.2: The FLAMES Spectra 51

Figure 4.2: Final spectra of star BL239, in each of the three GIRAFFE setups, HR10, HR13 and HR14. Overplotted on it is the continuum that DAOSPEC used for the EW measurement. 52 chapter 4: Abundances with the FLAMES multi-fibre instrument

Table 4.3: FLAMES Exposure time log date Exposure time (s) HR10 HR13 HR14 2003-09-29 0 14400 6225 2003-09-30 3600 0 14102 2003-10-01 10800 0 6900 2004-01-14 3600 0 0 2004-01-15 0 3600 0 2004-01-19 0 7200 0 2004-01-20 0 3600 0 2004-01-21 0 0 7200 2004-01-22 0 0 7503 2004-01-23 3600 0 3600 2004-01-24 3600 0 0 2004-01-26 3600 0 0 Total 8h 8h 12h39m

4.2.3 Determining the radial velocities (Vrad)

In total, we have six independent measurements of the Vrad per star, one per setup (HR10, HR13 and HR14) and one per period (September and January). Listed in Table 4.4 are the final Vrad(weighted average of the six values) with their associated error (the standard deviation of the six). These measurements were made with the girBLDRS routine called giCrossC.py, which takes a template spectrum (G2-type star in our case) and cross- correlates it with each observed star. There were no systematic difference from setup to setup, nor for epoch to epoch. Our mean heliocentric velocity (Vrad) is 55.9 km/s with a line of sight velocity dispersion σ = 14.2. The typical (median) error we have on a velocity (see Table 4.4) is ' 0.55 km/s. We can compare our values to the larger sample of Battaglia et al. (2006), with a mean heliocentric velocity Vrad = 54.1 ± 0.5 and a line of sight velocity dispersion of σ = 11.4 ± 0.4. Figure 4.3 shows the histogram of the distribution of Vrad, with the central velocity and the 3σ membership cut-off, in solid lines (this work) in dashed lines for Battaglia et al. (2006) values. Only one star is a clear non member (BL109), with a negative Vrad, and three others are just on the edge of membership. These three stars have been analysed as if they were confirmed members, since they were on the edge of our membership cut-off. Our subsequent analysis shows them to be consistent with RGB stars and not foreground dwarf stars.

4.2.4 Measuring the Equivalent Widths EW s are classically and arguably most reliably measured line by line by hand using SPLOT in IRAF. This can also be automated with DAOSPEC∗, a software tool that was optimised to work with GIRAFFE HR spectra. DAOSPEC ﬁts a continuum over an entire spectrum, subtracts it and then measures the EW s for all the lines. It will also ﬁnd the radial velocity of the star by cross-correlating the spectrum with a line list. In both

∗ http://cadcwww.dao.nrc.ca/stetson/daospec/ 4.2: The FLAMES Spectra 53

Table 4.4: Vrad and associated errors for our Fornax targets. BOLD are possible foreground stars, since they fall outside the accepted range of Vrad for membership in Fornax. Struck-out stars have been rejected because their September average velocity was signiﬁcantly diﬀerent (>3 km/s) to their January velocity. This suggests the possibility that they are binary stars and they were therefore rejected from our analysis at this point.

Star Vrad Stddev Star Vrad Stddev Star Vrad Stddev BL038 62.86 0.74 BL147 52.39 0.61 BL221 54.25 0.32 BL045 79.54 0.23 BL148 65.48 0.65 BL227 52.15 0.67 BL052 30.32 1.25 BL149 50.70 0.40 BL228 55.00 0.34 BL061 82.38 1.12 BL150 68.70 1.12 BL229 63.09 0.57 BL065 63.86 0.78 BL151 41.41 0.58 BL231 63.28 1.11 BL069 68.85 2.63 BL153 50.68 0.53 BL233 58.91 0.40 BL070 60.43 2.85 BL155 46.18 0.21 BL239 47.47 0.35 BL076 52.83 0.78 BL156 30.63 0.55 BL242 39.48 0.55 BL077 56.23 1.49 BL158 41.26 0.67 BL247 40.40 0.63 BL079 56.31 0.56 BL160 53.50 0.70 BL249 63.30 0.73 BL081 50.58 0.22 BL163 41.29 0.48 BL250 42.06 0.36 BL084 58.92 0.80 BL165 70.01 0.71 BL251 55.14 0.45 BL085 88.97 1.17 BL166 57.36 0.51 BL253 67.48 0.39 BL091 52.79 0.44 BL168 75.50 0.78 BL254 73.05 0.26 BL092 53.99 0.37 BL171 58.27 1.19 BL257 56.51 0.36 BL094 47.45 0.43 BL173 44.93 1.60 BL258 49.41 0.79 BL096 72.90 0.37 BL180 72.51 0.21 BL260 61.11 0.70 BL097 47.47 0.18 BL181 57.46 1.89 BL261 71.35 1.27 BL100 55.92 0.46 BL183 19.09 0.32 BL262 47.72 0.52 BL104 51.36 0.31 BL185 55.32 0.44 BL266 52.47 1.43 BL107 63.97 0.42 BL189 82.78 0.47 BL267 59.04 0.48 BL109 -15.20 0.42 BL190 57.31 0.24 BL269 45.33 0.30 BL112 53.90 2.36 BL195 39.03 0.46 BL273 63.77 6.18 BL113 59.21 0.31 BL196 79.09 0.25 BL274 41.22 1.65 BL115 70.05 0.30 BL197 46.27 1.02 BL278 52.43 0.56 BL119 56.25 1.95 BL198 45.06 11.29 BL279 67.66 0.73 BL122 65.88 0.75 BL203 56.40 0.92 BL293 56.83 0.99 BL123 73.85 0.71 BL204 59.30 0.81 BL295 42.56 0.37 BL125 58.23 0.77 BL205 59.91 0.27 BL298 62.09 1.43 BL127 46.60 0.67 BL207 48.82 0.42 BL300 69.67 0.41 BL132 40.34 0.48 BL208 54.35 0.53 BL304 71.21 0.26 BL135 45.64 0.57 BL210 51.51 0.32 BL311 50.86 0.23 BL138 56.52 0.33 BL211 50.35 0.20 BL315 53.15 0.85 BL140 46.13 0.69 BL213 71.82 0.45 BL323 52.72 1.15 BL141 76.47 0.35 BL216 66.92 0.26 BL325 18.70 0.80 BL146 50.01 0.61 BL218 63.33 0.46 54 chapter 4: Abundances with the FLAMES multi-fibre instrument

Figure 4.3: Histogram of For- nax radial velocity measurements, Vrad. The central value ± 3σ cut-oﬀ are shown in solid lines (this work) and the dashed lines (Battaglia et al. 2006).

cases, the placement of the continuum is critical because it influences the derived EW s. In the case of the three GIRAFFE setups, a badly determined local continuum level will create an offset between the average value for lines of the same element measured in different setups. Since DAOSPEC is run independently on each of the three setups, the continuum levels in each setup is independently determined, so continuum placement problems will show up as a systematic differences in the abundances deduced in different setups. Figure 4.2 shows where the continuum was placed for one star of our sample.

Figure 4.4, shows a comparison between the equivalent widths measured by SPLOT and DAOSPEC, on a UVES spectrum and on a GIRAFFE spectrum for the same star. The stars selected for the comparison are two of the three stars we have in common with Shetrone et al. (2003). There does not appear to be a correlation of the quality of the match between SPLOT and DAOSPEC with the signal to noise ratio (SNR), shown in the bottom corner of each panel. The SNR have been calculated in the same region of the UVES CCD Lower and Upper (l,u) and on GIRAFFE HR 10, 13 and 14. There is good agreement between the two methods, the EW s from both the UVES and GIRAFFE spectra seems to be scattered around the slope = 1 line. It can be seen that especially for EW . 200 mÅ, there is a good agreement between the two methods, which means that under these conditions, DAOSPEC can be used instead of SPLOT for EW measurements. 4.2: The FLAMES Spectra 55

Figure 4.4: Comparison of the equivalent width measurements (mÅ) for two reference stars, BL239 (top) and BL278 (bottom) in different cases. The Y-axis is the SPLOT measurements while the X-axis is DAOSPEC. The left column EW measurements are from UVES spectra while the right column EW s are from GIRAFFE. We plotted a full line (with slope = 1) for the perfect correlation, a dashed line with an offset of ±6 mÅ from this line and dotted lines as 10% error convolved with the 6 mÅ error. Also printed on the figure are the mean difference between two measurements and the standard deviation. 56 chapter 4: Abundances with the FLAMES multi-fibre instrument

4.2.5 Cleaning up the spectra

Telluric absorption lines

Telluric lines are absorption lines caused by molecules (O2,H2O) in our atmosphere. We cannot predict their strength very well (since our atmosphere is not static) but we know at which wavelength they occur (same rest frame for every exposure). Since no special calibration (e.g. fast rotating hot star) was acquired together with our observations, we could not correct for telluric absorption lines. The only option we have is to flag (our atomic) lines affected by telluric lines and remove them from our abundance analysis. This has to be done star by star, because the position of stellar lines changes with Vrad. This means that lines that are rejected for one star might be used in another star, because of their different radial velocities. Because our GIRAFFE observations were made at two different times, (September 2003 and January 2004), twice as many lines per star can be affected by tellurics and therefore rejected.

Leaks from Calibration Lamps

There was a problem with the calibration lamps in setup HR14 during our January 2004 run. This was not a problem in our September observations, but after this the grating of FLAMES was upgraded to a higher efficiency. Saturated light from the calibration lamps leaked on our stellar spectra, as seen in Figure 4.5. This extra light artificially increased the flux of the adjacent stellar spectra. Only a small fraction of our sample were affected, and for these stars, all lines in the affected regions were simply removed from our abundance analysis.

Figure 4.5: Portion of a raw CCD image (HR14, January 2004) displaying the vertical spectra aligned side by side. The two brightest ones are the saturated calibration lamps (thorium lines), leaking onto their neighbouring spectra. This portion of spectrum is about 50Å long, starting close to the Hα absorption line at the bottom of the ﬁgure. 4.3: Selecting our stellar parameters 57 4.3 Selecting our stellar parameters

Once we have accurately measured the radial velocities of our stars and the EW s for all our absorption lines, the data reduction is ﬁnished and we can proceed with the analysis. In this section, we will summarise how we assign stellar parameters to our results.

4.3.1 Photometric gravity When setting stellar parameters to an observed star, Fe ii lines can be used to constrain the gravity of the star, by ensuring that there is ionisation equilibrium between Fe ii and Fe i. Unfortunately, our Fe ii lines are too uncertain (not enough lines are available and they have a large scatter) to be used for an accurate ionisation balance determination. Thus we used equation 3.9 to determine a photometric estimate and use this gravity as a final value, without adjusting them to balance Fe i and Fe ii. The distance modulus we used in this equation is 20.65, the mass of our stars was set to 1.2 M and the extinction E(B −V ) = 0.03 (Bersier 2000). A standard reddening law (A(V )/E(B −V ) = 3.24) was adopted, and the bolometric corrections were computed for each star using the calibration of Alonso et al. (1999b). The gravity does not significantly affect our abundance: lowering log g by 0.5 dex will lower the Fe i by ≈ 0.1 dex and Fe ii by ≈ 0.3 dex. We thought it was preferable to use the same log g scale for all stars rather than modifying our sample using a sometimes poorly determined [Fe ii/H].

4.3.2 Photometric Teﬀ

As explained in section 3.2.1, we are using photometric colours to estimate Teff , using the calibrations found in Ramírez & Meléndez (2005). Using optical (V , I) data from our photometric survey and infrared (J, H, K) data that were kindly provided to us ahead of publication for 60% of our stars (Gullieuszik et al., in prep.), we calculated four Teff based on four colours, V − I, V − J, V − H and V − K. The equation to calculate 0 the temperature coefficient θeff is the following (taken directly from Ramírez & Meléndez 2005):

0 2 2 θeff = a0 + a1X + a2X + a3X[Fe/H] + a4[Fe/H] + a5[Fe/H] (4.1) where X represents the photometric colours, V − I, V − J, V − H, V − K, ai are the coefficients of the fit. The coefficients of the equations are given in Table 4.A1. This relates to Teff in the following way: 5040 Teff = 0 + P (X, [Fe/H]) (4.2) θeff the correction polynomial P , which is colour and metallicity dependent, is given as

2 3 P = P0 + P1X + P2X + P3X (4.3) where the coefficients (P0,P1,P2,P3) are given in Table 4.A2. Our three comparison stars, for which we have UVES spectra, were used to make sure that our temperature is on the correct scale (abundance is independent of χex). For these three stars, the spectroscopic Teff were in perfect agreement with the Teff (V − I). There was small offset for the temperature derived with the IR colours (Teff (V − {J, H, K})). We shifted them 58 chapter 4: Abundances with the FLAMES multi-fibre instrument by a constant to make them agree with the spectroscopic Teff . Doing this is the equivalent of zero-pointing our photometry, to be in agreement with the spectroscopy. This is represented graphically in Figure 4.6. The gray circles are for those stars for which we have V , I, J, H and K photometry. The black diamonds are those for which we only have V and I, where in this case we estimated their corresponding IR colours based on the linear regression with their V − I colour, as indicated by the solid line. From the right hand side of Figure 4.6, we see that most estimates of Teff for the same star are the same with a precision of ± 50 K.

The four Teff (V − X) as a function of (V − X) are shown in Figures 4.7, 4.8, 4.9, and 4.10. In these figures are also plotted (large symbols) the model predicted (V − X) of stellar atmosphere models at different temperature with different metallicity and gravity. These four plots clearly illustrate the relation of colour versus temperature. The predicted colours from the models, at the metallicity and gravity range of our sample sit well on the observed points, a nice self-consistency check. The models that we use to deduce our abundances also produce the right colours for our stars. The final Teff we used is the average of the four Teff , and they are presented in Table 4.A3. 4.3: Selecting our stellar parameters 59

Figure 4.6: Teff (V − I) as a function of Teff (V − {J, H, K}) and the difference between the two. The Teff (V − {J, H, K}) have been shifted onto the Teff (V − I) scale. 60 chapter 4: Abundances with the FLAMES multi-fibre instrument

Figure 4.7: Teﬀ (V − I) as a function of colour (V −I). Also plotted are the predicted colours from models of diﬀerent metallicity and gravity.

Figure 4.8: Teﬀ (V − J) as a function of colour (V −J). Also plotted are the predicted colours from models of diﬀerent metallicity and gravity. 4.3: Selecting our stellar parameters 61

Figure 4.9: Teﬀ (V −H) as a function of colour (V −H). Also plotted are the predicted colours from models of diﬀerent metallicity and gravity.

Figure 4.10: Teﬀ (V − K) as a function of colour (V − K). Also plotted are the predicted colours from models of diﬀerent metallicity and gravity. 62 chapter 4: Abundances with the FLAMES multi-fibre instrument

4.3.3 Iterating on the parameters Stellar parameters were chosen for our stars in an iterative process. We first took the photometric parameters, log g and Teff , as explained in sub-sections 4.3.1 and 4.3.2. As a first guess, we set [Fe i/H] and vt to be -1.0 dex and 2.1 km/s (typical values for our sample) for all stars. We modified these starting values until we obtained good fit for each parameter of the model. The vt will be correct when the slope measured in [Fe i/H] vs EW (slopeW ) becomes zero, as shown in the bottom panel of Figure 3.5. A fast way to converge to the solution, is to take advantage of the linearity (and symmetry) of slopeW on vt, (as will be demonstrated later in section 4.3.4, Figure 4.14) The simple correction based on the measured slope is the following:

vt = vt + slopeW /0.0055 (4.4)

At the same time as we converge on vt, we also adjust the metallicity of the model used to reflect the average value of all the Fe i lines of the star. We did that for all the stars and our results are summarised in Figure 4.11. At this point, we still have our photometric first guesses for Teff and log g. This figure shows the quality of our model fitting to all the stars, where the top panel displays the values of the slope in the [Fe/H] vs χex plane for each star, and in the [Fe/H] vs EW plane in the bottom panel. In each panel, there are two plots, where the the X-axis is always the absolute value of the slopes, the parameter we want to minimise. The top part of each panel is a simple histogram, while the bottom one is a scatter plot of the (absolute) value of the slope against its error. This is used to compare the quality of the fit from star to star, easily identifying outliers. These slopes (definition and relevance) are first introduced in chapter 3 (Fig- ure 3.5). The two shaded areas represent regions where: slope > σ and slope > 2σ . We have drawn two (solid) lines for the (empirically determined) maximum acceptable slope and error in order to still consider a good model fit. The stars outside these limits, those with large slopes and/or large scatter (σ) were not used in our analysis (chapter 6).

Once we were satisfied with vt and [Fe/H], we iterated on Teff , trying to minimise the slopeχ. Unlike vt, our Teff estimates are based on reliable photometry. We have seen from Figure 4.6 that our precision on Teff is of the order of ± 50 K. This limits the amount by which we want to depart from the photometric value. Figure 4.12 show the difference in slopeχ when we change the temperature by 100 K (twice our photometric precision) for the stars with large slopeχ. This allowed us to get some slopeχ closer to zero, therefore having a stellar model that is a better representation of the observed star. We stopped iterating at this point and accepted these parameters for our models, rejecting stars with a slopeχ > 0.06 and associated σ(slopeχ) > 0.05. For the vt, we rejected stars with a slopeW > 0.00035 and associated σ(slopeW ) > 0.00090. 4.3: Selecting our stellar parameters 63

Figure 4.11: Status of the slopes in χex (top) and in EW (bottom) after many iterations on vt and metallicity. The shaded areas correspond to slope > {1 and 2} σ. This is a graphical way of summarising our attempt to minimise the two slopes. The solid lines on the scatter plots are the maximum tolerated values in the slope and error. 64 chapter 4: Abundances with the FLAMES multi-fibre instrument

Figure 4.12: Status of the slopes in χex (top) and in EW (bottom) after we allowed the Teﬀ to change from the photometric value by a max of 100 K (our photometric precision). 4.3: Selecting our stellar parameters 65

4.3.4 Precision and error estimates DAOSPEC gives an error estimate for each EW it measures, δEW which is used to calculate the difference in abundance between EW and EW ± δEW . The largest error of the two is then adopted as our DAOSPEC abundance error, δDAO. The final adopted measurement error for each element was calculated in the following way. In order to avoid really small error estimates due to low number statistics, we use the dispersion of Fe i as a lower limit of the dispersion for an element. For the final error on each of our [X/H] ratio, we have adopted the maximum of these three values: ! σ(Fe I) σ(X) δ([X/H]) = MAX δDAO, p , p (4.5) (NX) (NX) and for the [X/Fe] ratios, we just take the quadratic sum of [X/H] and [Fe/H]. This con- servative estimate includes all sources of errors due to the measurement. There is also the uncertainty caused by our choice of stellar parameters, which is linked to the method, not the measurements and for this reason, is not included in our error bars but presented separately in Table 4.5. The dependencies on model atmosphere parameters is different for each element and each ionisation state. In the case of Fe, the only element used to constrain our stellar parameters, it is also used for weights when trying to determine the slopes (Figure 3.5).

To determine the precision of our abundance results, we took one star (BL239) and modified the model parameters many steps away from our adopted fit to check how this affected the derived Fe abundances. This illustrates the precision of the stellar parameters, and by how much they can change and still be consistent. In Figure 4.13, we plotted the slopeχ as a function of the change in Teff (by steps of ± 100 K), along with the corresponding change in [Fe i/H] (label above the points). It is clear from this figure that the change (in abundance) is not symmetrical, as cooling the star by 100 K, 200 K and 300 K does not significantly change [Fe i/H] while warming it up by steps of 100 K increases the abundance by at least 0.05 dex at each step. The values below the points corresponds to the the slope/σ, which is an indication of how far we are from a good parameter fit. The linear regression and the estimation of its error was made using the routine described in Fasano & Vio (1988).

From looking at Figure 4.13 and at the slopeχ numeric values, it could be argued that Teff + 100 K would be a better choice than our adopted value (4123 K). Indeed, warming up all of our Teff by 100 K could make all our slopes closer to zero but the resultant Teff would not be in agreement with our photometry and the spectroscopically determined Teff from the 3 Shetrone stars. This effect is probably caused by the combination of our resolution and/or choice of lines and it means that the precision of our spectroscopic Teff is not as good as our photometric one. An other way to check our Teff scale is to look at the two curves of growth (CoG) of Figure 4.15. The top plot is for the model at Teff = 4123 K and the bottom plot for a model at 4523 K. Although not immediately obvious, the hotter model has the tendency to separate the lines of different χex, with squares (, low χex) being artificially higher (perpendicular to the curve of growth) than the pluses (+, high χex), compared to the cold CoG. A representative curve of growth should not show such separation between different χex. 66 chapter 4: Abundances with the FLAMES multi-fibre instrument

Figure 4.13: Change of Teff for star BL239, where ∆Teff = 0 corresponds to Teff = 4123 K and each step is 100 K. On the Y-axis is the slope in abundance of Fe i vs χex, where slope = 0 is a good model fit for this parameter. On top of each point is the resultant Fe i abundance (differential) derived with these modified stellar parameters, and on the bottom is the slope/σ value, which is an indication of how far we are from a good parameter fit.

Figure 4.14: Change of vt for star BL239, where ∆vt = 0 corresponds to vt = 2.1 km/s, and each step is 0.1 km/s. On the Y-axis is the slope in abundance of Fe i vs EW , where slope = 0 is a good model fit for this parameter. On top of each point is the resultant Fe i abundance (differential) derived with these modified stellar parameters, and on the bottom is the slope/σ value, which is an indication of how far we are from a good parameter fit. 4.3: Selecting our stellar parameters 67

Figure 4.15: Two curves of growth (CoG) for the same star, BL239. The upper example has the optimum stellar parameters while the lower example had its Teff increased by +400 K, showing the different placement of the lines according to their χex on the curve. On the hotter CoG, lines with squares (, low χex) are artificially slightly higher (perpendicular to the curve of growth) than the pluses (+, high χex), compared to the cold CoG. This is an indication that the hotter CoG is not a perfect match for the data. 68 chapter 4: Abundances with the FLAMES multi-fibre instrument

Table 4.5: Errors due to stellar parameters uncertainties.

Element ∆ Teﬀ = +200 K ∆ log g = -0.5 dex ∆ vt = +0.2 km/s Combined [Ba ii/H] -0.03 0.15 0.26 0.30 [Ca i/H] -0.23 0.04 0.07 0.24 [Cr i/H] -0.32 0.11 0.09 0.35 [Eu ii/H] 0.03 0.22 0.04 0.22 [Fe i/H] -0.04 0.12 0.08 0.15 [Fe ii/H] 0.32 0.28 0.04 0.43 [La ii/H] -0.05 0.23 0.04 0.23 [Mg i/H] -0.06 0.00 0.06 0.08 [Na i/H] -0.18 0.01 0.01 0.18 [Nd ii/H] 0.00 0.20 0.02 0.20 [Ni i/H] -0.03 0.14 0.05 0.15 [Si i/H] 0.14 0.12 0.02 0.19 [Ti i/H] -0.33 0.08 0.04 0.34 [Ti ii/H] 0.10 0.21 0.04 0.24 [Y ii/H] 0.06 0.19 0.02 0.20

We made the same plot for vt, illustrated in Figure 4.14. Unlike Teﬀ , the changes in vt are symmetrical, making it easy to ﬁnd the right vt to represent our star. ∆vt = 0 corresponds to vt = 2.1 km/s and from this plot, looking only at the value of the slopes, vt = 2.2 would also be acceptable. Both of them have slopeW σ(slopeW ), showing the limit of our precision, which should be of the order of 0.2 km/s.

4.4 Systematics and corrections

In this section we discuss the systematic effects due to the method of analysis and the effect of hyperfine structure on the Eu line at 6645.1 .

4.4.1 Systematics To measure systematics caused by different technical choices, we have used five different methods and compared them over a range of elements. We are interested in measuring the effect of changing the stellar models, the line list, the resolution and the way of measuring the EW s. We analysed the 3 Shetrone stars in five different ways, always re-determining the stellar parameters to obtain a good fit:

• The original 2003 abundance analysis, using UVES spectra, the Shetrone line list, splot to measure the EW s and extrapolated (in log g and Teﬀ ), plane-parallel models (Gustafsson et al. 1975).

• Same as above but using spherical MARCS 2005 models instead of the plane-parallel models of Gustafsson et al. (1975). 4.4: Systematics and corrections 69

Figure 4.16: Abundance differences between the 3 Shetrone stars, using different methods of analysis on the UVES spectra, (open symbols) compared to the GIRAFFE analysis (filled dots). We plotted the original 2003 results (inverted triangles) our re-analysis with the 2005 models, keeping the old line list + splot (triangles), new line list + splot (diamonds) and new line list + daospec (circle). 70 chapter 4: Abundances with the FLAMES multi-fibre instrument

Figure 4.17: Curve of growth for BL239 (Fnx- M25), showing a lack of weak lines, leading to an inaccurate abundance.

• Same as above but using our new line list, better suited for the more metal rich stars in Fornax, restricting the analysis to the same wavelength coverage as FLAMES HR 10, 13 and 14.

• Same as above but measuring the EW s automatically with DAOSPEC instead of manually with splot. • Same as above but instead of using the same UVES spectra like the four previous cases, we used our new, independent GIRAFFE spectra, eﬀectively reducing the resolution by a factor two (from R ' 40 000 to R ' 20 000).

We compare the abundance determined in each these cases for 14 different elements or ionisation states in Figure 4.16. The most striking feature in this plot is the huge difference between different analysis (0.4 dex) in [Fe/H] for star BL239, the one with the [Fe/H] varying from -0.9 to -1.3. The reason why the GIRAFFE analysis differs so much from the original UVES analysis (Shetrone et al. 2003) is due to the fact that they lack weak lines, as can be seen in Figure 4.17, where the lack of lines on the left part of the CoG is evident and can lead to bias stellar parameters and therefore abundances. In this particular case, the lack of weak lines drove the microturbulence to lower values, biassing the metallicity towards lower values. The other two stars do not show a similar problem because the most metal-poor star had enough weak lines, and Shetrone et al. themselves added extra weak lines to the analysis of the metal-rich one.

Old models versus new models There are several differences between the old (Gustafsson et al. 1975) and new (Gustafs- son et al. 2003) family of models. The two most important are the different geometry used (plane-parallel versus spherical) and the physics involved determining the opacities. Abundances calculated with these two methods are similar, most of the time within the error bars. Heiter & Eriksson (2006) studied extensively the effect of using different geometry in determining the abundance. We cannot directly compare to their result because as we summarise in our section 3.3.2 they compare both p2005_p and s2005_p to the fully consistent s2005_s while we are trying to compare plane-parallel and spherical 4.4: Systematics and corrections 71 models with different physics, both of them treated in the plane-parallel approximation for the line formation code (p1975_p to s2005_p). According to their conclusion, if the “ideal” case is the fully consistent case s_s, using mixed geometry (s_p) is preferable (closer to s_s) than the full p_p case. Therefore, the effect of geometry alone is mea- surable and is partially responsible for the difference in abundance between the triangles and inverted triangles symbols of Figure 4.16.

Choice of Lines to include in the analysis To compare the effects of different line lists (see section 3.4), compare the triangles and the diamonds in Figure 4.16. This is usually the biggest difference between two methods, showing the importance of having a common line list for comparing abundance results. Generally, there is a difference of 0.3 dex for [Fe i/H] in BL239 and a smaller difference (0.15 dex) for the other two stars. But as long as there are enough lines in a given line list, the derived abundance is usually reliable. Only when the number of lines goes down there can be a problem with the abundance, where an outlier can really affect the derived value. But this has to be checked star by star, as it can be metallicity and/or signal to noise dependent.

DAOSPEC vs SPLOT for measuring EW s The method used to measure the EW s also aﬀects the abundance results but only by a small amount, in most cases well within the error estimate. This can be seen by looking at the circles and the diamonds in the Figure 4.16. This check reinforces our conﬁdence using the automatic DAOSPEC measurements allowing us to go from hand-measurement to a much faster automated processing.

UVES vs GIRAFFE, the eﬀect of resolution and wavelength coverage Comparing UVES and GIRAFFE results shows that even with a loss of a factor two in resolution, it is possible to determine accurate abundances. By comparing the empty circles to the solid dots in Figure 4.16, it is clear that the results are identical, within the errors.

4.4.2 Hyperfine splitting correction The hyperfine structure is a small perturbation in the energy levels of an atoms due to the interaction of the nuclear magnetic dipole with the magnetic field of the electron. The electron moving around the nucleus has a magnetic dipole moment, because it is charged. Its interaction with the magnetic moment of the nucleus (due to its spin) leads to hyperfine splitting of the energy level. This phenomenon is energy level dependent, therefore different for each line. We can correct for hyperfine splitting with spectral synthesis, as presented for the Eu abundance in chapter 5 and illustrated in Figure 5.4. For our FLAMES data, we only corrected for the HFS in our Eu line at λ = 6645.1 Å. Here is a summary of the method we used.

A line that has hyperﬁne structure consist of many small lines close to each other instead a line arising from a single energy level. These lines can be used to create a 72 chapter 4: Abundances with the FLAMES multi-fibre instrument

Figure 4.18: HFS correction (dex) for the Eu line at 6645.1Å, tested on a plane-parallel model of Teﬀ = 4300, log g = 0.6, [Fe/H] = -1.5 and vt = 1.7. synthetic spectrum, where all these split lines will be superposed together. All the stellar parameters are known except the abundance of Eu, which we can modify until it represents the observed spectra. Ignoring HFS tends to overestimate the abundance so correcting for HFS will typically lower the “direct EW ” measured abundance.

Making synthetic spectra and taking into account the HFS information to ﬁnd the abundance of a line is a good method but it is too time consuming for large sample of stars or for elements with many lines. To measure the Eu abundance in the many GIRAFFE stars, we rather designed an HFS correction to be applied to the abundance deduced from the EW treated as a single line. We found out that this correction, in our stellar parameter and abundances range, depends dominantly on the strength (EW ) of the line, and very little on other parameters. The validity of this correction was tested in the following ranges:

• -2.0 < [Fe/H] < -1.0

• 1.7 < vt < 2.5 • 0.3 < log g < 1.2

• 4000 < Teﬀ < 4800 The HFS correction becomes EW -dependent only and is applied after the the abundance has been derived from the non-corrected line. In Figure 4.18 we present the correction∗ for the Eu line at λ = 6645.1 Å.

Appendix 4.A Large tables

∗ Thanks to Kim Venn for calculating the correction in this range of parameters 4.A: Large tables 73

Table 4.A1: Coeﬃcients of the giant star colour calibration

Colour a0 a1 a2 a3 a4 a5 (V − I) 0.3575 +0.9069 -0.2025 +0.0395 -0.0551 -0.0061 (V − J) 0.2943 +0.5604 -0.0677 +0.0179 -0.0532 -0.0088 (V − H) 0.4354 +0.3405 -0.0263 -0.0012 -0.0049 -0.0027 (V − K) 0.4405 +0.3272 -0.0252 -0.0016 -0.0053 -0.0040

Table 4.A2: Metallicity dependence of the Polynomial correction.

Colour P0 P1 P2 P3 Validity (V − I) +0.42933 ...... −0.5 ≤ [Fe/H] ≤ 0.5 (V − J) -122.595 +76.4847 ...... 00 (V − H) -377.022 +334.733 -69.8093 ... 00 (V − K) -72.6664 +36.5361 ...... 00 (V − I) -0.14180 ...... −1.5 ≤ [Fe/H] ≤ −0.5 (V − J) -10.3848 ...... 00 (V − H) +71.7949 -55.5383 +9.61821 ... 00 (V − K) +86.0358 -65.4928 +10.8901 ... 00 (V − I) +9.31011 ...... −2.5 ≤ [Fe/H] ≤ −1.5 (V − J) +4.18695 +13.8937 ...... 00 (V − H) -27.4190 +20.7082 ...... 00 (V − K) -6.96153 +14.3298 ...... 00 (V − I) -23.0514 ...... −4.0 ≤ [Fe/H] ≤ −2.5 (V − J) -67.7716 +28.9202 ...... 00 (V − H) -46.2946 +20.1061 ...... 00 (V − K) -943.925 +1497.64 -795.867 +138.965 00

Table 4.A3: Coordinates and photometric temperatures in different colours, shifted on the Teff (V − I) scale. Teff is the average of the four different TV −X .

Star RA (J2000) DEC (J2000) TV −I TV −J TV −H TV −K Teﬀ log g BCV MBol BL038 02 40 20.45 -34 24 00.1 3984 3976 3996 3964 3980 0.69 -0.93 -3.32 BL045 02 40 07.52 -34 23 31.8 4114 4134 4128 4110 4122 0.85 -0.77 -3.07 BL052 02 40 10.42 -34 25 17.6 4009 3999 3995 3984 3997 0.72 -0.91 -3.27 BL061 02 39 28.59 -34 18 38.0 4316 4341 4335 4333 4331 0.84 -0.60 -3.30 BL065 02 39 22.15 -34 19 40.3 4315 4340 4334 4332 4330 0.97 -0.59 -2.99 BL069 02 39 29.56 -34 25 10.6 4143 3999 4033 3994 4042 0.72 -0.86 -3.29 BL070 02 39 40.46 -34 19 38.8 3939 3928 3931 3935 3933 0.68 -1.00 -3.29 BL076 02 39 31.49 -34 23 05.1 4088 4061 4063 4048 4065 0.83 -0.83 -3.05 BL077 02 39 51.42 -34 21 20.9 4027 4024 4025 4028 4026 0.80 -0.88 -3.08 BL079 02 39 19.60 -34 24 49.3 3986 3903 3922 3934 3936 0.76 -1.01 -3.08 BL081 02 39 56.01 -34 24 10.5 4102 4045 4060 4042 4062 0.82 -0.84 -3.07 BL084 02 38 42.96 -34 25 49.1 3972 3964 3966 3970 3968 0.72 -0.95 -3.23 BL085 02 38 55.53 -34 25 36.3 4277 4299 4294 4293 4291 0.87 -0.63 -3.19 Continued on next page 74 chapter 4: Abundances with the FLAMES multi-fibre instrument

Star RA (J2000) DEC (J2000) TV −I TV −J TV −H TV −K Teﬀ log g BCV MBol BL091 02 39 04.31 -34 25 18.8 4155 4165 4163 4164 4162 0.86 -0.73 -3.09 BL092 02 38 49.28 -34 24 04.9 3965 3956 3959 3963 3961 0.74 -0.96 -3.16 BL093 02 38 39.87 -34 25 57.1 4068 4070 4070 4072 4070 0.90 -0.82 -2.87 BL094 02 38 53.89 -34 25 06.7 3990 3984 3985 3989 3987 0.72 -0.92 -3.24 BL096 02 39 14.33 -34 22 41.5 4012 4008 4009 4013 4010 0.75 -0.90 -3.19 BL097 02 39 04.07 -34 23 52.7 4058 4059 4059 4062 4060 0.82 -0.84 -3.06 BL100 02 38 56.00 -34 24 44.8 4043 4042 4043 4046 4044 0.84 -0.86 -3.01 BL104 02 39 14.59 -34 23 21.0 4014 4011 4012 4015 4013 0.77 -0.89 -3.15 BL107 02 38 54.37 -34 31 23.5 4366 4396 4389 4387 4384 0.83 -0.56 -3.37 BL109 02 39 04.08 -34 37 58.7 4334 4360 4354 4352 4350 0.93 -0.58 -3.09 BL112 02 38 45.68 -34 34 47.5 4096 4100 4099 4101 4099 0.76 -0.79 -3.27 BL113 02 39 08.16 -34 36 53.3 4179 4191 4188 4189 4187 0.83 -0.72 -3.19 BL115 02 38 43.45 -34 32 05.3 4112 4117 4116 4118 4116 0.79 -0.77 -3.22 BL119 02 38 42.17 -34 29 50.8 3964 3956 3959 3963 3960 0.73 -0.96 -3.19 BL122 02 39 02.91 -34 31 12.0 4051 4051 4052 4054 4052 0.76 -0.84 -3.22 BL123 02 38 53.75 -34 30 06.6 3995 3990 3991 3995 3993 0.71 -0.92 -3.27 BL125 02 39 08.50 -34 30 55.4 4078 4080 4080 4082 4080 0.79 -0.82 -3.17 BL127 02 39 04.00 -34 37 26.1 3963 3955 3957 3961 3959 0.71 -0.96 -3.25 BL132 02 38 53.62 -34 33 04.5 3916 3903 3906 3911 3909 0.65 -1.04 -3.33 BL135 02 39 01.55 -34 36 48.7 4057 4058 4058 4061 4058 0.83 -0.84 -3.06 BL138 02 39 16.20 -34 37 00.2 3989 3906 3924 3937 3939 0.71 -0.99 -3.21 BL140 02 39 11.56 -34 30 44.7 3995 3989 3991 3994 3992 0.75 -0.92 -3.17 BL141 02 39 10.99 -34 28 34.4 4076 4078 4078 4080 4078 0.84 -0.82 -3.04 BL146 02 38 41.76 -34 28 58.4 4074 4076 4076 4078 4076 0.84 -0.82 -3.04 BL147 02 38 44.02 -34 30 51.8 4186 4199 4196 4197 4194 0.94 -0.70 -2.92 BL148 02 39 11.05 -34 39 08.6 3929 3917 3920 3925 3923 0.72 -1.03 -3.16 BL149 02 38 57.19 -34 35 39.8 4097 4101 4100 4102 4100 0.88 -0.80 -2.97 BL150 02 39 13.94 -34 28 36.1 4026 4023 4024 4027 4025 0.80 -0.88 -3.08 BL151 02 39 00.29 -34 30 30.0 4025 4022 4023 4026 4024 0.81 -0.88 -3.06 BL153 02 39 08.35 -34 32 44.8 4035 4033 4034 4037 4035 0.82 -0.86 -3.04 BL155 02 38 40.44 -34 35 25.0 4059 4060 4060 4063 4060 0.90 -0.84 -2.87 BL156 02 39 05.62 -34 26 30.6 4096 4100 4099 4101 4099 0.89 -0.79 -2.95 BL157 02 39 02.44 -34 27 35.4 4073 4075 4075 4077 4075 0.88 -0.82 -2.93 BL158 02 39 06.84 -34 35 45.7 4076 4079 4078 4081 4078 0.87 -0.82 -2.96 BL160 02 39 01.74 -34 27 19.2 4027 4025 4026 4029 4027 0.84 -0.88 -3.00 BL163 02 38 43.54 -34 32 25.4 4025 4022 4023 4026 4024 0.88 -0.88 -2.89 BL165 02 39 13.37 -34 40 15.8 3967 3959 3961 3965 3963 0.77 -0.95 -3.08 BL166 02 39 16.01 -34 30 12.9 4070 4091 4074 4108 4086 0.84 -0.81 -3.06 BL168 02 39 14.78 -34 27 29.8 4014 4011 4012 4015 4013 0.83 -0.89 -3.00 BL171 02 39 15.48 -34 30 46.5 4047 4047 4047 4050 4048 0.87 -0.85 -2.94 BL173 02 39 14.36 -34 34 42.3 3991 3985 3987 3990 3988 0.85 -0.93 -2.92 BL180 02 39 33.38 -34 33 58.4 4115 4102 4107 4134 4114 0.77 -0.78 -3.25 BL181 02 39 16.76 -34 34 49.9 4016 3962 3969 3969 3979 0.80 -0.93 -3.04 BL183 02 39 47.94 -34 26 43.4 4073 4199 4223 4230 4181 0.83 -0.72 -3.16 BL185 02 39 27.67 -34 37 48.5 4026 3980 3999 3988 3998 0.69 -0.92 -3.32 BL189 02 39 38.24 -34 31 20.3 4167 4222 4272 4274 4234 0.91 -0.67 -3.03 BL190 02 39 17.80 -34 30 56.8 4004 3954 3967 3990 3979 0.82 -0.94 -2.99 BL195 02 39 20.21 -34 31 57.0 4153 4153 4147 4191 4161 0.91 -0.73 -2.95 BL196 02 39 34.07 -34 33 33.2 4011 4010 4010 4029 4015 0.76 -0.89 -3.17 BL197 02 39 29.33 -34 26 35.5 3958 3953 3963 3948 3956 0.68 -0.97 -3.30 Continued on next page 4.A: Large tables 75

Star RA (J2000) DEC (J2000) TV −I TV −J TV −H TV −K Teﬀ log g BCV MBol BL198 02 39 19.83 -34 26 27.1 4077 3988 4000 3992 4014 0.78 -0.89 -3.11 BL203 02 39 50.64 -34 26 47.0 4043 4035 4046 4025 4037 0.79 -0.87 -3.13 BL204 02 39 16.23 -34 32 29.9 4076 4184 4161 4134 4139 0.97 -0.75 -2.78 BL205 02 39 38.06 -34 37 06.2 4229 4237 4263 4244 4243 0.96 -0.67 -2.92 BL207 02 39 26.29 -34 29 03.4 3988 3979 3977 3992 3984 0.74 -0.93 -3.19 BL208 02 39 32.94 -34 32 06.2 4140 4166 4161 4170 4159 0.89 -0.74 -3.00 BL210 02 39 47.65 -34 27 05.4 4057 4064 4072 4057 4062 0.81 -0.84 -3.10 BL211 02 39 51.18 -34 29 58.6 3940 3961 3964 3966 3958 0.69 -0.97 -3.28 BL213 02 39 50.18 -34 35 59.3 4039 4045 4017 4025 4032 0.78 -0.87 -3.14 BL216 02 39 41.92 -34 30 35.9 3970 4008 3995 4019 3998 0.75 -0.92 -3.18 BL218 02 39 50.77 -34 28 36.5 3920 3943 3941 3953 3939 0.67 -1.00 -3.31 BL221 02 39 31.85 -34 29 19.9 4051 4051 4057 4064 4056 0.81 -0.84 -3.08 BL227 02 39 45.25 -34 31 57.8 4020 4064 4043 4055 4046 0.84 -0.85 -3.01 BL228 02 39 53.84 -34 29 56.4 3975 3998 3990 4007 3992 0.71 -0.92 -3.28 BL229 02 39 24.80 -34 34 38.1 4022 4008 4012 4013 4014 0.80 -0.90 -3.07 BL231 02 39 18.96 -34 26 43.9 4087 4033 4040 4052 4053 0.86 -0.84 -2.96 BL233 02 39 53.58 -34 37 50.1 4064 4028 4039 4063 4048 0.83 -0.86 -3.03 BL239 02 39 47.09 -34 31 49.8 4083 4145 4127 4137 4123 0.89 -0.77 -2.96 BL242 02 39 28.03 -34 34 01.2 4064 4058 4057 4074 4063 0.85 -0.83 -3.00 BL247 02 39 43.07 -34 40 18.4 4050 4031 4018 4027 4032 0.85 -0.87 -2.98 BL249 02 39 54.24 -34 35 11.2 4032 4034 4009 4019 4024 0.82 -0.88 -3.03 BL250 02 39 45.03 -34 40 11.1 3866 3829 3849 3833 3844 0.65 -1.16 -3.27 BL251 02 39 30.81 -34 35 45.1 3998 3970 3977 3985 3982 0.81 -0.93 -3.00 BL253 02 39 34.08 -34 33 09.6 3989 3995 4006 4022 4003 0.82 -0.91 -3.02 BL254 02 39 41.78 -34 34 16.4 4025 4013 3990 4017 4011 0.83 -0.89 -3.01 BL257 02 39 57.30 -34 31 20.8 3981 3984 3994 4015 3994 0.78 -0.93 -3.10 BL258 02 39 49.68 -34 28 50.4 4008 4045 4026 4039 4030 0.85 -0.88 -2.96 BL260 02 39 55.38 -34 29 54.6 3985 4026 4016 4010 4009 0.79 -0.90 -3.09 BL261 02 39 57.80 -34 26 48.8 4028 4053 4066 4039 4046 0.82 -0.86 -3.06 BL262 02 39 38.42 -34 26 10.3 4054 4021 4027 4018 4030 0.84 -0.88 -2.98 BL266 02 40 10.00 -34 29 58.8 4157 4241 4225 4223 4212 0.83 -0.68 -3.20 BL267 02 40 17.50 -34 26 06.1 4197 4210 4211 4186 4201 0.80 -0.70 -3.27 BL269 02 39 58.20 -34 32 05.3 3965 3992 4008 3997 3990 0.75 -0.92 -3.18 BL273 02 40 09.37 -34 36 17.1 4123 4158 4124 4149 4138 0.81 -0.76 -3.18 BL274 02 40 06.14 -34 28 52.0 4011 4041 4032 4019 4026 0.72 -0.87 -3.29 BL278 02 40 04.38 -34 27 11.3 3964 3967 3977 3980 3972 0.64 -0.95 -3.43 BL279 02 40 02.70 -34 38 29.9 4228 4298 4274 4286 4272 0.95 -0.64 -2.97 BL293 02 40 01.77 -34 27 47.9 4030 4031 4024 4026 4028 0.72 -0.87 -3.29 BL295 02 40 26.72 -34 26 56.8 3994 3976 3975 3975 3980 0.70 -0.94 -3.28 BL298 02 40 13.49 -34 30 02.0 4118 4145 4129 4125 4129 0.85 -0.76 -3.07 BL300 02 40 17.90 -34 27 00.7 3989 3991 4001 3977 3990 0.71 -0.92 -3.27 BL304 02 40 05.49 -34 32 42.7 3959 3936 3941 3965 3950 0.70 -0.97 -3.26 BL311 02 40 22.64 -34 31 31.0 4032 4022 4022 4032 4027 0.79 -0.88 -3.12 BL315 02 40 24.22 -34 26 20.0 4173 4135 4115 4133 4139 0.86 -0.76 -3.06 BL323 02 40 16.76 -34 29 34.4 3897 3859 3881 3887 3881 0.66 -1.08 -3.28 BL325 02 40 27.00 -34 26 44.1 4082 4064 4036 4050 4058 0.81 -0.84 -3.09

Chapter 5 High resolution spectroscopy in Fornax Globular Clusters

Published as A&A 2006 453 547 VLT/UVES spectroscopy of individual stars in three globular clusters in the Fornax dwarf spheroidal galaxy∗ B. Letarte, V. Hill, P. Jablonka, E. Tolstoy, P. François, G. Meylan ABSTRACT– We present a high resolution (R∼ 43 000) abundance analysis of a total of nine stars in three of the five globular clusters associated with the nearby Fornax dwarf spheroidal galaxy. These three clusters (1, 2 and 3) trace the oldest, most metal-poor stellar populations in Fornax. We determine abundances of O, Mg, Ca, Ti, Cr, Mn, Fe, Ni, Zn, Y, Ba, Nd and Eu in most of these stars, and for some stars also Mn and La. We demonstrate that classical indirect methods (isochrone fitting and integrated spectra) of metallicity determination lead to values of [Fe/H] which are 0.3 to 0.5 dex too high, and that this is primarily due to the underlying reference calibration typically used by these studies. We show that Cluster 1, with [Fe /H]=−2.5, now holds the record for the lowest metallicity globular cluster. We also measure an over-abundance of Eu in Cluster 3 stars that has only been previously detected in a subgroup of stars in M15. We find that the Fornax globular cluster properties are a global match to what is found in their Galactic counterparts; including deep mixing abundance patterns in two stars. We conclude that at the epoch of formation of globular clusters both the Milky Way and the Fornax dwarf spheroidal galaxy shared the same initial conditions, presumably pre-enriched by the same processes, with identical nucleosynthesis patterns.

∗ Based on UVES observations collected at the European Southern Observatory, proposal number 70.B-0775 78 chapter 5: HR spectroscopy in Fornax Globular Clusters 5.1 Introduction

t is now established that some dwarf galaxies have globular cluster systems around I them (Lotz et al. 2004, van den Bergh 2006, Seth et al. 2004). Their possible common origin with clusters in larger parent galaxies, the link between the dwarf galaxy ﬁeld and globular cluster stars are open questions to be addressed. The largest samples of dwarf galaxies with globular cluster systems are however distant, and this restricts the analyses to using integrated properties.

Fornax and Sagittarius are the nearest dwarf spheroidal galaxies (dSph) with globular clusters and can be resolved into individual stars. The Fornax dSph contains ﬁve star clusters (Shapley 1938; Hodge 1961) and while the Sagittarius dSph is obscured by dust and confused by merging with our Galaxy, Fornax is high above the Galactic plane, therefore oﬀering a uniquely useful target for investigation, see Figure 5.1.

Figure 5.1: A ≈ 850 × 620 DSS image of the Fornax dSph. North is up and East is to the left, as indicated. We have marked the position of the 5 GCs using the numbering scheme deﬁned by Shapley 1938 and Hodge 1961

The ages of the Fornax globular clusters have been determined by fitting isochrones to deep HST Colour-Magnitude Diagrams [CMDs] (Buonanno et al. 1998, 1999). They are found to be the same age as old metal-poor Galactic clusters M92 and M68 (around 13 Gyr old) to within ± 1 Gyr, with the exception of Cluster 4, which seems buried in the 5.2: Observations 79 center of Fornax and maybe younger by about 3 Gyr. The cluster metallicities have been estimated with different techniques ranging from fitting a slope to the Red Giant Branch (RGB) to high and medium resolution spectroscopy of the integrated light of the cluster. Conclusions vary from one work to another, as summarized in Strader et al. (2003), but the clusters definitely appear more metal-poor than the bulk of the galaxy field stellar population, with bluer RGBs, well populated blue horizontal branches (HB) and a range of HB morphology (Buonanno et al. 1998 and 1999). Saviane et al. (2000) showed that the Fornax dSph field star colour distribution is well fitted by two Gaussian functions, best interpreted as a bi-modal metallicity distribution, with the older population having a wide abundance range between −2.2 and −1.4. Stars as young as 108 Myr have also been discovered in the field of Fornax (Stetson et al. 1998). In this framework, the globular clusters of Fornax dSph trace the first stages of star formation in the galaxy.

High resolution spectroscopy of individual stars in the clusters is the only way to assess the abundances of individual chemical species. Alpha, iron-peak, heavy -elements provide essential clues on (i) the conditions of formation of the globular clusters in a dwarf galaxy, including epoch and time scales (ii) to probe the nucleosynthesis in a galactic system with a star formation history that is fundamentally diﬀerent from that of the Milky Way. We present here a VLT/UVES spectroscopic analysis of a total sample of nine stars in three Fornax dSph clusters.

5.2 Observations

We targeted Cluster 1, Cluster 2 and Cluster 3, to span the Fornax globular cluster system range of distances from the galaxy centre avoiding regions of heavy crowding. We also sampled a range of HB morphology, as well as the metallicity and concentration ranges. Cluster 1, at a radial distance of 43 arcmin (or 1.75 kpc at the distance of Fornax dSph) from the galaxy center, is diﬀuse, with low surface brightness, most of its HB is red. Cluster 2, located at 25 arcmin (1 kpc) from the galaxy center, is slightly more concentrated and exhibits a more extended HB. Finally, Cluster 3 at a galactocentric radial distance of 13 arcmin (530 pc) is very dense and has an extended HB.

We used the red arm of UT2/UVES, CD#3, centered at 580nm, with a wavelength range of 480-680nm (Dekker et al. 2000) in visitor mode in October 2002. We obtained spectra with a resolution of ∼43 000 and average S/N ∼ 20 − 30 per pixel with an integration time of 2 − 6 hours for each of the nine individual stars in Fornax dSph globular Clusters 1, 2 and 3. The stars were selected to be on the RGB from CMDs, (Buonanno et al. 1985; Demers et al. 1990; Jorgensen & Jimenez 1997 and Buonanno et al. 1998). Their individual ﬁnding charts are shown in Figure 5.2. We also observed 5 calibration red giant branch stars in the well studied globular cluster M15 (Sneden et al. 1997). The observations of M15 stars provide an independent check on our data reduction and analysis methods. Details of the observations are shown in Table 5.1, including the derived radial velocities Vrad and S/N ratios. 80 chapter 5: HR spectroscopy in Fornax Globular Clusters Comments rad V from Kustner (1921) from Sandage (1970) e f ID ID Observation Log ...... 3900 ...... carbon star Table 5.1: degrees degrees (s) 670 nm (km/s) from Buonanno et al. (1998) from Jorgensen & Jimenez (1997) c d 39.25495839.254609 -33.17790 -34.17875 14400 14400 23 50 57.6 60.2 same slit as D68 same slit as D56 39.25755439.67738839.684917 -34.1862839.685203 -34.80412 -34.80301 18000 -34.80303 21600 30 14400 30 14400 ... 60.0 30 63.4 ... 64.1 same slit as B74 Too faint 39.68214339.942803 -34.80801 -34.25855 7200 10800 40 30 64.0 59.7 39.957271 -34.25782 21600 30 63.7 39.951057 -34.25277 14400 40 64.8 322.484344322.481920322.509599 12.21002322.543661 12.21231322.457798 12.18986 12.16832 900 1200 12.13489 113 750 115 1000 122 -106.4 900 -111.3 96 -101.4 83 spec. -100.0 double star -100.7 ID ID d c , B713 f c c c f f f f c c , J23 d d b a a a , J9 , J3 , J31 , J65 , J24 a a b b a , K431 , K387 , K825 , K146 , K1040 e e e e , D68 e a , D164 a from Buonanno et al. (1985) from Demers et al. (1990) b a ID Our idCl1-D56Cl1-D68Cl1-D164 Litterature IDs B18 B51 D56 RA (J2000) DEC (J2000) exp. time S/N @ M15-S1M15-S3M15-S4M15-S6M15-S7 S1 ID S3 S4 S6 S7 Cl2-B71Cl2-B74Cl2-B77Cl2-B200Cl2-B226 B71 B74 B77 B200 B226 Cl3-B59Cl3-B61 B59 B61 Cl3-B82 B82 5.3: Data Reduction and Analysis 81

Figure 5.2: The ﬁnding charts for our observations of the Fornax GCs, from 1 (left) to 3 (right). North is up and East is left, as indicated. Note that star Cl3-B59 is outside of the cluster 3 HST ﬁeld, to the west.

5.3 Data Reduction and Analysis

The spectra were extracted with the standard UVES pipeline, except for two pairs of stars on the same slit which we had to reduce interactively using the UVES context within MIDAS (see Table 5.1). At the telescope we already identiﬁed Cl2-B200 as a carbon star, and it was discarded from further analysis. As already noted by Sneden et al. (1997), M15-S4 is probably a spectroscopic double star, as all lines are signiﬁcantly wider (larger Full Width Half Maximum [FWHM]) than the other stars of M15. It was not used for our abundance analysis.

For each of our targets we made equivalent width (EW) measurements with SPLOT in IRAF, except for the lines with a small EW (. 50 mÅ). For these weak lines, we noticed that SPLOT was giving very unstable FWHM measurements. A home-made gaussian-fitting program was used for these lines to fix the FWHM at the instrumental value. We also used DAOSPEC∗, a new programme that automatically measures EWs by iteratively fitting gaussians of fixed FWHM to all lines in the spectrum and removing the continuum signature (Stetson & Pancino, in preparation). Having confirmed that DAOSPEC gives results compatible with those obtained by hand for lines of moderate strength (EW ≤ 150 mÅ)†, we used DAOSPEC measured EWs for M15.

We can detect a range of elements in our coadded spectra: Fe i, Fe ii, Ti i, Ti ii, O i, Mg i, Ca i, Cr i, Mn i, Ni i, Zn i,Y ii, Ba ii, La ii, Nd ii and Eu ii, which allows us to achieve a comprehensive abundance analysis. The most important is Fe, with an average of 50 measured lines for Fe i and 10 lines for Fe ii. Line parameters and EW measurements for all stars are reported in Table 5.A1. Abundances for the diﬀerent

∗ http://cadcwww.dao.nrc.ca/stetson/daospec/ † We note here for completeness that, at this high resolution, the fixed FWHM gaussian hypothesis adopted by DAOSPEC does not hold for the strongest lines (EW > 150 mÅ) where departures from the gaussianity and natural broadening play a significant role. 82 chapter 5: HR spectroscopy in Fornax Globular Clusters elements were calculated with CALRAI, originally described in Spite (1967) with many improvements over the years. The stellar atmospheres models are those of Plez (private communications, 2000 and 2002, described in Gustafsson et al. 1975, Plez et al. 1992, Gustafsson et al. 2003). Spectral synthesis was required for some elements: Eu, Zn, Mg, Na, O and Ba to account for hyperfine splitting (Eu, Ba); weak lines (Zn, O) and strong, possibility saturated lines (Na, Mg).

Initial guesses were made for the stellar effective temperature (Teff ) using V −I and/or B − V colours, using the Alonso et al. (2001) calibration and a reddening of E(B − V ) = 0.065. The surface gravity (log g) was estimated assuming a 0.8 M mass for the stars, a distance modulus of (m-M)=20.85 mag and bolometric corrections from Alonso et al. (2001). However, the quality of the photometric data we gathered for these stars turned out to be too poor to constrain firmly the star’s effective temperature (only 2 stars had HST photometry in Buonnano et al. 1998, and the other photometric sources were ground-based, suffering from crowding and not all in a homogeneous photometric system). We therefore chose to base our analysis solely on spectroscopic criteria. The Teff , log g and micro-turbulence velocity (vt) were adjusted to insure that we had the ionisation balance of Fe i and Fe ii and that the Fe i abundance is independent of both line strength and excitation potential of the line. Figure 5.3 illustrates the quality of our

Figure 5.3: Observed Curve of growth for Fe i in Cl3-B82. The dotted line marks the [Fe/H]=0 location, while the full line is the theoretical curve of growth for a typical Fe i line with the stellar parameters adopted for this star. 5.3: Data Reduction and Analysis 83 solution by showing the curve of growth obtained for Cl3-B82, where we notice that every part of the curve of growth is well populated. The ﬁnal set of stellar parameters used for each star are shown in Table 5.2. The [Fe/H] in this table is the metallicity of the model used to compute the abundances, not the ﬁnal abundance value of Fe i or Fe ii of the star.

As an additional test, since the S/N reached in the individual spectra was rather limited, we also co-added the spectra of stars with similar parameters within each cluster (all three stars of Cluster 3 on the one hand, and the two cooler stars of Cluster 1 on the other hand), and repeated the analysis. The results are fully consistent with the analysis of the individual stars: Teﬀ , log g and vt are undistinguishable, while the mean [Fe/H] is recovered within 0.02 dex, and most of the other abundance ratios fall well within the star to star scatter.

A significant source of error in our analysis is the uncertainty in measuring the EW. Our errors in the EW determinations were estimated by propagating the EW error estimates (from splot) through the abundance computation (the abundances EW + δEW and EW − δEW were computed and compared to the central adopted value). For elements which were computed by spectral synthesis, the error is estimated by eye, plotting a range of acceptable fits, as illustrated in Figure 5.4 for the weak Eu line in Cl3-B59. Another way to estimate the measurement errors affecting the abundance is to consider

Figure 5.4: The Synthetic spectra for the Eu line at λ = 6645.1 overlaid on the data for Cl3-B59. The middle line is the adopted ﬁt, while the lower and upper ones are the error estimate of ± 0.1 dex. The larger line on the left is a Ni line. 84 chapter 5: HR spectroscopy in Fornax Globular Clusters the dispersion (rms) around the mean. For species with suﬃcient numbers of lines measured (>3), this dispersion was adopted whereas the direct measurement error was used for species probed by fewer lines.

Table 5.2: Adopted parameters of the stellar atmosphere model for each star

Star ID Teﬀ (K) Log g [Fe/H] vt(km/s) Cl1-D56 4600 1.0 -2.60 2.1 Cl1-D68 4350 0.5 -2.60 2.0 Cl1-D164 4400 0.8 -2.60 2.1 Cl2-B71 4450 0.7 -2.10 1.8 Cl2-B77 4350 0.7 -2.10 1.7 Cl2-B226 4250 0.6 -2.10 2.0 Cl3-B59 4400 0.5 -2.30 2.0 Cl3-B61 4400 0.8 -2.30 1.8 Cl3-B82 4350 0.5 -2.30 2.0 M15-S1 4350 0.5 -2.40 1.9 M15-S3 4400 0.6 -2.40 1.8 M15-S4 4150 0.6 -2.30 2.3 M15-S6 4400 0.7 -2.40 1.8 M15-S7 4400 0.4 -2.50 1.9

Table 5.3: Dependencies on model atmosphere parameters

∆ Teﬀ ∆ Log g ∆ vt Combined -200 K -0.3 -0.2 km s−1 [Ba ii/Fe i] -0.11 0.11 -0.06 0.17 [Ca i/Fe i] 0.00 -0.01 0.05 0.05 [Cr i/Fe i] 0.21 -0.03 -0.02 0.21 [Eu ii/Fe i] -0.16 0.13 0.08 0.22 [Fe i/H] 0.27 -0.05 -0.10 0.29 [Fe ii/H] -0.05 0.07 -0.06 0.10 [La ii/Fe i] -0.12 0.12 0.08 0.19 [Mg i/Fe i] -0.06 -0.08 0.02 0.10 [Mn i/Fe i] -0.02 -0.01 0.08 0.08 [Na i/Fe i] 0.18 -0.04 -0.04 0.19 [Nd ii/Fe i] -0.14 0.10 0.04 0.18 [Ni i/Fe i] 0.02 -0.02 -0.03 0.04 [O i/Fe i] -0.11 0.13 0.08 0.19 [Ti i/Fe i] 0.27 -0.02 0.02 0.27 [Ti ii/Fe i] -0.21 0.09 0.02 0.23 [Y ii/Fe i] -0.18 0.09 0.04 0.21 [Zn i/Fe i] -0.28 0.03 0.06 0.29 5.4: Interpretation 85

However, there is more than just the measurement error to consider. The chosen stellar model will also affect the derived abundances. The three important parameters in the model are: temperature, gravity and micro-turbulence velocity. Each of these influences the final abundance in a different way. We estimated the uncertainty in each of these three parameters using the corresponding statistical errors on the slopes and the (Fe i - Fe ii) difference and computed the resultant change in abundance for all elemental ratios. Table 5.3 shows the abundance offset generated by each parameter, and the combined effect of all three (added quadratically).

A summary of our abundance analysis is available in Table 5.A2 (Fornax) and Ta- ble 5.A3 (M15) where we present all of our elements with the associated error estimates and the number of lines used to compute the ratio. Only the EW measurement error is used in plots and tables, and Fe i is used to determine our [el/Fe] ratios.

5.4 Interpretation 5.4.1 The Iron abundance Table 5.4 compares our mean [Fe/H] with the latest results of two different classical methods: RGB slope fitting and integrated spectroscopy. Our abundances appear 0.3 to 0.5 dex lower than previous estimates. Most of this discrepancy is attributable to different reference calibrators. Indeed, both the integrated spectroscopy and isochrone fitting are based on the Zinn & West (1984) metallicity scale which places M15 at h[Fe/H]i= −2.15 and M92 at h[Fe/H]i= −2.24. In contrast, high resolution spectroscopic analyses consis- tently find h[Fe/H]i=−2.4 for M15, including Sneden et al. (1997) and this present work. Meanwhile, M92 is found to be h[Fe/H]i=−2.34 (Sneden et al., 2000). The difference between high resolution spectroscopy and the other indirect methods, due to differences in calibration, is therefore of the order of 0.25 dex, the rest of the discrepancy might be due to the propagation of errors, and indeed appears of the order of the quoted error bars (± 0.2dex). In conclusion, although the absolute value of metallicities presented in the works quoted in Table 5.4 do not appear accurate, the comparison made by the authors with M15 and M92, the most metal-poor clusters known in our Galaxy, is correct. Our analysis reveals that Cluster 1, at h[Fe/H]i= −2.5, is actually the most metal-poor globular cluster yet observed. It is clearly more metal-poor than M15, with weaker iron lines, as can be seen in Figure 5.5, where we compare Cl1-D68 and M15-S1, two RGB stars of similar temperature, surface gravity and micro-turbulence velocity.

Table 5.4: Recent metallicity estimates from diﬀerent methods

Cluster 1 Cluster 2 Cluster 3 Method Reference −2.5 ± 0.1 −2.1 ± 0.1 −2.4 ± 0.1 Individual stars, HR spectra This work N/A −1.76 ± 0.41 −1.84 ± 0.18 Integrated light spectra Strader et al. 2003 −2.20 ± 0.20 −1.78 ± 0.20 −1.96 ± 0.20 RGB Slope Buonanno et al. 1998 86 chapter 5: HR spectroscopy in Fornax Globular Clusters

Figure 5.5: The comparison between Cl1-D68 and M15-S1. These are two RGB stars with similar stellar parameters but a diﬀerence in [Fe/H] of 0.2 dex.

5.4.2 The Alpha elements

Alpha elements come predominantly from Type II supernovae, unlike Fe which comes predominantly from type Ia SN (McWilliam 1997, Tinsley 1979). The [α/Fe] ratios frequently display diﬀerent patterns with respect to Fe in diﬀerent environments (e.g., Shetrone et al. 2001). They are typically overabundant by +0.3 to +0.4 dex in Galactic globular cluster stars and halo stars with respect to solar, as expected in old components where only SNe II have had time to contribute to the chemical enrichment. 5.4: Interpretation 87

Figure 5.6: Alpha elements abundances as a function of [Fe/H]. Filled triangles are for Cluster 1, ﬁlled diamonds are for Cluster 2 and ﬁlled squares are for Cluster 3. Circles are for our M15 stars. Small grey dots are galactic stars and small empty circles are galactic GCs. Upper limits, when present, are shown with one sided arrows, replacing the error bars. See text for more details.

In Figure 5.6, we plot the abundance ratios for α-elements Ca, Mg and O in Fornax dSph globular clusters 1, 2 & 3. Also plotted are the four M15 control stars and, as smaller symbols, Galactic halo stars, taken from the compilation of Venn et al. (2004) and Galactic globular cluster stars from the compilation (averaged by cluster) of Pritzl et al. (2005), except for [O/Fe] points, which are from Shetrone et al. 1996a (individual stars, not averages.) The abundances of Ca, Mg and O are all above the solar value (just like the Galactic halo and globular cluster stars) with a small dispersion and small error 88 chapter 5: HR spectroscopy in Fornax Globular Clusters

Figure 5.7: Titanium abundances as a function of [Fe/H]. The symbols are the same as in Figure 5.6. Separated Ti i and Ti ii were not available for our halo stars, so a global [Ti/Fe] is used for these points. bars. There are a couple of Fornax dSph globular cluster stars with clearly anomalous O and Mg abundance, and they will be discussed later in section 5.4.3. The Fornax dSph globular cluster α-element ratios appear to follow the same patterns found in Galac- tic globular cluster stars, suggesting that the oldest epoch of globular cluster formation is very similar in these two diﬀerent environments. The overabundance of α-elements seen in Galactic globular clusters stars may be interpreted as the number of massive stars present in the early history of our Galaxy assuming that the main contributor to α-elements is SNe II explosions from massive stars. The same over abundance is seen in Fornax dSph globular clusters so this enrichment pattern is not only present in our Galaxy.

Titanium is shown in Figure 5.7, where we chose to compare our results with the Galactic globular clusters studied by Shetrone et al. (2003) rather than the compilation of Pritzl et al. (2005) for homogeneity purposes. At first glance, Ti seems to be underabundant in the Fornax clusters with respect to halo stars in the Milky Way. However, they fall right on top of our M15 and Shetrone’s M30, M68 and M55, close to a solar Ti/Fe ratio. However, we would like to stress that [Ti/Fe] ratios of different authors can be on different scales (depending on the set of Ti lines used and the adopted log gfs), as well illustrated by M15: the [Ti/Fe] ratios in M15 found by Sneden et al. (1997, included in the Pritzl compilation) are ∼0.4 dex higher than in our own analysis of M15, but using 5.4: Interpretation 89

Figure 5.8: Here we show the “Deep-Mixing” abundance anomaly. An anti-correlation of O-Na on the left and a correlation of O-Mg on the right. The symbols are the same as in Figure 5.6.

Sneden’s Ti lines, log gfs and EWs (in the stars we have in common), our analysis yields the same value as Sneden’s. We also notice a small systematic difference between the ratio of Ti i and Ti ii over iron (∼0.2 dex), that could be caused by log gfs (that could be on different scales for Ti i and Ti ii) and/or non-LTE effects . We therefore conclude that, based on the comparison of our Fornax globular clusters with a fully compatible analysis of galactic globular clusters (our analysis of M15 and three other clusters by Shetrone et al. 2003), there is no difference in the Ti/Fe ratios observed in Fornax and MW globular clusters.

5.4.3 Deep mixing pattern

Deep-mixing occurs when material processed deep inside a star finds its way to the upper atmosphere, thus modifying the original abundance pattern. Proton-capture nucleosynthesis converts O, N, Ne to Na, and Mg to Al in the H fusion layer of evolved RGB stars. This means that a significant atmospheric depletion of O caused by deep-mixing should be accompanied by an enhancements of Na (Langer et al. 1993) and similarly an enhancement in Al should cause observable Mg depletion (Langer & Hoffman 1995). Such patterns (anti-correlations of O-Na and Mg-Al) are found in galactic globular cluster stars but not in comparable field stars of our Galaxy (Gratton et al. 2004), or any other (e.g., Shetrone et al. 2001). It is assumed that this is caused by environmental effects within a star cluster but whether it is the result of deep-mixing within the RGB stars that are observed or the fossil traces of self-pollution of the globular cluster during its formation process, or a combination of the two, is not well understood. 90 chapter 5: HR spectroscopy in Fornax Globular Clusters

Figure 5.8 shows that deep mixing patterns are not only found in galactic globular clusters and the old clusters of the Large Magellanic Cloud [LMC] (Hill et al. 2000) but also in clusters of much smaller dwarf spheroidal galaxies like Fornax. The anti- correlation O-Na is visible in two (Cl3-B82 and Cl1-D164) of the nine stars we observed in the Fornax globular clusters displaying high Na and low O abundances (left panel), accompanied by low Mg abundances (correlation O-Mg, right panel). We cannot check directly whether the Mg-Al anti-correlation also exists in these clusters, since we did not detect Al in our Fornax dSph globular cluster spectra, because the Al lines present in our spectral range are too weak. Our detection limit is about 14 mÅ, which translates into an upper limit to [Al/Fe] of 1.4. Shetrone et al. (1996b) found that the usual enhancement of Al ranges from 0.5 to 1.0 dex, thus largely consistent with our upper limit.

5.4.4 Iron-peak elements The Fe-peak elements we observed in the Fornax dSph globular clusters are Cr, Ni and Zn, and they are shown in Figure 5.9. Comparison points for Galactic globular clusters are from the compilation of Pritzl et al. (2005), and Galactic halo stars are from the Hamburg-ESO (HERES) survey (Barklem et al. 2005) for Cr (top panel), from the compilation of Venn et al. (2004) for Ni (middle panel), and from Sneden et al.(1991) and Barklem et al. (2005) for Zn (lower panel).

Cr is believed to be produced mainly by incomplete explosive silicon burning (Woosley & Weaver 1995). Despite large error bars in our measurements of the Fornax dSph globular cluster stars, there seems to be an increase (by ∼0.3 dex) of the [Cr/Fe] ratio between the two more metal-poor clusters and the more metal-rich Cluster 2. Such a trend of increasing [Cr/Fe] with increasing [Fe/H] has been observed in Galactic ﬁeld stars (McWilliam et al. 1995, Carretta et al. 2002), leading to a similar ∼0.3 dex increase, but over a much wider metallicity range (−3.5 to −2.). Newer, high quality observations by Cayrel et al. (2004) of Galactic halo stars further reduced the observed slope of increasing [Cr/Fe] with increasing metallicity to ∼0.15 dex over a [Fe/H] range from −2.5 to −4 dex, with an extremely small intrinsic scatter (σ = 0.05 dex). The higher [Cr/Fe] abundance observed in Cluster 2 therefore seems unlikely, and is probably caused by our observational errors.

Ni is believed to be produced in complete explosive silicon burning. We don’t expect any relation between [Ni/Fe] as a function of [Fe/H], based on what we see in the MW. Even at this low metallicity, the relation is ﬂat with a value close to zero, within the error bars, as we can see in Figure 5.9. This is consistent with the majority of Galactic globular clusters, open clusters and halo stars (Sneden et al. 2004). So yet again, the Fornax dSph globular clusters are similar to the normal Galactic globular clusters.

Zn has the same origin as Ni, but it has been suggested (Heger & Woosley 2002) that it could also be formed by neutron capture, and be either an r-process or an s- process element. Our results, more than half of which are upper limits, are consistent with Galactic values. 5.4: Interpretation 91

Figure 5.9: Iron-peak elements abundances as a function of [Fe/H]. The symbols are the same as in Figure 5.6. 92 chapter 5: HR spectroscopy in Fornax Globular Clusters

Figure 5.10: Heavy elements abundances as a function of [Fe/H]. The symbols are the same as in Figure 5.6.

5.4.5 Heavy elements

The heavy elements Y, Ba and Eu in the Fornax dSph globular clusters are plotted in Figure 5.10. [Y/Fe] appears to be consistent with what is observed in Galactic globular clusters. However, Cluster 1 and 3 (the two most metal-poor) appear to have higher [Ba/Fe] than average for Galactic globular clusters. As shown in Figure 5.4), europium is measured from a single weak line, and could only be detected in Cluster 3 (all other Fornax points in this plot are upper limits), in which [Eu/Fe] is particularly high, above the typical range for Galactic globular clusters. 5.4: Interpretation 93

Ba and Y are neutron-capture elements which are, in the solar system, dominated by the s-process, a process due to low to intermediate-mass Asymptotic Giant Branch (AGB) stars, with only a minor contribution from the r-process. Eu on the other hand is almost entirely dominated by the r-process, which requires more extreme neutron ﬂuxes, such as SN II explosions (McWilliam 1997) associated with massive stars. In the Milky Way, with decreasing metallicities the s-process contribution gradually decreases (consistent with the timescale of AGB evolution) so that below ∼ −2.5dex, both in ﬁeld and globular clusters stars, all heavy elements are dominated exclusively by the r-process. (Johnson et al. 2001, James et al. 2004, Barklem et al. 2005). In Cluster 3, we detect, not only Eu, but also other heavy elements represented by weak lines preventing detection in the other clusters: Nd and La.

In Figure 5.11, we compare Cluster 3 log ()∗ values to the solar system r− and s− process abundances (Burris et al. 2000). The solar system elemental abundances have been shifted by the difference between the mean values of Eu for Cluster 3 and the solar system abundance distribution (−1.55 dex). Clearly, the abundances of most elements in the Fornax globular clusters match the solar system r-process pattern within the observational uncertainties (with the exception of La which seems to be matched by neither the r− nor the s− process patterns). Cluster 3 stars are obviously very close to the r-process expectations, confirming that, similarly to the most metal-poor globular clusters in the galactic Halo (M15, M92, M68), cluster 3 is also dominated by the r-process. This is also confirmed by the [Ba/Eu] ratio observed in the three Cluster 3 stars [Ba/Eu] = −0.62, −0.69, −0.7 (±0.20), very close to the −0.69 for the r-process component in the solar system (as compared to +1.15 for the s-process, Arlandini et al. 1999). The upper limits for Eu in Clusters 1 and 2, although not decisive, are also compatible with a pure r-process enrichment ([Ba/Eu]> −1.00 to −0.82). This result indicates that, in Fornax dSph as in our Galaxy, heavier neutron capture elements in the lowest metallicity stars have only very weak s-process contribution. Or in other words, that heavy elements in Fornax dSph globular clusters, as in M15, are formed principally through the r-process.

The high neutron-capture element content of Cluster 3, that we attribute to the r- process, is similar for all three stars, and above the upper edge of the range of values traditionally covered by the Galactic globular clusters (Pritzl et al. 2005). R-process enrichments of this order or even higher are found in Galactic halo ﬁeld stars (Barklem et al. 2005), but as far as Galactic globular clusters are concerned, the only case known to date is M15. Sneden et al. (1997, 2000) have established that M15 has a stellar bi-modality with one group being strongly overabundant in [Eu/Fe] (and [Ba/Fe]) compared to the other. Our observations of three stars in Cluster 3 do not provide suﬃcient statistics to determine if this cluster also has a bimodality (with our 3 stars by chance happening to belong to the high Eu group) or if all stars in Cluster 3 are Eu-rich.

Finally, despite the dispersion in Ba that seems to exist among the three Fornax globular clusters (Cluster 3 being the most Ba and Eu rich), Y is very similar from cluster to cluster, and comparable to the Galactic abundances of this element (globular clusters and ﬁeld stars). This also leads Cluster 3 to have a Ba/Y ratio higher than in the two other clusters ([Ba/Y]=+0.43 compared to +0.0 in Cluster 2, and marginally

∗ The scale used for log () is the standard astronomical scale (log10(Nel/NH) + 12). 94 chapter 5: HR spectroscopy in Fornax Globular Clusters

Figure 5.11: The relative contributions of the r− and s− processes for the heavy elements in Cluster 3 (ﬁlled circle). The solar r− and s− process abundances, traced by a dotted and a full line respectively, are taken from Burris et al. (2000). They are shifted by the diﬀerence between Cluster 3 and the solar system abundance for Eu (−1.55 dex).

higher than the +0.28 dex observed in Cluster 1). Interestingly, the [Ba/Y] observed in the three Fornax clusters are yet again very similar to that of the Galactic globular clusters and halo ﬁeld stars, whereas the (on average more metal-rich) ﬁeld stars in dwarf spheroidal galaxies have been shown to display systematically higher [Ba/Y] than their galactic counterparts (Venn et al. 2004). 5.5: Conclusions 95

Figure 5.12: The cluster mean elemental abundances of the three Fornax dSph globular clusters and M15. Each individual stellar abundance has being weighted by its error. Cluster 1 is identiﬁed by a ﬁlled square, Cluster 2 by a star, Cluster 3 by a cross and M15 by a triangle.

5.5 Conclusions

We have compared the properties of the globular clusters belonging to the Fornax dSph with those of the Milky Way with unprecedented accuracy. The Fornax dSph contains clusters with a range of properties such as metallicity, central concentration and Horizon- tal Branch structure. For the first time detailed chemical abundances have been derived for a sample of stars in a globular cluster system in an external galaxy, apart from the Magellanic Clouds. Despite their very different mass, morphology and global star formation history, the Fornax dSph and the Milky Way appear to have experienced the same 96 chapter 5: HR spectroscopy in Fornax Globular Clusters very early enrichment conditions and in particular similar nucleosynthesis. This is summarised in Figure 5.12, where the mean elemental abundances, each being weighted by its error, of the three Fornax globular clusters and M15 are compared. The abundance patterns of the individual stars in Milky Way globular clusters and Fornax globular clusters match each other almost perfectly. We find that the star-to-star abundance dispersion in the Fornax clusters is modest and compatible with similar observations of Galactic globular clusters.

We have definitively established that the Fornax globular Clusters 1, 2 and 3 are very metal-poor, slightly poorer than previous estimates, with respectively h[Fe/H]i=−2.5, −2.1 and −2.4. Part of the discrepancy with previous studies is explained by the different reference calibrations used. Cluster 1 is now the most metal-poor globular cluster known, however the difference between Cluster 1 and M92 or M15 in the Milky Way is small. There seems to be universal lower limit to the metallicity at which star clusters form, which is higher than that of field stars in the halo of our Galaxy, where significant numbers of stars are found with [Fe/H] < −4. It is also clear, that as in our Galaxy, the ratio of the number of globular cluster to the number of field stars strongly decreases with rising metallicity (Harris & Harris 2002).

Clusters 1, 2 and 3 were clearly formed promptly and early in the history of Fornax dSph, alike the Milky Way globular clusters. They are over abundant in α-elements (O, Mg, Ca) at a similar level to Galactic clusters at identical [Fe/H], and the heavy element abundances (Y, Ba, Eu) in the 3 clusters are compatible with dominant r-process enrichment. Finally, the Fe-peak elements are also very similar to Galactic globular cluster values, with [Ni/Fe] being unambiguously solar in all three clusters and Zn and Cr are also compatible with Galactic values.

The analogy between Galactic and Fornax dSph holds even in the rare cases and anomalies: (i) Eu is extremely overabundant in Cluster 3 stars. The only Galactic counterpart known to date is M15. (ii) Cl1-D164 and Cl3-B82 show low O and Mg associated with a high Na abundance, thus establishing an O-Na anti-correlation and O-Mg correlation. This is the same deep-mixing pattern observed in Galactic star clusters, and old LMC clusters.

The effort towards a comprehensive description of the formation and evolution of the Fornax dSph will soon benefit from the analysis of VLT/FLAMES high resolution spectra of a hundred field stars (Letarte et al., in preparation). It will then be possible to describe the chemical enrichment and nucleosynthetic processes dominant for the field star population compared to that found in the globular clusters, and to see when and if the similarities in enrichment patterns with our Galaxy end. 5.A: Large tables 97 Acknowledgements:

We gratefully acknowledge Carlo Emanuele Corsi and Roberto Buonanno for providing their photometry for our target selection. ET gratefully acknowledges support from a fellowship of the Royal Netherlands Academy of Arts and Sciences. BL is funded by a grant from the Netherlands Organisation for Scientiﬁc Research (NWO). PJ and GM gratefully acknowledge support from the Swiss National Science Foundation (SNSF).

Appendix 5.A Large tables

Table 5.A1: Line parameters and equivalent widths for the Fornax globular clusters and M15. When there is a * in the S column, it indicates that a synthetic spectra was used for the abundance determination. HFS indicates a line with hyperﬁne splitting, so no individual EW measurement for that line is available. (Part 1)

λ El χex log gf S Cl1-D56 Cl1-D68 Cl1-D164 Cl2-B71 Cl2-B77 Cl2-B226 4934.12 Ba ii 0.00 -0.703 * HFS HFS HFS HFS ...... 5853.69 Ba ii 0.60 -1.010 * HFS HFS HFS HFS HFS HFS 6141.73 Ba ii 0.70 -0.077 * HFS HFS HFS HFS HFS HFS 6496.91 Ba ii 0.60 -0.380 * 103.7 136.8 123.6 121.5 115.0 138.2 6102.73 Ca i 1.88 -0.790 50.4 54.5 50.8 91.9 97.3 116.0 6122.23 Ca i 1.89 -0.320 85.0 93.5 86.5 129.0 119.2 160.1 6161.30 Ca i 2.52 -1.270 ...... 26.3 6166.44 Ca i 2.52 -1.140 ...... 27.7 ... 32.8 6169.04 Ca i 2.52 -0.800 30.2 14.6 12.7 33.3 41.7 51.5 6169.56 Ca i 2.52 -0.480 ... 24.1 17.9 41.1 45.7 58.8 6439.08 Ca i 2.52 0.390 77.5 69.2 72.4 103.1 114.0 139.8 6455.60 Ca i 2.52 -1.290 ...... 30.3 6499.65 Ca i 2.52 -0.820 24.1 22.1 11.2 31.3 35.1 58.9 5206.04 Cr i 0.94 0.019 100.0 124.2 109.0 126.4 170.5 214.4 5409.80 Cr i 1.03 -0.720 66.3 74.2 74.8 95.5 114.1 148.9 6645.13 Eu ii 1.37 0.200 * ... HFS HFS HFS HFS HFS 4966.10 Fe i 3.33 -0.890 51.3 54.0 39.8 68.2 84.0 85.1 5006.12 Fe i 2.83 -0.628 82.0 91.4 108.2 117.1 112.4 140.0 5079.75 Fe i 0.99 -3.240 ... 127.1 119.6 141.9 136.6 193.0 5083.35 Fe i 0.96 -2.862 100.3 124.7 122.3 138.8 135.0 181.7 5150.85 Fe i 0.99 -3.030 84.7 114.4 112.7 117.4 129.9 178.3 5151.92 Fe i 1.01 -3.326 69.7 115.4 90.3 104.4 118.9 155.9 5162.29 Fe i 4.18 0.020 64.5 32.0 44.7 76.3 67.2 90.2 5166.28 Fe i 0.00 -4.200 110.5 ... 130.8 140.5 180.3 198.8 5171.61 Fe i 1.48 -1.751 126.3 133.7 126.4 132.6 149.0 178.1 5192.34 Fe i 3.00 -0.520 98.0 89.1 79.0 104.7 124.7 134.1 5196.08 Fe i 4.26 -0.450 ...... 41.8 9.3 33.2 5215.19 Fe i 3.27 -0.930 41.4 35.8 43.9 76.4 81.9 86.7 5216.28 Fe i 1.61 -2.102 97.6 104.9 107.0 115.1 128.2 165.4 5217.30 Fe i 3.21 -1.270 36.0 ... 45.4 63.4 ... 93.3 5232.95 Fe i 2.94 -0.067 122.3 122.8 104.6 130.9 135.9 166.2 5250.21 Fe i 0.12 -4.700 59.0 78.4 78.8 95.0 99.2 145.1 Continued on next page 98 chapter 5: HR spectroscopy in Fornax Globular Clusters

λ El χex log gf S Cl1-D56 Cl1-D68 Cl1-D164 Cl2-B71 Cl2-B77 Cl2-B226 5307.37 Fe i 1.61 -2.812 58.2 76.3 70.9 91.7 92.4 129.1 5324.19 Fe i 3.21 -0.100 86.4 84.0 100.6 127.7 114.8 146.2 5339.93 Fe i 3.27 -0.680 65.8 59.3 52.1 111.3 99.2 123.6 5364.86 Fe i 4.45 0.220 ...... 32.8 52.1 92.8 72.2 5367.48 Fe i 4.42 0.550 34.3 ... 37.7 54.7 56.7 78.2 5369.96 Fe i 4.37 0.540 64.2 ... 26.0 74.3 68.3 88.6 5371.50 Fe i 0.96 -1.644 205.6 195.4 188.1 171.9 196.1 273.1 5383.37 Fe i 4.31 0.500 49.7 49.1 51.4 79.1 59.9 102.9 5393.17 Fe i 3.24 -0.920 58.3 57.2 70.4 80.9 116.0 111.7 5397.14 Fe i 0.91 -1.992 161.7 181.1 172.3 182.3 198.0 246.3 5405.79 Fe i 0.99 -1.852 159.2 190.7 163.8 173.6 201.6 230.3 5415.19 Fe i 4.39 0.510 54.3 ... 50.3 68.3 91.9 96.0 5424.07 Fe i 4.32 0.520 72.8 67.7 57.2 80.9 94.1 91.2 5501.48 Fe i 0.96 -3.050 114.1 132.9 124.7 143.4 140.8 174.0 5506.79 Fe i 0.99 -2.790 111.2 131.9 134.3 157.9 149.1 206.1 5615.66 Fe i 3.33 0.050 84.7 88.8 94.9 115.0 124.5 149.4 5956.70 Fe i 0.86 -4.570 40.8 37.5 ... 51.1 ... 99.2 6024.05 Fe i 4.55 -0.110 18.6 12.0 ... 28.8 49.1 54.3 6136.62 Fe i 2.45 -1.500 98.8 91.3 92.3 119.8 109.6 159.3 6137.70 Fe i 2.59 -1.366 77.7 77.1 83.7 109.1 113.7 143.9 6157.75 Fe i 4.07 -1.260 ...... 6173.34 Fe i 2.22 -2.850 29.8 21.5 21.8 45.8 62.2 80.5 6191.57 Fe i 2.43 -1.416 76.3 92.6 93.4 121.7 117.8 140.8 6213.43 Fe i 2.22 -2.660 34.6 39.9 41.0 68.5 59.0 91.0 6219.29 Fe i 2.20 -2.438 46.6 60.3 52.4 87.5 86.6 109.0 6229.23 Fe i 2.84 -2.900 ...... 33.0 24.7 6230.74 Fe i 2.56 -1.276 84.1 93.2 94.6 115.6 49.3 156.8 6232.64 Fe i 3.65 -0.960 ...... 16.5 40.3 45.1 45.3 6252.57 Fe i 2.40 -1.757 82.8 82.9 98.1 110.2 124.0 146.5 6270.23 Fe i 2.85 -2.610 ...... 16.9 24.1 37.5 6297.80 Fe i 2.22 -2.740 ... 57.3 61.8 68.8 80.5 113.5 6301.50 Fe i 3.65 -0.720 62.0 51.3 44.4 68.0 81.4 106.5 6302.49 Fe i 3.69 -1.150 ... 9.2 8.8 23.8 51.0 57.1 6393.61 Fe i 2.43 -1.630 68.8 94.3 95.9 86.4 137.7 154.7 6421.36 Fe i 2.28 -2.014 52.9 72.9 71.2 90.4 136.0 142.0 6430.86 Fe i 2.18 -1.946 66.9 87.8 88.5 94.7 ... 148.3 6481.87 Fe i 2.27 -2.980 ... 27.9 23.9 42.8 57.0 79.8 6498.94 Fe i 0.96 -4.690 ...... 30.1 38.1 57.7 98.7 6518.37 Fe i 2.83 -2.460 ...... 23.8 47.7 6574.23 Fe i 0.99 -5.020 ... 11.5 7.5 ... 38.2 64.3 6593.88 Fe i 2.43 -2.390 ... 33.6 27.2 69.8 66.1 89.1 6609.12 Fe i 2.56 -2.660 ... 15.1 6.7 37.4 55.7 63.6 4923.92 Fe ii 2.89 -1.320 117.7 125.4 122.9 130.4 114.8 132.4 5197.57 Fe ii 3.23 -2.100 58.3 43.8 49.7 65.6 77.3 78.4 5234.63 Fe ii 3.22 -2.118 64.1 52.6 52.1 ... 69.2 80.0 5276.00 Fe ii 3.20 -1.950 96.5 50.1 55.3 87.9 93.2 92.4 5284.10 Fe ii 2.89 -3.190 ...... 21.2 ... 37.0 52.8 5325.56 Fe ii 3.22 -2.600 ...... 5425.25 Fe ii 3.20 -3.360 ...... 22.8 34.5 ...... 5534.85 Fe ii 3.24 -2.920 27.9 27.8 24.4 46.8 42.7 43.9 Continued on next page 5.A: Large tables 99

λ El χex log gf S Cl1-D56 Cl1-D68 Cl1-D164 Cl2-B71 Cl2-B77 Cl2-B226 5991.38 Fe ii 3.15 -3.740 ...... 6149.25 Fe ii 3.89 -2.720 ...... 24.1 ... 6238.38 Fe ii 3.89 -2.480 ...... 6369.46 Fe ii 2.89 -4.250 ...... 6432.68 Fe ii 2.89 -3.710 ... 17.5 ... 19.0 ...... 6456.39 Fe ii 3.90 -2.080 ... 6.8 19.8 ...... 32.2 6516.08 Fe ii 2.89 -3.450 ...... 35.9 35.1 40.8 5301.97 La ii 0.40 -1.140 * ...... 5303.52 La ii 0.32 -1.350 * ...... 6320.43 La ii 0.17 -1.562 * ...... 6390.46 La ii 0.32 -1.400 * ...... 6774.27 La ii 0.13 -1.708 * ...... 5172.70 Mg i 2.71 -0.390 * 283.0 266.1 241.8 312.6 340.9 ... 5528.41 Mg i 4.35 -0.357 * 110.2 93.3 79.0 130.7 141.1 135.8 5711.09 Mg i 4.35 -1.728 * 26.2 12.1 8.1 48.5 46.9 53.1 6013.51 Mn i 3.07 -0.252 * ...... 25.9 6021.82 Mn i 3.08 0.035 * ...... 40.9 5889.97 Na i 0.00 0.122 * 230.5 219.7 249.6 230.4 231.4 ... 5895.94 Na i 0.00 -0.184 * 204.0 185.1 240.0 216.1 215.6 ... 6154.23 Na i 2.10 -1.560 ...... 5249.59 Nd ii 0.98 0.217 * ... 27.5 ...... 21.1 5319.82 Nd ii 0.55 -0.194 * 37.9 30.8 30.7 ...... 33.5 5476.92 Ni i 1.83 -0.890 152.1 124.5 96.3 96.5 131.6 158.1 6176.82 Ni i 4.09 -0.430 ...... 8.6 22.1 ...... 6177.25 Ni i 1.83 -3.500 ...... 3.4 ...... 21.1 6300.31 O i 0.00 -9.760 * 19.3 11.9 10.0 21.6 23.8 45.0 4840.87 Ti i 0.90 -0.450 40.6 23.4 17.6 48.8 64.8 69.8 4913.62 Ti i 1.87 0.216 ...... 18.5 39.0 ... 41.2 5014.24 Ti i 0.81 0.910 97.2 96.3 95.3 154.4 173.8 204.8 5016.16 Ti i 0.85 -0.510 31.4 26.5 ... 43.7 69.2 80.3 5064.65 Ti i 0.05 -0.930 60.9 73.3 63.8 86.1 99.7 148.0 5210.39 Ti i 0.05 -0.580 60.5 76.3 77.2 98.1 115.0 164.9 4798.53 Ti ii 1.08 -2.670 54.1 36.3 ... 45.6 ...... 5129.16 Ti ii 1.89 -1.390 63.2 54.4 55.5 63.4 104.1 96.5 5154.07 Ti ii 1.57 -1.520 62.2 69.6 71.2 ... 78.7 96.1 5226.55 Ti ii 1.57 -1.000 90.2 86.7 97.6 105.2 124.6 126.2 5381.01 Ti ii 1.57 -1.780 54.2 39.9 47.2 71.8 81.5 97.1 5418.77 Ti ii 1.58 -2.110 ... 40.6 55.5 67.4 60.5 67.9 4883.69 Y ii 1.08 0.070 41.4 41.0 37.2 56.8 82.3 73.4 4900.11 Y ii 1.03 -0.090 80.2 ...... 65.6 ... 160.8 5087.43 Y ii 1.08 -0.170 ... 32.9 29.2 66.4 38.9 64.0 5200.42 Y ii 0.99 -0.570 34.5 29.9 ... 49.4 40.9 53.5 4810.54 Zn i 4.08 -0.170 * 22.0 40.9 31.9 32.9 43.0 50.0 100 chapter 5: HR spectroscopy in Fornax Globular Clusters

λ El χex log gf S Cl3-B59 Cl3-B61 Cl3-B82 M15S1 M15S3 M15S6 M15S7 4934.12 Ba ii 0.00 -0.703 * HFS ...... HFS HFS HFS HFS 5853.69 Ba ii 0.60 -1.010 * ... HFS ... HFS HFS HFS HFS 6141.73 Ba ii 0.70 -0.077 * HFS HFS HFS HFS HFS HFS HFS 6496.91 Ba ii 0.60 -0.380 * 144.8 123.1 147.2 124.8 136.2 149.2 118.4 6102.73 Ca i 1.88 -0.790 60.2 63.4 72.4 77.7 74.8 79.9 71.5 6122.23 Ca i 1.89 -0.320 107.0 104.5 103.8 116.1 110.6 114.2 105.9 6161.30 Ca i 2.52 -1.270 ...... 11.8 14.0 9.7 9.0 6166.44 Ca i 2.52 -1.140 ...... 17.6 16.7 ... 16.9 6169.04 Ca i 2.52 -0.800 ... 21.4 33.6 28.4 28.4 32.0 24.7 6169.56 Ca i 2.52 -0.480 ... 41.0 ... 43.4 39.2 44.2 38.3 6439.08 Ca i 2.52 0.390 84.5 88.2 97.1 99.7 95.7 99.4 95.9 6455.60 Ca i 2.52 -1.290 26.0 ...... 13.1 10.5 9.5 10.0 6499.65 Ca i 2.52 -0.820 ... 20.0 25.6 26.8 23.4 26.4 23.6 5206.04 Cr i 0.94 0.019 136.3 129.2 129.3 126.9 115.7 121.3 112.2 5409.80 Cr i 1.03 -0.720 89.3 85.3 86.3 88.1 78.8 84.5 74.0 6645.13 Eu ii 1.37 0.200 * HFS HFS HFS HFS HFS HFS HFS 4966.10 Fe i 3.33 -0.890 53.5 42.6 57.8 63.3 60.4 64.2 51.9 5006.12 Fe i 2.83 -0.628 119.5 109.7 113.7 107.0 100.9 106.7 97.3 5079.75 Fe i 0.99 -3.240 133.1 116.6 130.0 118.6 109.7 115.2 101.7 5083.35 Fe i 0.96 -2.862 126.3 114.5 134.0 130.5 120.6 122.2 114.8 5150.85 Fe i 0.99 -3.030 150.6 114.2 128.0 122.4 106.3 111.9 100.6 5151.92 Fe i 1.01 -3.326 114.1 120.3 143.8 111.5 ... 102.6 91.4 5162.29 Fe i 4.18 0.020 50.0 62.3 46.7 49.9 44.1 49.9 43.0 5166.28 Fe i 0.00 -4.200 153.3 159.4 146.4 144.6 128.0 131.9 121.0 5171.61 Fe i 1.48 -1.751 141.7 142.0 142.3 145.8 133.8 142.0 133.7 5192.34 Fe i 3.00 -0.520 122.7 82.0 110.5 102.2 96.8 97.4 96.0 5196.08 Fe i 4.26 -0.450 18.7 23.4 ... 11.8 9.4 8.8 9.1 5215.19 Fe i 3.27 -0.930 58.2 47.0 47.3 53.8 52.7 56.0 47.6 5216.28 Fe i 1.61 -2.102 130.4 119.2 135.9 123.1 110.6 118.4 108.5 5217.30 Fe i 3.21 -1.270 44.5 43.2 ... 50.5 48.7 49.9 ... 5232.95 Fe i 2.94 -0.067 125.1 113.3 123.2 124.5 118.8 122.1 112.3 5250.21 Fe i 0.12 -4.700 102.3 87.7 95.2 92.5 76.5 83.3 67.9 5307.37 Fe i 1.61 -2.812 74.3 74.1 83.6 77.8 68.2 73.3 65.8 5324.19 Fe i 3.21 -0.100 106.4 100.7 105.9 105.7 99.6 104.3 84.3 5339.93 Fe i 3.27 -0.680 86.6 71.5 92.7 72.7 66.7 72.1 59.0 5364.86 Fe i 4.45 0.220 53.6 45.5 40.7 40.5 37.6 40.5 32.3 5367.48 Fe i 4.42 0.550 51.5 47.8 42.6 45.8 42.8 46.5 38.8 5369.96 Fe i 4.37 0.540 72.1 55.3 59.0 49.9 50.6 50.5 47.5 5371.50 Fe i 0.96 -1.644 191.1 184.0 201.7 187.2 177.4 180.7 171.3 5383.37 Fe i 4.31 0.500 60.3 60.0 64.6 61.6 57.7 59.6 50.9 5393.17 Fe i 3.24 -0.920 76.1 70.8 65.3 71.6 63.6 69.0 61.2 5397.14 Fe i 0.91 -1.992 171.5 180.6 190.1 177.9 163.5 169.4 157.8 5405.79 Fe i 0.99 -1.852 183.9 184.4 182.0 175.3 165.0 172.6 157.7 5415.19 Fe i 4.39 0.510 48.4 41.0 58.6 59.9 55.5 57.1 50.4 Continued on next page 5.A: Large tables 101

λ El χex log gf S Cl3-B59 Cl3-B61 Cl3-B82 M15S1 M15S3 M15S6 M15S7 5424.07 Fe i 4.32 0.520 78.3 72.4 71.0 65.1 63.0 66.0 55.5 5501.48 Fe i 0.96 -3.050 133.0 137.9 149.6 135.2 120.4 127.1 114.5 5506.79 Fe i 0.99 -2.790 142.3 122.2 150.1 144.3 131.0 135.3 127.8 5615.66 Fe i 3.33 0.050 104.0 95.0 111.9 105.8 98.5 102.8 95.0 5956.70 Fe i 0.86 -4.570 49.8 61.8 55.8 49.9 37.9 37.0 29.0 6024.05 Fe i 4.55 -0.110 32.1 29.1 22.7 25.3 20.7 24.6 19.8 6136.62 Fe i 2.45 -1.500 100.0 104.5 122.5 110.5 98.9 106.1 94.4 6137.70 Fe i 2.59 -1.366 104.0 88.9 105.5 98.8 93.0 99.0 89.4 6157.75 Fe i 4.07 -1.260 ...... 8.8 8.5 9.6 6.3 6173.34 Fe i 2.22 -2.850 34.4 32.3 36.8 40.2 32.9 39.9 33.0 6191.57 Fe i 2.43 -1.416 123.1 86.8 109.4 109.0 100.7 101.4 94.7 6213.43 Fe i 2.22 -2.660 45.2 50.0 47.8 57.9 52.1 55.5 45.5 6219.29 Fe i 2.20 -2.438 78.6 64.7 60.6 66.6 59.5 68.7 53.0 6229.23 Fe i 2.84 -2.900 ...... 16.7 8.3 7.7 11.5 7.0 6230.74 Fe i 2.56 -1.276 107.2 105.2 116.5 115.2 105.4 112.6 102.2 6232.64 Fe i 3.65 -0.960 ... 19.0 ...... 6252.57 Fe i 2.40 -1.757 105.2 100.7 100.9 101.7 95.0 94.3 86.8 6270.23 Fe i 2.85 -2.610 ...... 20.4 ...... 6297.80 Fe i 2.22 -2.740 ... 70.0 59.9 69.0 65.7 ...... 6301.50 Fe i 3.65 -0.720 50.7 53.1 50.1 45.2 43.7 44.7 36.3 6302.49 Fe i 3.69 -1.150 ...... 21.9 24.8 ... 6393.61 Fe i 2.43 -1.630 94.8 72.0 102.6 104.9 93.2 99.3 91.1 6421.36 Fe i 2.28 -2.014 91.1 80.7 90.5 94.7 83.9 91.0 79.8 6430.86 Fe i 2.18 -1.946 98.4 97.5 110.6 105.3 92.1 98.5 90.3 6481.87 Fe i 2.27 -2.980 25.7 37.5 35.0 ...... 6498.94 Fe i 0.96 -4.690 35.8 37.0 40.1 41.0 28.6 34.3 23.3 6518.37 Fe i 2.83 -2.460 ... 19.2 17.1 28.7 29.4 ...... 6574.23 Fe i 0.99 -5.020 ... 19.9 29.2 23.2 16.5 20.8 13.9 6593.88 Fe i 2.43 -2.390 57.6 42.6 59.5 52.9 43.2 53.2 39.4 6609.12 Fe i 2.56 -2.660 ... 36.6 ... 24.7 20.4 20.8 16.3 4923.92 Fe ii 2.89 -1.320 117.3 115.6 117.7 122.4 121.7 124.8 118.6 5197.57 Fe ii 3.23 -2.100 38.6 52.9 45.4 54.9 51.6 57.5 51.2 5234.63 Fe ii 3.22 -2.118 55.1 121.1 ... 59.0 58.9 59.7 51.4 5276.00 Fe ii 3.20 -1.950 87.8 68.5 77.3 69.7 69.7 68.3 61.4 5284.10 Fe ii 2.89 -3.190 40.7 41.9 40.5 29.9 27.5 31.6 24.6 5325.56 Fe ii 3.22 -2.600 9.1 ...... 14.9 14.2 15.4 14.2 5425.25 Fe ii 3.20 -3.360 ...... 14.1 10.8 13.0 9.3 5534.85 Fe ii 3.24 -2.920 38.6 ... 27.8 28.0 27.7 26.2 24.0 5991.38 Fe ii 3.15 -3.740 ...... 14.2 8.4 14.8 10.1 6149.25 Fe ii 3.89 -2.720 ...... 6.5 4.7 8.8 8.6 6238.38 Fe ii 3.89 -2.480 ...... 13.8 12.2 14.9 11.8 6369.46 Fe ii 2.89 -4.250 ...... 10.1 ...... 6432.68 Fe ii 2.89 -3.710 ...... 20.0 18.1 16.7 20.3 18.2 6456.39 Fe ii 3.90 -2.080 ...... 26.6 28.0 28.6 26.5 6516.08 Fe ii 2.89 -3.450 ...... 27.9 31.6 26.7 26.4 23.3 5301.97 La ii 0.40 -1.140 * 16.0 ... 22.3 ...... 5303.52 La ii 0.32 -1.350 * 12.9 ... 13.7 ...... 6320.43 La ii 0.17 -1.562 * 19.3 15.6 16.9 ...... 6390.46 La ii 0.32 -1.400 * 18.2 25.8 23.4 ...... 6774.27 La ii 0.13 -1.708 * 14.2 9.2 15.2 ...... Continued on next page 102 chapter 5: HR spectroscopy in Fornax Globular Clusters

λ El χex log gf S Cl3-B59 Cl3-B61 Cl3-B82 M15S1 M15S3 M15S6 M15S7 5172.70 Mg i 2.71 -0.390 * 276.4 260.0 204.2 264.1 266.4 237.1 247.9 5528.41 Mg i 4.35 -0.357 * 96.0 102.6 60.0 106.7 101.5 81.1 108.1 5711.09 Mg i 4.35 -1.728 * 18.1 27.6 8.6 22.5 20.0 11.1 19.6 6013.51 Mn i 3.07 -0.252 * ... 15.0 15.6 ...... 6021.82 Mn i 3.08 0.035 * ... 25.0 26.2 ...... 5889.97 Na i 0.00 0.122 * 242.1 206.2 251.6 240.6 296.2 302.3 237.7 5895.94 Na i 0.00 -0.184 * 217.3 199.4 271.0 216.2 257.3 261.2 194.3 6154.23 Na i 2.10 -1.560 ...... 5249.59 Nd ii 0.98 0.217 * 45.2 14.4 46.3 12.5 21.8 23.0 8.3 5319.82 Nd ii 0.55 -0.194 * 55.2 37.2 62.8 ... 31.8 38.0 ... 5476.92 Ni i 1.83 -0.890 134.7 110.2 123.4 115.9 108.4 108.8 100.9 6176.82 Ni i 4.09 -0.430 ...... 9.4 7.8 9.2 8.8 6177.25 Ni i 1.83 -3.500 ...... 9.0 5.8 6.2 5.2 6300.31 O i 0.00 -9.760 * 20.8 22.7 11.9 10.7 ... 7.2 10.0 4840.87 Ti i 0.90 -0.450 33.9 ... 37.9 40.1 36.2 35.8 30.2 4913.62 Ti i 1.87 0.216 25.5 22.0 ... 14.4 5.7 15.3 10.9 5014.24 Ti i 0.81 0.910 122.3 122.0 124.6 ...... 5016.16 Ti i 0.85 -0.510 43.2 ... 36.8 45.7 36.3 41.5 37.4 5064.65 Ti i 0.05 -0.930 88.4 78.8 97.6 89.2 75.6 81.8 74.6 5210.39 Ti i 0.05 -0.580 111.1 105.7 98.6 94.2 83.5 84.3 77.6 4798.53 Ti ii 1.08 -2.670 54.6 45.1 ...... 5129.16 Ti ii 1.89 -1.390 79.4 74.1 63.7 74.5 72.1 73.1 68.7 5154.07 Ti ii 1.57 -1.520 80.2 92.2 81.2 77.3 73.2 75.3 66.6 5226.55 Ti ii 1.57 -1.000 109.7 91.4 103.6 106.2 101.1 104.5 98.2 5381.01 Ti ii 1.57 -1.780 78.0 67.5 65.4 64.2 56.8 62.5 54.5 5418.77 Ti ii 1.58 -2.110 50.9 54.8 42.4 54.5 52.4 51.8 47.9 4883.69 Y ii 1.08 0.070 58.9 55.0 56.3 53.1 58.9 64.9 47.1 4900.11 Y ii 1.03 -0.090 ...... 64.1 77.2 81.4 53.8 5087.43 Y ii 1.08 -0.170 47.0 30.2 42.0 36.3 39.8 47.6 34.9 5200.42 Y ii 0.99 -0.570 ...... 36.1 23.4 ...... 20.0 4810.54 Zn i 4.08 -0.170 * 36.9 34.8 36.6 30.2 33.2 30.9 26.4 5.A: Large tables 103

Table 5.A2: Fornax Globular Clusters elemental ratios

Cl1-D56 σ Nlines Cl1-D68 σ Nlines Cl1-D164 σ Nlines [Ba ii/Fe i] -0.13 0.06 4 0.06 0.09 4 0.07 0.07 4 [Ca i/Fe i] 0.27 0.09 5 0.18 0.07 5 0.09 0.05 4 [Cr i/Fe i] -0.36 0.26 2 -0.35 0.12 2 -0.43 0.16 2 [Eu ii/Fe i] ...... 0 <1.04 0.00 1 <0.89 0.00 1 [Fe i/H] -2.40 0.03 40 -2.55 0.03 39 -2.59 0.03 45 [Fe ii/H] -2.43 0.06 4 -2.62 0.11 5 -2.56 0.06 7 [La ii/Fe i] ...... 0 ...... 0 ...... 0 [Mg i/Fe i] 0.52 0.12 3 0.38 0.07 3 0.08 0.07 3 [Mn i/Fe i] ...... 0 ...... 0 ...... 0 [Na i/Fe i] -0.10 0.12 2 -0.15 0.08 2 0.42 0.12 2 [Nd ii/Fe i] 0.60 0.20 1 0.40 0.07 2 0.49 0.10 1 [Ni i/Fe i] ...... 0 0.27 0.23 1 -0.20 0.21 1 [O i/Fe i] <0.68 0.00 1 0.28 0.10 1 0.37 0.10 1 [Ti i/Fe i] 0.07 0.12 5 -0.22 0.09 5 -0.09 0.14 5 [Ti ii/Fe i] 0.06 0.10 5 -0.02 0.06 6 0.18 0.08 5 [Y ii/Fe i] -0.21 0.28 2 -0.29 0.14 3 -0.33 0.26 2 [Zn i/Fe i] <-0.10 0.00 1 <0.45 0.00 1 <0.29 0.00 1

Cl2-B71 σ Nlines Cl2-B77 σ Nlines Cl2-B226 σ Nlines [Ba ii/Fe i] -0.19 0.10 4 -0.12 0.10 3 -0.38 0.10 3 [Ca i/Fe i] 0.27 0.07 7 0.16 0.05 6 0.21 0.04 9 [Cr i/Fe i] -0.38 0.19 2 -0.10 0.16 2 -0.03 0.22 2 [Eu ii/Fe i] <0.63 0.00 1 <0.88 0.00 1 <0.60 0.00 1 [Fe i/H] -2.14 0.03 50 -2.09 0.04 47 -2.01 0.02 47 [Fe ii/H] -2.06 0.06 6 -2.03 0.07 6 -2.01 0.04 8 [La ii/Fe i] ...... 0 ...... 0 ...... 0 [Mg i/Fe i] 0.53 0.08 3 0.43 0.08 3 0.28 0.07 2 [Mn i/Fe i] ...... 0 ...... 0 -0.28 0.07 2 [Na i/Fe i] -0.08 0.09 2 -0.25 0.12 2 ...... 0 [Nd ii/Fe i] ...... 0 ...... 0 -0.13 0.12 2 [Ni i/Fe i] 0.09 0.29 2 -0.01 0.25 1 0.09 0.26 2 [O i/Fe i] 0.37 0.15 1 0.32 0.20 1 0.49 0.10 1 [Ti i/Fe i] -0.03 0.13 5 -0.13 0.10 4 -0.08 0.06 5 [Ti ii/Fe i] 0.05 0.07 5 0.18 0.09 4 0.16 0.06 5 [Y ii/Fe i] -0.10 0.15 4 -0.44 0.23 2 -0.25 0.18 3 [Zn i/Fe i] -0.11 0.20 1 0.09 0.20 1 0.11 0.20 1

Cl3-B59 σ Nlines Cl3-B61 σ Nlines Cl3-B82 σ Nlines [Ba ii/Fe i] 0.27 0.09 3 0.09 0.10 3 0.27 0.09 2 [Ca i/Fe i] 0.27 0.17 4 0.21 0.03 6 0.23 0.05 5 [Cr i/Fe i] -0.30 0.20 2 -0.28 0.20 2 -0.50 0.15 2 [Eu ii/Fe i] 0.89 0.10 1 0.78 0.15 1 0.97 0.10 1 [Fe i/H] -2.35 0.02 44 -2.42 0.03 50 -2.38 0.03 48 [Fe ii/H] -2.30 0.09 5 -2.31 0.09 4 -2.36 0.06 8 [La ii/Fe i] <0.52 0.04 5 0.95 0.10 1 0.62 0.06 3 [Mg i/Fe i] 0.19 0.09 3 0.37 0.07 3 -0.35 0.08 3 [Mn i/Fe i] ...... 0 0.03 0.07 2 -0.01 0.07 2 [Na i/Fe i] 0.05 0.14 2 -0.25 0.12 2 0.48 0.13 2 [Nd ii/Fe i] 0.65 0.07 2 0.44 0.09 2 0.73 0.07 2 [Ni i/Fe i] 0.22 0.27 1 -0.04 0.24 1 -0.02 0.23 1 [O i/Fe i] 0.43 0.10 1 0.65 0.10 1 0.16 0.10 1 [Ti i/Fe i] 0.02 0.08 6 0.04 0.09 4 -0.17 0.07 5 [Ti ii/Fe i] 0.18 0.05 6 0.30 0.08 6 0.05 0.03 5 [Y ii/Fe i] -0.23 0.30 2 -0.19 0.19 2 -0.24 0.18 3 [Zn i/Fe i] <0.15 0.00 1 <0.22 0.00 1 0.18 0.10 1 104 chapter 5: HR spectroscopy in Fornax Globular Clusters

Table 5.A3: M15 elemental ratios

M15S1 σ Nlines M15S3 σ Nlines [Ba ii/Fe i] -0.15 0.03 4 0.29 0.06 4 [Ca i/Fe i] 0.32 0.03 7 0.37 0.03 7 [Cr i/Fe i] -0.43 0.05 2 -0.33 0.05 2 [Eu ii/Fe i] 0.30 0.20 1 0.65 0.10 1 [Fe i/H] -2.36 0.02 55 -2.41 0.02 52 [Fe ii/H] -2.37 0.03 12 -2.35 0.04 12 [La ii/Fe i] 0.09 0.10 1 0.36 0.07 3 [Mg i/Fe i] 0.33 0.07 3 0.28 0.07 3 [Mn i/Fe i] -0.59 0.05 5 -0.38 0.07 6 [Na i/Fe i] 0.03 0.14 2 0.68 0.11 2 [Nd ii/Fe i] -0.14 0.12 1 0.30 0.04 2 [Ni i/Fe i] 0.07 0.06 3 0.11 0.06 3 [O i/Fe i] 0.04 0.10 1 ...... 0 [Ti i/Fe i] -0.08 0.09 5 -0.06 0.07 5 [Ti ii/Fe i] 0.13 0.05 5 0.16 0.07 5 [Y ii/Fe i] -0.42 0.02 4 -0.20 0.04 3 [Zn i/Fe i] 0.03 0.05 1 0.15 0.09 1

M15S6 σ Nlines M15S7 σ Nlines [Ba ii/Fe i] 0.39 0.04 4 -0.20 0.06 4 [Ca i/Fe i] 0.33 0.03 6 0.41 0.03 7 [Cr i/Fe i] -0.39 0.06 2 -0.39 0.04 2 [Eu ii/Fe i] 0.81 0.10 1 0.41 0.10 1 [Fe i/H] -2.32 0.03 56 -2.47 0.04 49 [Fe ii/H] -2.35 0.06 13 -2.50 0.04 11 [La ii/Fe i] 0.37 0.12 2 0.09 0.18 2 [Mg i/Fe i] -0.15 0.07 3 0.43 0.07 3 [Mn i/Fe i] -0.41 0.06 5 -0.45 0.20 1 [Na i/Fe i] 0.59 0.14 2 0.08 0.12 2 [Nd ii/Fe i] 0.28 0.03 2 -0.23 0.16 1 [Ni i/Fe i] 0.06 0.06 3 0.00 0.06 3 [O i/Fe i] -0.05 0.10 1 0.15 0.10 1 [Ti i/Fe i] -0.09 0.07 6 -0.04 0.11 4 [Ti ii/Fe i] 0.14 0.05 5 0.07 0.07 5 [Y ii/Fe i] -0.15 0.04 3 -0.37 0.04 4 [Zn i/Fe i] 0.02 0.10 1 0.04 0.08 1 Chapter 6 A high resolution spectroscopic study of Fornax Field Stars

paper in preparation∗

B. Letarte, V. Hill, E. Tolstoy, and DART

ABSTRACT– In this chapter, we present the results of our high resolution abundance analysis of 81 individual stars in the central region of Fornax. Using the FLAMES/GIRAFFE spectrograph on the VLT, we obtained high resolution (R ∼20 000) spectra for 81 Red Giant Branch stars in the central 250 of the Fornax dSph (see chapter 4 and Table 4.3 for a description of the observations). Chapters 3 and 4 describe the methods we used to determine abundances, and in this chapter we present the sample selection and discuss the abundances that we have obtained, including α-elements (Mg and Ca, Si, Ti, O), iron-peak (Fe, Ni and Cr) and heavy (Y, Ba, Eu, La and Nd) elements. This is consistent with the fact that we randomly selected our sample from the RGB and that the more metal rich stars are centrally concentrated (e.g. Battaglia et al. 2006). We compare our results with Milky Way (MW) studies, and to recent VLT/UVES abundance determinations of nine individual stars in Fornax globular clusters (chapter 5 and Letarte et al. 2006) and to the Sculptor dSph (Hill et al. in prep). Fornax stars are found to have unusually low α-elements ratios, as well as Ni and Na abundances. The role of metal poor AGB in the creation of s-process elements is clearly seen by our high [Ba/Y] compared to the Milky Way.

∗ Based on FLAMES observations collected at the European Southern Observatory, proposal number 171.B-0588 106 chapter 6: HR spectroscopic study of Fornax Field Stars

Figure 6.1: The spatial distribution of the HR spectroscopic targets in the central region of Fornax. The squares are stars for which we have successfully derived abundances and circles for those that were rejected during our analysis. The ellipses represent Rc, the core radius and 2 Rc (Battaglia et al. 2006).

6.1 Sample selection

s first illustrated in Figure 1.3 in chapter 1, our FLAMES HR survey covers the A central 250 region, see Figure 6.1. The radial velocities of our targets are presented in chapter 4, Table 4.4, and the coordinates and photometrically determined Teff of individual stars are presented in Table 4.A3. Our 107 spectroscopic targets are represented by squares and circles; the squares for stars for which we obtained reliable abundances and the circles for those that were rejected during our analysis. Stars were rejected for various reasons: their Vrad suggested that they are not members of Fornax (1 star); velocity offsets in spectra taken at different times making it impossible to stack the different exposures (possibly these are binary stars, 8 stars); or the [Fe/H] coming from different lines were too scattered to perform a conclusive analysis because they are not RGB stars (possibly foreground dwarfs) and/or have a too low signal to noise (17 stars).

Figure 6.2 shows a Colour Magnitude Diagram of a circular region of 100 radius in the centre of Fornax, matching the FLAMES ﬁeld of view, on which we identify our targets, including the rejected ones. As can be seen, we selected our stars to include the entire RGB colour range, going as far to the blue and red side as possible and thus (hopefully) the entire age and metallicity range. We cross-correlated our potential target list with known carbon stars∗ in Fornax so we could minimise the number of AGB stars that we would observe. Still, we were expecting that the bluest and reddest stars might be foreground dwarf or AGB stars but we are conﬁdent we didn’t exclude extremely metal poor stars through our sample selection and subsequent analysis. This is supported by our single metal poor star BL085 at [Fe/H] = -2.58 which was analysed without any problems.

∗ We are grateful to Serge Demers for providing us with his list 6.2: Results 107

Figure 6.2: A CMD of the FLAMES ﬁeld of view in the centre of Fornax coming from our WFI data. Our spectroscopic targets are represented by squares for stars for which we were able to determine abundances and circles for those that were rejected during our analysis.

6.2 Results

In this section, we present the abundance ratios for the 81 stars which survived the selection process described in section 6.1 The stellar parameters we assigned to each star, following the method presented in chapter 4 can be found in Table 6.A1. The line list used, along with the measured EW s for each star (for every element) can be found in Table 6.A2 and all the abundance ratios used throughout this chapter are listed in Table 6.A3.

6.2.1 Iron abundance One of the most straightforward results is an accurate determination of the iron abundance, [Fe/H], which in our case is derived from ∼40+ lines per star. We present our [Fe/H] distribution in Figure 6.3 and compare it to the distribution of Battaglia et al. (2006), where they used low resolution (R = 6500) Ca II triplet measurements to determine [Fe/H]. Our distribution peaks sharply at [Fe/H] ' -0.8 and is clearly skewed towards more metal rich stars. There are a few stars outside the main distribution (on the metal poor side) but the centre of Fornax is clearly dominated by stars with [Fe/H] > -1. However, we still sample two orders of magnitude in [Fe/H], from -2.5 . [Fe/H] . -0.5.

As seen in Figure 1.3 (chapter 1), our FLAMES ﬁeld included Fornax globular cluster 4 but there is no sign that our ﬁeld sample includes any star from GC 4. This can be seen from the spatial distribution and abundance measurements. The single, really metal poor star we have in our sample ([Fe/H] ' -2.5) is the only star to overlap with Fornax 108 chapter 6: HR spectroscopic study of Fornax Field Stars

Figure 6.3: The [Fe/H] distribution of our Fornax sample (solid line) of stars, with a peak at [Fe/H] ≈ -0.8, richer than the sample of Battaglia et al. (2006) (dashed line).

GCs metallicity range and it is not spatially close to GC 4. Therefore, it seems likely that it is a metal poor ﬁeld star, the only representative of Fornax oldest population in our ﬁeld star sample.

6.2.2 Alpha Elements The evolution of chemical abundances in a galaxy is linked to its star formation history, SFH, (e.g. Tinsley 1979; Pagel 1997) and the [α/Fe] ratio is a useful tool to study this evolution. Alpha elements, which include Calcium (Ca), Magnesium (Mg), Titanium (Ti) Oxygen (O) and Silicon (Si) are predominantly produced by high mass, (> 8 M ) short lifetime Type II supernovae explosions (SNe II). The ratio [α/Fe] is a way of tracing the relative contribution of SN II and SN Ia products that were available when the star formed. Lower mass SNe II (8-12 M ) result in lower yields than more massive SNe II (Woosley & Weaver 1995). The stars that form after the ISM has been enriched by a SNe II event should have an enhanced [α/Fe] while those that form after the SNe Ia have started to contribute signiﬁcant amounts of Fe to the ISM should have a lower [α/Fe].

The [Fe/H] at which the [α/Fe] ratio starts decreasing (the “knee”) in a galaxy depends on several factors: the SFH; the initial mass function (IMF); the time it takes for the first SN Ia to explode and the time it takes for the mixing of SNe Ia and SNe II products back into the ISM (e.g. Matteucci 2003). The arrival of the first SN Ia should be constant in age (commonly believed to happen after ∼1 Gyr) for all galaxies since it’s the result of the binary interaction between two evolved stars. The higher in [Fe/H] this “knee” occurs, the more efficient the system was at enriching its gas before the arrival of the first SNe Ia. A plateau is reached when (presumably) there is balance between the contribution of SNe II and SNe Ia. 6.2: Results 109

Figure 6.4: [α/Fe] as a function of [Fe/H]. We computed α as the average of Mg, Ca and Ti i abundances. The Milky Way points were taken from the compilation of Venn et al. (2004), with diﬀerent symbols for thin disk (dark grey squares), thick disk (light grey triangles), halo (empty diamonds) stars and our Fornax Field star results (black circles). In subsequent plots in this chapter, there will be no distinction between the MW components, for clarity.

In the Milky Way, α-elements are typically overabundant in GCs and metal poor halo stars relative to disk stars. To obtain an average measure of α-element abundance, the average of Mg, Ca and Ti is taken and is shown in Figure 6.4, where we compare MW halo, thin and thick disk stars to Fornax Field stars. It is obvious from Figure 6.4 that our [α/Fe] are significantly underabundant compared to the MW at the same [Fe/H], a sign that the evolutionary history of is Fornax is significantly different from the MW. Note that the Milky Way stars in the [Fe/H] range of Fornax are a mix of halo, thin and thick disk stars, a population possibly not as uniform as what we might expect from the centre of a dwarf galaxy like Fornax. A simple comparison in [Fe/H] of Fornax versus the MW does not take into account the different evolutionary processes in the MW different components, but is it obvious that Fornax does not overlap with any part of the MW.

Individual [α/Fe] ratios are presented in Figure 6.5 ([Mg/Fe], [Si/Fe] and [Ca/Fe]) and Figure 6.6 ([Ti i/Fe] and [Ti ii/Fe]). Along with our Fornax GC and MW points, we introduce the eight peculiar∗ halo stars observed by Nissen & Schuster (1997) (NS97). These stars were found to display low [α/Fe] and [Ni/Fe] (along with other chemical peculiarities) when compared to normal MW halo stars, and this has been suggested that they might have been accreted by the MW from a dwarf galaxy.

∗ These stars have unusual kinematics and orbital parameters: large maximum distance from the Galactic centre (Rmax) and large distance from the Galactic plane (zmax). 110 chapter 6: HR spectroscopic study of Fornax Field Stars

Figure 6.5: [Mg/Fe], [Si/Fe] and [Ca/Fe] as a function of [Fe/H]. The Fornax ﬁeld stars are plotted as solid circles, the galactic stars of Venn et al. (2004) as small grey squares, the Fornax globular clusters of chapter 5 as triangles and the eight peculiar halo stars from Nissen & Schuster (1997) as empty squares. There is a representative (average) error bar for the Fornax ﬁeld star abundances in the bottom right corner of each panel. This is the quadratic sum of [element/H] + [Fe/H], (measurement errors) taken from Table 6.A3. The same axis scale will be used for every [element/Fe] vs [Fe/H] plots presented in this chapter, for easy comparison. 6.2: Results 111

Figure 6.6: [Ti i/Fe] and ([Ti ii/Fe] as a function of [Fe/H]. Symbols are deﬁned in Figure 6.5.

It can be seen from Figure 6.5 that globally, the metal poor star and the GCs of Fornax show a typical overabundance in α-elements but as the [Fe/H] increases, [α/Fe] decreases, reaching a signiﬁcant underabundance for the highest metallicities. The [Mg/Fe] are typically close to zero, lower than Galactic stars by at least 0.2 to 0.5 dex and the highest values just overlap NS97. The [Ca/Fe] ratios are globally lower than other α-elements, especially for the handful of stars close to [Fe/H] = -1.5. At higher metallicity ([Fe/H] & -0.7), our [Ca/Fe] ratios seems to increase, as if the balance in SN II/SN Ia has been broken and suddenly there was less contribution from SN II, which seems unlikely, since Ca is the only element for which we notice such behaviour. The [Si/Fe] ratios are the highest for the α-elements (if we do not consider Ti ii, which will be discussed later), with a mean value close to zero. The [Si/Fe] in Fornax agree with the eight peculiar halo stars of NS97, the only α-element for which this is clearly the case.

In our Teff range, (. 4400 K) atoms of Ti and O can be trapped in molecules, like TiO and CO, artificially lowering our derived abundances. This is not a problem in hotter atmospheres, where the the molecules will always be broken up into their atomic form. We have tested the influence of molecules on atomic abundances by using a line forma- 112 chapter 6: HR spectroscopic study of Fornax Field Stars

Figure 6.7: Iron peak elements in Fornax, [Ni/Fe] and [Cr/Fe] as a function of [Fe/H]. Symbols are defined in Figure 6.5. tion code that takes molecules into account∗ to derive our abundances and compare these with our standard analysis. We chose stars to cover the Teff range of the Fornax field stars.

We observed a significant difference (∼0.3 dex) in our O/H ratios, confirming that at this Teff , there is a significant fraction of oxygen locked in the form of CO. The O abundance is not only affected by molecules, but it also has a problem with the lines used for this analysis. They are two O lines in our wavelength range, the forbidden line at 6300 Å and another one at 6363 Å. The more reliable line at 6300 Åis not suitable for Fornax stars as the typical Vrad means this matches a telluric absorption line, rendering it unusable for most of our stars. The second one is a very weak line, falling in a region that is affected by a Ca i auto-ionisation broad feature at 6362Å, and would deserve much more attention than could be awarded in the course of this thesis, to provide an accurate O abundance indicator.

∗ TURBOSPECTRUM (Alvarez & Plez 1998), used in the plane-parallel approximation 6.2: Results 113

In contrast to oxygen, we found that Ti abundances are typically not affected by the amount of Ti locked in the form of TiO. However, it can be seen from Figure 6.6 that [Ti i/Fe] is quite different from [Ti ii/Fe], a sign that there are significant non-LTE effects, affecting more Ti i abundances (since most of the Ti i lines have a low χex) than Ti ii. However, our Ti i abundances are statistically more reliable than Ti ii, since Ti i was calculated using on average ∼8-9 lines compared to only ∼2-3 lines for Ti ii. This is reflected in the much larger error bars on Ti ii. If we consider only Ti i, it’s behaviour is similar to a typical α-element like Ca of Mg.

6.2.3 Iron peak elements According to nucleosynthetic predictions, iron peak elements like Iron (Fe), Chromium (Cr), and Nickel (Ni) are believed to be formed predominantly from explosive nucleosynthesis, in SN Ia (Iwamoto et al. 1999; Travaglio et al. 2005). In Figure 6.7, we present the [Ni/Fe] and [Cr/Fe] abundance ratios for Fornax stars as a function of [Fe/H]. Both ratios behave in more or less the same way, but Ni has much smaller error bars due to the larger number of lines available; we have ∼15 lines of Ni per star, compared to only 1 line of Cr. [Cr/Fe] seems to be more scattered than [Ni/Fe] but consistent with the larger error bars.

Our metal poor star (BL085, [Fe/H] = -2.58) has a signiﬁcantly higher [Ni/Fe] than the other Fornax ﬁeld stars, comparable to MW halo stars and Fornax GC stars. At this low metallicity, we do not detect as many Ni lines as for the more metal rich stars, only 4 instead of ∼15 lines, giving it a much larger error bar than the average shown in the bottom corner of the plot as can be seen in Table 6.A3.

For [Cr/Fe], we performed the abundance determination on an Arcturus spectrum, known to have [Cr/Fe] ' 0.0, similar to other MW stars with [Fe/H] ' -0.5, and with the same line, we obtain [Cr/Fe] = -0.2. The same analysis was done for the Ni lines and we obtained the expected [Ni/Fe] = 0.0 for Arcturus. This is a sign that the Cr line could have an erroneous log gf, leading to lower a abundance. We therefore expect that the true [Cr/Fe] could be some ∼0.2 higher than what is currently displayed in Figure 6.7, with an average closer to the [Ni/Fe] value.

At [Fe/H] . -1.5, [Ni/Fe] appears to be Galactic halo-like, while near [Fe/H] ' -1.0, it decreases. This [Ni/Fe] underabundance has also been observed in the eight peculiar halo stars of Nissen & Schuster (1997) (but not for [Cr/Fe]). Similar to the NS97 stars, our [Ni/Fe] underabundance in Fornax is also accompanied by a moderate decrease in [Na/Fe] and [α/Fe] in this metallicity range. These low values of [Ni/Fe] and [Cr/Fe] cannot be easily explained with our current understanding of nucleosynthesis. The [Ni/Fe] ratios should be zero and constant for all [Fe/H] since the two elements are believed to be predominantly created in the same production site, SN Ia (Travaglio et al. 2005). To witness this diﬀerent behaviour is an indication that the production factors for each iron-peak element are not the same and depend on the evolutionary history of the parent population. Maybe the SNe Ia Ni yields are linearly dependant on the original metallicity of the white dwarf progenitor, as suggested by Timmes et al. (2003), or some elements (like Ni but not Fe) are more aﬀected by winds, causing preferential metal loss in the ISM. This result is surprising and will need to be investigated further, as Fornax is not 114 chapter 6: HR spectroscopic study of Fornax Field Stars

Figure 6.8: [Mg/Fe] plotted against [Na/Fe]. Unlike two of the Fornax GC stars, there is no sign deep mixing pattern in the ﬁeld. the only galaxy in which this behaviour is observed. This under-abundance patterns in [Ni/Fe] and [Cr/Fe] have been observed in the Large Magellanic Cloud (LMC) disk stars by Pompeia et al. (2006).

This under-abundance of Ni and Cr was not observed by Shetrone et al. (2003), with which we have three stars in common. We attribute this difference to a combination of systematic effects, where the most important one is the use of a different line list. See section 4.4.1 and Figure 4.16 for more detail on the systematics in abundance determination.

6.2.4 Deep-mixing pattern The so-called deep mixing pattern is believed to be caused by self pollution in a star that modiﬁes the upper atmosphere abundances. Proton-capture nucleosynthesis can convert O, N and Ne to Na, and Mg to Al in the H fusion layer of evolved RGB stars. A deep-mixing pattern can thus cause a decrease in O associated to an increase in Na and a decrease in Mg with an increase in Al. This has only been observed in stars belonging to globular clusters, in a range of diﬀerent galaxies, like the Milky Way, the LMC and Fornax (see chapter 5). 6.2: Results 115

Figure 6.9: The Na-Ni relationship. Symbols are deﬁned in Figure 6.5.

In the Fornax field stars, we do not find evidence of deep mixing pattern, see Fig- ure 6.8. Unfortunately, our FLAMES observation wavelength coverage (usable) stops at λ ' 6690 Å, just missing two Al lines at 6696.03 Å and 6698.67 Å. It is therefore not possible to say anything about Al. Also, as mentioned in section 6.2.2, we do not yet have a reliable O abundance for our Fornax stars. However, it is clear from Figure 6.8 that there is no enhanced Na. All [Na/Fe] values show a significant underabundance, much lower than MW stars of similar metallicity, going as low as [Na/Fe] = -1.0 in the extreme cases. There are only two outliers that have higher [Na/Fe] but these are compatible with normal MW stars, showing no hints of deep mixing. There are some low [Mg/Fe] values in Fornax, but without a high [Na/Fe], this is not deep-mixing. In Fornax only the GC stars show deep mixing pattern.

6.2.5 The Na-Ni relationship Where does Ni come from? As mentioned in section 6.2.3, Ni comes predominantly from SN Ia, but at earlier times, before the ﬁrst SN Ia start to enrich the ISM, it is linked to the Na production in SNe II. As explained in section 5.6 of Clayton (1983), a correlation between Na and Ni is a natural result of nucleosynthesis in massive stars. Timmes et al. (1995) suggest that Na is delivered to the ISM when massive stars explode as SNe II and the amount of Na produced is controlled by the neutron excess, where 23Na is the only stable neutron-rich isotope produced in signiﬁcant quantity during C and O burning stage. 116 chapter 6: HR spectroscopic study of Fornax Field Stars

During the SN II event, the elements are photodissociated to protons and neutrons, which will recombine to form 56Ni, which β decays to 56Fe, the dominant isotope of iron. 54Fe and/or 58Ni can also be produced at this stage, depending on the abundance of the neutron-rich elements (e.g. 23Na). The amount of 54Fe made is small compared with the total yield of iron (dominated by the 56Fe production), but this is the main source for 58Ni, the stable isotope of nickel. In summary, the Ni production depends on the neutron excess during the photodissociation of the core during the SN II event, and the neutron excess will depend primarily on the amount of 23Na produced earlier. So the Na-Ni relationship is expected when SN II enrichment dominates. The arrival of SN Ia can break (or ﬂatten) this relationship, as Ni is produced without Na in the standard model of SNe Ia (Tsujimoto et al. 1995). But the Na yields are still a matter of discussion, because SN Ia involve binary star interactions, and the outcome is dependent on the accreted star. So it is hard to accurately predict the SNe Ia contribution to the Na-Ni correlation.

This has two implications of interest to us: i) In SNe II, the dominant source of Ni is independent of the dominant source of Fe, possibly allowing an underabundance in [Ni/Fe] at low metallicities, before the SN Ia start to enrich the ISM. ii) If the ﬁrst SNe II were not neutron rich, then there will be a lack of Na (therefore also Ni) created. The next generation of stars that will be formed could carry this signature, even at higher metallicity. Our low Ni at high metallicity is possibly the consequence of low Na in earlier generations, caused by neutron-poor SNe II.

For Galactic stars, it has been observed that stars with low [α/Fe] also have low [Na/Fe] and low [Ni/Fe]. Nissen & Schuster (1997) first found this behaviour for eight peculiar halo stars. They suggested that these stars might have been accreted from a nearby dwarf galaxy (their abundance patterns show that they are chemically different from the MW). Other studies, by Fulbright (2002), suggested something similar for stars at large galactocentric distances. Stephens & Boesgaard (2002) suggest that there might be a gradient (0.1 dex over 10 kpc) in our galaxy, reflecting the different local conditions where the stars form.

If we look at Figure 6.9, our extremely low [Na/Fe] values go beyond the previous gradient (to lower [Na/Fe]) observed by Venn et al. (2004) (their Figure 5). As they have done, we have restricted the iron abundance of the stars plotted to -1.5 < [Fe/H] < -0.5, where the Na-Ni relation can be seen. This suggests that the chemical evolution of this metal rich sub-sample of Fornax stars is relatively similar to the MW, at least in how Na is linked to Ni in this [Fe/H] range, with the NS97 points ﬁlling the gap between Fornax and the MW.

6.2.6 Heavy elements Heavy elements are those with atomic number Z > 30, like Yttrium (Y), Barium (Ba), Europium (Eu), Lanthanum (La) and Neodymium (Nd). They are neutron capture elements built from elements that are exposed to the high neutron ﬂux. Iron peak elements, like 56Fe, are the most eﬃcient seeds to capture neutrons to create heavier elements. There are two main types of neutron capture, the s-process (or slow process) 6.2: Results 117

Figure 6.10: Heavy elements: [Ba/Fe], [Y/Fe] and [La/Fe] as a function of [Fe/H]. Symbols are deﬁned in Figure 6.5. 118 chapter 6: HR spectroscopic study of Fornax Field Stars

Figure 6.11: Heavy elements in Fornax: [Nd/Fe] and [Eu/Fe] as a function of [Fe/H]. Symbols are deﬁned in Figure 6.5. The [Nd/Fe] points are from Burris et al. (2000). and the r-process (rapid process). When a seed accumulates neutrons and leads to the production of a β-unstable nucleus, like 59Fe, the outcome will depend on neutron-capture timescales. In the s-process, most of the unstable nuclei will have time to undergo β- decay before capturing other neutrons and building up heavier elements (Pagel 1997). The main contribution of the s-process is believed to be from thermal pulses in 2-4 M AGB (cool giants) stars (Truran 1981). The r-process by contrast usually involves more extreme conditions, very high temperature and neutron densities, possibly found in low mass (8-12 M ) SN II explosions.

Tracing the s-process and r-process contribution We present the abundance ratios of [Ba/Fe], [Y/Fe], [La/Fe], [Nd/Fe] and [Eu/Fe] as a function of [Fe/H] in Figures 6.10 and 6.11 in decreasing order of s-process contribution. Ba is the element with the larger s-process contribution (for the main∗ s-process) in the Sun, with a fraction of 88% s-process (Kappeler et al. 1989). Then, Y has an s-process

∗ Contribution from the weak s-process (helium burning in massive star) is 4% for Y, 1% for Ba, La, Nd and 0% for Eu. 6.2: Results 119

Figure 6.12: [Ba/Y] as a function of [Fe/H]. Symbols are deﬁned in Figure 6.5. contribution of 85% (Raiteri et al. 1992), The La s-process fraction is 75%, Nd 46% (Kappeler et al. 1989) and only . 5% for Eu, as it is 95% r-process (Burris et al. 2000). Their relative abundance is thus a good discriminant of the dominant neutron capture processes in Fornax stars.

According to nucleosynthesis calculations based on hydrodynamical simulations, the s-process does not occur before [Fe/H] & -2.0 and is not significant before [Fe/H] ' -1.0 (Travaglio et al. 1999, 2004). This is consistent with what we observe in Fornax, where [Ba/Fe] increases significantly only at [Fe/H] & -1.0. This is different than what we observe for [Y/Fe], also a s-process-dominated element. From Figure 6.10, we see that the [Y/Fe] ratio distribution is flat, scattered around zero in almost the same way as MW stars. This means that the s-process-enrichment does not uniformly contribute to the creation of different s-process-elements which, in the Sun, have the same s-process contribution. The behaviour of [La/Fe] is similar to [Ba/Fe], except that the rise in [La/Fe] with increasing [Fe/H] is not as prominent, probably due to the smaller contribution from s-process to La. The same reasoning goes for [Nd/Fe] in Figure 6.11: with a smaller s-process contribution comes a smaller increase (barely noticeable) with increasing [Fe/H]. And finally, [Eu/Fe] shows no increase at all with [Fe/H], confirming that Eu is r-process dominated. In summary, from Figures 6.10 and 6.11, we observe that the s-process becomes clearly dominant at [Fe/H] & -1.0.

Dominant role of metal poor AGB [Ba/Y] behaves quite diﬀerently in Fornax than in the MW, as can be seen in Figure 6.12, which clearly shows that Fornax favoured the creation of Ba over Y compared to the MW. Both elements are supposed to be s-process dominated from our knowledge of s-process contribution in the Sun. Ba (Z=56) belongs to the 2nd peak in the distribution of neutron magic numbers and Y (Z=39) belongs to the 1st peak. These magic numbers are related to low neutron-capture cross-sections which lead to abundance peaks close to Sr (Z=38, next to Y), Ba and Pb (Z=82). 120 chapter 6: HR spectroscopic study of Fornax Field Stars

Figure 6.13: [α/Eu] as a function of [Fe/H]. Symbols are deﬁned in Figure 6.5.

The high [Ba/Y] in Fornax could be explained in a scenario where Fornax has a larger contribution from metal poor AGB stars compared to the MW. This would mean that the AGB that created the s-process elements were more metal poor in Fornax, favouring the creation of heavier elements, bypassing ﬁrst peak elements (Y) in favour of second peak elements (Ba). A more metal poor environment has fewer nuclei (seeds) to absorb the neutrons available from the AGB envelop so they have to accumulate on fewer targets (creating more high-Z elements). In a more metal rich environment, the neutrons will have more targets on which to distribute, creating more low-Z elements.

6.3 Discussion

In summary, because it reached low [α/Fe] values at much lower [Fe/H] than the MW, Fornax was most likely less eﬃcient in enriching its gas, possibly due to galactic winds. The lack of Fornax stars (in our sample) with -2.5 < [Fe/H] < -1.0 prevents us from making a more conclusive statement, as we cannot trace the decline (the knee) from high to low [α/Fe]. The [α/Fe] deﬁciency at [Fe/H] & -1.0 could be explained as the result of an early burst of star formation in the evolutionary history of Fornax followed by a dormant period with predominantly SN Ia enrichment before the relatively recent star formation episode that formed most of the stars in our sample (1-3 Gyr ago).

The underabundance of [α/Fe] in Fornax with respect to the MW could be attributed SNe Ia starting to contribute Fe at lower metallicities in Fornax, resulting in a lower ratio. Also, perhaps because Fornax is a small galaxy with a low average star formation rate, α-elements were most likely only produced by low-mass SNe II (8-12 M ) which result in lower yields than more massive SNe II (Woosley & Weaver 1995). This suggest an “eﬀectively” truncated IMF. This is plausible since intuitively, a small system like Fornax is less likely to form giant molecular clouds (that are believed to be required to obtain massive stars) than a much larger system like the MW. 6.3: Discussion 121

Figure 6.14: [α/Fe] ratio as a function of [Fe/H], including points for the Sculptor dSph (Hill et al. in prep.).

It has been argued that the r-process production occurs in low mass SNe II (Mathews et al. 1992). The high [Eu/Fe] ratios in Fornax (reaching slightly higher than the MW) suggest an important contribution of these low mass SNe II in the enrichment of Fornax. This is compatible with low [α/Fe], for which we also need low mass SNe II. When we plot [Eu/α] ratios (Figure 6.13), it is clear that the sites and relative contribution of α- and r-process-elements creation differ in Fornax and the (average) MW, where the Fornax values are ' 0.7 dex lower than the MW. If both r-process and α−elements were created in the same way in every galaxy, the ratio of [Eu/α], should be the same in every system. But we know this is not the case, as shown in chapter 5, the stars of Fornax cluster 3 are r-process enhanced, (Eu-rich) something not common but still seen in many M 15 stars (Sneden et al. 1997). The reason why all Fornax stars have low [α/Eu] is not the same for the field stars and the GC. The stars of Cluster 3 are Eu-rich and the field stars are α-low.

6.3.1 Comparison of Fornax and Sculptor In Figure 6.14 we compare the average α-element abundance in Fornax with those found in a similar FLAMES study of the Sculptor dSph (Hill et al., in prep). Sculptor is a faint dSph galaxy (LFnx ' 7 × LScl) dominated by old (10-12 Gyr) stars. Using the combined data set we clearly see a decline in [α/Fe] as SN Ia become more important. It appears that the stars observed in Fornax (except for the single metal poor star) are all lying in the “plateau” where a balance has been achieved between SN Ia and SN II element production. To understand if this is a recent minimum or the end product of the entire evolutionary history of Fornax we need to observe more stars in Fornax in the metallicity range -2.0 < [Fe/H] < -1.0. This would also allow us to make a more detailed comparison between the metal enrichment history of Scl and Fnx, which are two galaxies with very diﬀerent star formation histories (cf. Tolstoy et al. 2003). 122 chapter 6: HR spectroscopic study of Fornax Field Stars

Figure 6.15: [Ba/Eu] ratio as a function of [Fe/H], including points for the Sculptor dSph (Hill et al. in prep.).

In Figure 6.15, we compare [Ba/Eu] in Fornax to the same Sculptor results. Using [Ba/Eu] we track the slow increase in the s-process contribution with increasing [Fe/H]. As the stellar population becomes more metal rich there is a steady rise in the Ba abundance. In Figure 6.15 we also show the Galactic measurements, and it can be seen that the s-process is a much stronger contribution to chemical evolution of Fornax than it is to the MW, with a startling divergence at [Fe/H] ' -1.0. This suggests that stellar winds (e.g., from AGB stars) have played a uniquely important role in the (recent, 2- 4 Gyr ago) enrichment history of Fornax. It is also clear that Sculptor with it’s much shorter star formation history never reached this stage where stellar winds from evolved stars had a strong eﬀect on the enrichment of the ISM.

6.3.2 Age and [Fe/H] In Figure 6.16 we show a Colour-Magnitude diagram where we have colour-coded the stars depending upon their spectroscopically determined [Fe/H]. From this plot we can see that the metallicity distribution is strongly peaked, with most stars in a relatively small metallicity range which are quite spread out in colour over the RGB. Broadly speaking the high metallicity stars (the darker points) are redder than the low metallicity stars (lighter points). However, it is clear that there are exceptions, and that it can be risky to use the RGB to determine the metallicity of a complex stellar population.

Using isochrones of the appropriate metallicity we can determine ages for each star (see Battaglia et al. 2006, for details). In Figure 6.17 we plot the ages and metallicities for both the HR (this work) and the LR (Battaglia et al.) [Fe/H] measurements of Fornax ﬁeld stars, including the age-metallicity relation (dashed line) determined by Battaglia et al.. We compare our observations to the age-metallicity relation of Gallart et al. (2005), which is determined from a detailed Colour-Magnitude Diagram analysis. It is not easy to make strong conclusions on the basis of this comparison, as the spectroscopic samples consist of only RGB stars (which are thus always more than 1 Gyr old), whereas 6.4: Conclusions 123

Figure 6.16: (left): Colour-Magnitude Diagram of the central (100 radius) region of Fornax, with a box around the area where we selected our targets. (right): Close-up of the CMD where our targets are colour coded (grey-scale) according to their metallicity in the following way. Black for stars of -0.5 > [Fe/H] > -0.7, dark grey for -0.7 > [Fe/H] > -0.9, light grey for -0.9 > [Fe/H] > -1.1 and white for -1.1 > [Fe/H] > -2.6. the CMD analysis also includes stars as young as ∼200 Myr old. However, it is clear that, although there is a large spread in [Fe/H] especially at relatively young ages, the spectroscopic measurements suggest that the majority of 1.5-2 Gyr old stars in Fornax are more metal rich than the chemical evolutionary history from CMD analysis suggests they should be. This discrepancy needs further detailed investigation to understand what is going on. It might be that this highlights a very complex metallicity distribution in the younger stars in Fornax, possibly as the result of a recent merger with a galaxy having quite a diﬀerent (higher) average metallicity. This would likely confuse the determination of a chemical evolution history from CMD analysis.

6.4 Conclusions

The young metal poor stars of Fornax have low [α/Fe] ratios associated to a sub-zero [Ni/Fe]. The [α/Fe] dependence on [Fe/H] is different from the Milky Way, showing a different efficiency in gas enrichment. Fornax is dominated by s−process at high metallicity, showing the strong role of (metal poor) AGB in its evolution. The oldest most metal poor field star in Fornax is almost indistinguishable from the Galactic halo stars at the same [Fe/H], except in the case of [La/Fe] which may indicate a problem in our La measurements rather than anything more fundamental. We expect that the hyperfine structure correction (which we have not applied yet) for La will not lower it enough to be compatible with MW stars. There is also a near perfect agreement between the abundance ratios observed in individual stars in Galactic globular clusters and in the Fornax globular clusters. However, observations of more low metallicity stars will be needed to better understand the full chemical evolution history of Fornax dSph. 124 chapter 6: HR spectroscopic study of Fornax Field Stars

Figure 6.17: Age and metallicity (Z = 10[Fe/H] × 0.02) of the 81 stars from the HR sample (diamonds), and the LR sample of Battaglia et al. (2006) (empty squares) with the average value for each age bin (dashed line). The age-metallicity relation (ﬁlled gray section) was taken from Gallart et al. (2005),

An interesting next step is to link the stellar abundances in the nearby dwarf galaxies, especially the old, low metallicity stars to studies carried out at high redshift. For example the damped Lyα systems (DLAs) give us similar information about chemical abundances of the gas in distant systems. These objects are potentially the precursors of what we observe in the Local Group today.

Further work on Fornax will be to investigate the different regions of this surprisingly complex dwarf galaxy in more detail. We would benefit from obtaining more resolution high abundances of individual stars. Although 81 stars is a dramatic improvement on the previous high resolution study (of 3 field stars) it also leaves several intriguing questions unanswered. Specifically we would like to study more metal poor field stars in this galaxy in both the central region and the outskirts. It would also be interesting to continue our detailed abundance studies of the globular clusters, obtaining high resolution spectra for stars in GC 4 & 5. Another intriguing open question is the nature of the “shell-like” structures found in and around Fornax, and high resolution abundances of individual stars may shed some light on their origins.

Appendix 6.A Large tables 6.A: Large tables 125 min EW t v used for each star. [Fe/H] EW g log eff T Star BL218 3939BL221 0.67 4056BL227 0.81 -0.62 4046 2.0BL228 0.84 -0.86 3992 2.1 16.17 BL229 0.71 -0.87 4014 2.0 13.80 BL233 0.80 -0.88 4048 2.3 15.40 BL239 0.83 -0.71 4123 2.3 13.31 BL242 0.89 -0.68 4063 2.2 15.73 BL247 0.85 -0.88 4032 2.1 14.52 BL250 0.85 -1.04 3944 2.1 13.80 BL253 0.65 -0.82 4003 2.3 12.98 BL257 0.82 -0.67 3994 2.2 18.64 BL258 0.78 -0.66 4030 2.3 19.14 BL260 0.85 -0.58 4009 2.3 14.46 BL261 0.79 -0.56 4046 2.2 16.77 BL262 0.82 -0.85 4130 2.4 16.88 BL266 0.84 -0.79 4212 2.0 14.90 BL267 0.83 -0.78 4201 2.1 16.99 BL269 0.80 -1.44 3990 2.0 15.68 BL278 0.75 -0.72 4072 2.1 12.10 BL279 0.64 -0.81 4272 2.0 15.62 BL295 0.95 -0.72 3980 2.3 13.86 BL300 0.70 -1.52 3990 1.6 15.51 BL304 0.71 -0.70 3950 2.3 13.31 BL311 0.70 -0.92 4027 2.2 17.38 BL315 0.79 -0.89 4139 2.3 15.68 BL323 0.86 -0.78 3881 2.2 14.41 0.66 -0.81 1.8 16.94 -0.88 2.3 16.72 14.46 min EW t v [Fe/H] g log eff T Star BL148 4023BL149 0.72 4100BL150 0.88 -0.62 4025 2.2BL151 0.80 -0.91 4024 2.2 20.62 BL155 0.81 -0.83 4060 2.2 14.30 BL156 0.90 -0.86 4099 2.1 15.29 BL158 0.89 -0.74 4078 2.3 13.37 BL160 0.87 -1.14 4027 2.2 17.21 BL163 0.84 -0.87 4124 2.0 13.48 BL166 0.88 -0.87 4086 2.2 15.79 BL168 0.84 -0.73 4113 2.3 14.30 BL171 0.83 -0.89 4048 2.3 15.62 BL173 0.87 -0.88 3988 2.1 14.24 BL180 0.85 -0.90 4114 2.4 15.40 BL185 0.77 -0.78 4098 2.3 15.79 BL190 0.69 -0.90 3979 2.2 17.77 BL195 0.82 -0.73 4261 2.1 13.09 BL196 0.91 -0.79 4015 2.3 15.84 BL197 0.76 -1.00 3956 2.1 13.37 BL203 0.68 -1.02 4037 2.4 11.99 BL204 0.79 -0.89 4139 2.2 11.93 BL205 0.97 -0.83 4243 2.1 14.19 BL208 0.96 -1.00 4159 2.4 15.07 BL210 0.89 -0.69 4062 2.2 17.54 BL211 0.81 -0.66 4058 2.1 14.13 BL213 0.69 -0.76 4032 2.2 13.59 BL216 0.78 -0.65 3998 2.1 15.62 0.75 -0.86 2.2 14.13 -0.72 2.2 13.53 14.08 min EW t v [Fe/H] g log Stellar parameters of the model used for our sample of 81 stars, along with the minimum eff T Star BL038 3980BL045 0.69 4122BL052 0.85 -0.88 3997 2.2BL065 0.72 -1.09 4330 2.1 14.24 BL076 0.97 -1.02 4065 2.4 12.15 BL077 0.83 -1.43 4026 1.9 14.24 BL079 0.80 -0.86 4036 2.2 11.33 BL081 0.76 -0.79 4062 2.2 13.04 BL084 0.82 -0.56 3968 2.1 12.71 BL085 0.72 -0.62 4291 2.1 17.98 BL091 0.87 -0.82 4162 2.1 14.08 BL092 0.86 -2.58 3961 2.2 12.10 BL096 0.74 -0.96 4010 2.1BL097 0.75 -0.95 9.18 4060 2.0 11.99 BL100 0.82 -0.75 4044 2.1 15.23 BL104 0.84 -0.92 4013 2.3 17.98 BL113 0.77 -0.92 4187 2.1 12.98 BL115 0.83 -0.96 4116 2.2 12.04 BL123 0.79 -0.75 3993 2.2 13.37 BL125 0.71 -1.44 4080 2.0 13.91 BL132 0.79 -0.97 3909 2.3 11.77 BL135 0.65 -0.73 4058 2.1 11.66 BL138 0.83 -0.85 3939 2.1 13.64 BL140 0.71 -0.95 3992 2.2 13.75 BL141 0.75 -1.01 4078 2.3 16.28 BL146 0.84 -0.86 4076 2.1 14.85 BL147 0.84 -0.82 4194 2.1 14.52 0.94 -0.92 2.3 13.64 -1.38 1.8 12.98 12.21 Table 6.A1: 126 chapter 6: HR spectroscopic study of Fornax Field Stars

Table 6.A2: Complete line list with parameters and associated EW s (in mÅ, measured by DAOSPEC) for all the stars. Part 1/5.

Equivalent witdh, one star per column, BLxxx λ(Å) elem χ gf 038 045 052 065 076 077 079 081 084 085 091 092 096 097 100 104 6141.73 Ba ii 0.70 -0.08 265.7 182.3 227.5 174.2 236.8 226.5 290.2 267.4 238.8 141.6 226.5 224.8 248.4 247.7 197.9 238.5 6496.91 Ba ii 0.60 -0.38 259.7 204.4 278.5 161.6 257.9 244.2 ... 287.1 272.3 134.5 231.4 246.9 260.8 274.8 228.3 252.4 6122.23 Ca i 1.89 -0.32 243.6 219.3 218.2 176.0 242.1 237.9 277.5 265.2 239.3 116.5 214.9 ... 261.2 ... 213.9 236.9 6156.03 Ca i 2.52 -2.39 37.9 ... 23.5 ... 28.3 19.7 52.0 27.6 17.0 ...... 24.2 36.5 25.5 ... 22.9 6161.30 Ca i 2.52 -1.27 144.3 97.1 106.5 44.0 112.7 121.4 167.2 142.5 130.2 17.2 119.5 117.6 141.9 123.9 103.5 131.9 6162.17 Ca i 1.90 -0.32 271.8 248.2 249.9 200.2 266.1 252.9 295.7 293.0 260.4 135.6 243.2 262.4 284.8 267.7 235.3 260.8 6166.44 Ca i 2.52 -1.14 141.2 97.5 107.2 51.6 126.1 114.7 134.7 132.2 122.5 ... 105.6 121.1 123.3 119.1 99.9 122.5 6169.04 Ca i 2.52 -0.80 125.4 121.0 120.8 91.9 146.0 143.9 128.0 151.6 144.6 22.7 134.4 130.2 ... 146.3 115.4 139.0 6169.56 Ca i 2.52 -0.48 149.8 131.0 135.8 112.8 145.6 164.6 160.8 174.6 158.9 35.6 153.7 156.5 ... 161.8 138.9 166.0 6439.08 Ca i 2.52 0.39 234.7 199.4 229.8 167.9 228.2 227.2 258.3 237.6 219.9 105.8 205.8 215.8 ... 228.6 201.2 217.9 6455.60 Ca i 2.52 -1.29 146.3 102.4 31.4 117.2 76.1 ... 40.6 148.5 134.7 ... 123.3 85.0 126.2 154.5 48.1 127.3 6471.67 Ca i 2.52 -0.76 ...... 20.7 ...... 6493.79 Ca i 2.52 -0.32 217.7 166.2 194.6 143.0 177.2 189.8 239.5 195.4 186.5 78.3 169.0 177.1 181.6 174.1 175.1 179.1 6499.65 Ca i 2.52 -0.82 137.9 101.5 130.4 59.6 142.0 117.8 165.6 136.4 137.1 25.3 118.4 133.8 153.3 118.0 124.2 127.4 6508.84 Ca i 2.52 -2.41 38.6 ... 24.6 ... 33.7 22.4 67.9 43.1 26.6 ...... 37.0 ... 37.3 23.1 38.8 6330.09 Cr i 0.94 -2.92 152.6 108.7 130.8 35.1 133.9 128.0 164.9 139.9 134.6 ... 115.6 131.3 137.0 122.6 120.7 106.4 6645.13 Eu ii 1.37 0.20 85.0 27.3 81.5 23.3 50.9 50.7 90.1 59.0 65.4 ... 57.9 62.2 79.0 60.7 ... 76.0 5369.96 Fe i 4.37 0.54 146.2 154.2 125.0 ... 175.2 177.0 ... 180.7 173.9 41.0 159.5 147.4 162.6 145.3 163.9 143.0 5383.37 Fe i 4.31 0.50 165.1 143.0 158.5 126.9 162.0 157.5 154.3 163.4 153.7 68.1 164.4 141.6 166.2 167.7 150.0 158.5 5386.34 Fe i 4.16 -1.74 37.7 62.7 30.2 31.0 ... 54.6 ... 33.9 66.2 ... 32.3 45.6 ... 45.1 39.1 ... 5393.17 Fe i 3.24 -0.92 191.0 191.9 173.7 149.5 196.3 202.0 210.3 208.5 206.8 77.1 205.9 175.9 189.2 206.5 176.0 195.2 5395.22 Fe i 4.45 -1.73 ... 22.3 ...... 41.1 ...... 21.4 ...... 23.4 16.7 ... 5405.79 Fe i 0.99 -1.85 ...... 268.7 ...... 205.0 ...... 5415.19 Fe i 4.39 0.51 115.7 155.6 119.1 137.7 158.7 166.2 148.2 165.9 161.9 71.0 164.8 142.3 149.4 163.3 154.7 145.3 5417.04 Fe i 4.42 -1.42 31.2 27.9 27.6 ... 33.1 40.6 31.0 63.1 37.8 24.9 49.4 33.9 28.8 37.1 32.1 46.6 5434.53 Fe i 1.01 -2.12 ...... 261.2 ...... 206.0 ...... 5436.30 Fe i 4.39 -1.35 ...... 5464.29 Fe i 4.14 -1.62 74.9 41.2 52.9 ... 51.1 57.9 86.4 72.4 73.4 ... 48.5 54.0 55.0 59.4 57.9 56.7 5470.09 Fe i 4.45 -1.60 43.2 21.8 26.7 ... 29.5 ... 34.9 41.2 ...... 18.7 30.8 27.0 25.6 20.8 22.2 5501.48 Fe i 0.96 -3.05 258.9 236.3 239.1 193.2 247.9 269.1 265.1 276.3 273.9 159.8 241.3 254.3 253.3 267.3 243.0 254.7 5506.79 Fe i 0.99 -2.79 ... 273.2 294.2 203.6 ...... 158.7 290.6 ...... 272.1 ... 5539.29 Fe i 3.64 -2.59 47.1 41.0 27.4 31.9 57.1 67.1 62.0 82.3 65.6 ... 56.6 42.3 53.9 63.5 54.8 54.8 5586.77 Fe i 3.37 -0.10 194.0 203.2 141.5 175.9 211.2 232.9 175.6 238.8 226.1 108.8 215.9 207.6 214.7 228.6 221.2 229.2 6120.26 Fe i 0.91 -5.94 70.3 67.7 92.1 33.9 87.0 85.8 87.4 106.2 83.2 ... 57.9 92.5 94.9 96.2 67.6 91.0 6136.62 Fe i 2.45 -1.50 178.0 ... 261.1 209.1 ...... 118.1 ...... 268.1 ...... 6137.00 Fe i 2.20 -2.95 ...... 181.4 ...... 165.2 ...... 6151.62 Fe i 2.18 -3.37 132.7 122.7 137.3 93.6 125.6 148.1 128.0 150.7 126.3 21.4 136.7 133.4 154.4 130.7 126.1 133.9 6157.75 Fe i 4.07 -1.26 125.5 94.5 102.3 51.0 110.7 107.7 137.9 122.8 112.6 ... 94.7 103.1 120.9 113.0 95.4 115.7 6159.38 Fe i 4.61 -1.97 47.8 ...... 28.9 ...... 27.8 ...... 25.2 6165.36 Fe i 4.14 -1.47 64.4 56.4 66.5 37.2 65.2 74.7 88.5 63.9 60.9 ... 66.1 62.7 73.7 68.4 68.0 50.6 6173.34 Fe i 2.22 -2.85 169.3 150.5 169.6 102.5 160.9 166.9 176.6 176.2 169.7 33.9 148.4 164.9 167.7 164.6 155.9 161.5 6180.20 Fe i 2.73 -2.78 49.8 119.4 93.6 79.4 119.5 131.8 112.9 ... 134.2 34.7 130.9 116.4 144.3 123.9 112.0 123.5 6187.99 Fe i 3.94 -1.58 81.4 65.1 72.7 32.1 76.7 81.8 90.9 96.2 78.7 ... 77.9 71.1 76.7 87.8 67.2 79.2 6200.31 Fe i 2.61 -2.44 138.8 138.5 151.4 114.0 151.2 151.5 147.9 164.4 150.6 39.3 144.0 145.4 157.9 152.0 140.0 158.7 6213.43 Fe i 2.22 -2.66 188.6 170.0 195.5 143.4 183.2 203.7 207.2 200.8 200.5 72.5 165.1 178.5 185.4 182.4 175.7 180.6 6219.29 Fe i 2.20 -2.44 199.9 185.7 209.7 135.0 187.8 199.5 203.6 210.5 206.1 68.3 190.2 195.3 205.4 191.4 179.8 190.3 6226.74 Fe i 3.88 -2.20 60.2 41.4 62.4 21.0 59.3 57.1 74.1 48.4 51.9 ... 58.8 70.3 72.2 55.6 57.4 61.7 6252.57 Fe i 2.40 -1.76 217.9 193.7 200.3 154.7 207.4 218.5 203.1 214.9 203.2 115.3 187.0 198.0 216.1 205.6 190.3 205.3 6265.13 Fe i 2.18 -2.55 178.1 167.1 180.8 113.5 179.0 194.9 190.2 186.1 183.0 66.5 170.4 190.9 192.1 196.7 181.2 171.2 6271.28 Fe i 3.32 -2.96 49.7 46.8 59.8 21.5 59.2 63.2 65.5 72.3 56.8 ... 53.4 49.4 67.4 65.0 44.8 49.2 6297.80 Fe i 2.22 -2.74 ...... 42.0 ...... 6301.50 Fe i 3.65 -0.72 ... 144.9 ...... 40.0 ...... 6307.85 Fe i 3.64 -3.27 ...... 33.6 26.1 ...... 15.8 ... 6322.69 Fe i 2.59 -2.43 163.0 140.4 158.5 116.1 160.4 174.6 169.0 170.1 155.4 36.1 146.2 146.2 150.4 159.7 153.5 165.0 6330.85 Fe i 4.73 -1.22 37.8 26.3 35.3 17.3 42.6 38.5 41.3 50.7 31.4 ... 34.7 44.3 34.0 35.6 36.5 17.1 6335.33 Fe i 2.20 -2.23 214.5 198.2 250.2 151.0 224.3 209.1 234.5 220.8 206.2 83.2 197.4 191.7 207.9 213.3 201.1 212.5 6336.82 Fe i 3.69 -1.05 144.3 144.9 145.5 108.8 140.5 139.9 157.7 144.0 142.3 33.7 150.9 129.9 138.0 134.0 130.4 128.6 6344.15 Fe i 2.43 -2.92 169.6 165.4 165.2 100.0 170.3 181.3 196.6 189.3 168.3 20.7 149.4 ... 180.6 170.0 143.6 165.7 6355.04 Fe i 2.84 -2.29 164.2 143.2 149.7 86.5 160.5 163.5 176.3 148.2 153.6 35.2 150.7 159.1 ... 162.0 150.4 160.1 6380.75 Fe i 4.19 -1.50 79.5 49.9 74.2 30.1 76.5 79.5 88.1 87.6 73.3 ... 72.3 78.5 66.7 73.8 63.3 88.9 6392.54 Fe i 2.28 -3.95 80.1 84.8 82.9 46.2 85.7 89.3 78.2 106.7 90.0 17.3 78.8 94.2 95.9 87.9 87.5 94.1 6393.61 Fe i 2.43 -1.63 236.5 228.0 221.5 194.6 241.1 248.4 249.6 260.1 236.6 120.8 240.6 204.9 240.4 251.8 221.3 251.5 6408.03 Fe i 3.69 -1.00 149.0 140.4 142.1 101.0 149.6 153.4 131.4 167.8 140.8 32.2 141.5 118.5 143.8 142.3 123.3 142.7 6419.96 Fe i 4.73 -0.24 112.8 91.1 115.3 61.7 99.9 115.5 132.5 118.3 105.1 ... 96.8 76.4 115.4 110.0 109.5 97.1 6421.36 Fe i 2.28 -2.01 228.9 217.8 233.8 163.9 237.0 234.2 255.5 234.9 223.8 92.9 213.5 218.3 226.9 222.6 214.7 223.8 6430.86 Fe i 2.18 -1.95 256.5 215.3 282.6 174.9 256.8 269.1 ... 272.7 244.7 111.1 222.4 241.4 276.1 263.7 235.9 242.2 6481.87 Fe i 2.27 -2.98 183.4 150.6 183.1 119.2 175.7 176.4 210.5 187.6 169.5 39.8 154.5 167.9 169.3 180.6 173.1 182.9 6498.94 Fe i 0.96 -4.69 205.1 151.9 215.6 110.2 178.6 198.4 243.4 203.5 191.3 28.5 163.7 186.0 208.9 185.0 176.6 205.9 6533.93 Fe i 4.55 -1.46 43.4 34.9 ...... 52.3 51.2 70.7 63.9 39.5 ... 40.0 55.9 50.8 32.2 40.7 37.9 6556.81 Fe i 4.79 -1.72 26.7 20.3 ...... 51.0 24.6 17.7 ...... 28.9 38.5 19.3 21.4 ... 6569.22 Fe i 4.73 -0.42 110.3 84.1 122.0 59.8 115.9 114.9 159.8 124.1 106.0 ... 100.0 104.6 110.5 102.6 111.7 100.3 6574.23 Fe i 0.99 -5.02 191.0 151.3 ... 101.1 177.6 180.7 169.8 191.4 180.5 32.7 150.5 162.1 151.7 186.5 158.2 166.9 6593.88 Fe i 2.43 -2.39 194.6 167.8 100.8 156.8 179.5 192.2 151.5 220.5 217.3 76.2 179.1 156.4 160.7 200.6 173.5 177.5 6597.56 Fe i 4.79 -1.07 45.3 44.8 53.7 ... 66.1 66.7 75.1 68.7 64.2 ... 60.3 43.5 56.1 54.4 58.7 48.3 6608.03 Fe i 2.28 -3.94 92.1 73.6 143.8 35.5 99.6 93.2 154.2 104.1 96.2 29.8 79.7 75.8 119.0 92.3 98.9 94.6 6609.12 Fe i 2.56 -2.66 158.0 149.2 206.9 107.1 167.1 171.1 208.3 179.3 170.5 36.6 156.9 149.4 184.9 159.7 174.4 151.7 6627.54 Fe i 4.54 -1.68 16.9 22.7 83.1 ... 29.7 32.9 75.1 43.1 ...... 25.0 37.4 55.4 25.2 27.6 28.6 6633.76 Fe i 4.56 -0.82 107.4 86.2 149.8 ... 112.2 94.4 155.8 115.3 85.9 ... 102.2 110.4 ... 107.5 95.4 88.3 6646.93 Fe i 2.60 -3.99 63.5 42.7 79.8 27.3 66.0 59.1 92.6 62.5 58.8 ... 49.3 67.5 89.5 52.7 59.8 49.2 6653.85 Fe i 4.15 -2.52 26.6 23.3 69.6 15.0 24.7 24.0 62.2 34.7 36.1 ... 26.0 ... 37.3 28.9 16.5 24.5 5414.08 Fe ii 3.22 -3.61 ... 23.3 ... 42.9 33.4 20.9 ...... 16.4 34.1 ... 34.4 34.1 20.7 ... 5425.25 Fe ii 3.20 -3.36 56.9 50.1 45.7 64.2 44.8 36.0 35.5 38.7 45.3 29.5 53.4 43.7 39.3 37.8 57.1 54.0 6149.25 Fe ii 3.89 -2.72 36.3 39.3 ... 23.8 ...... 25.0 22.6 ... 21.2 25.9 34.5 36.8 36.1 24.0 6432.68 Fe ii 2.89 -3.71 37.7 57.7 72.7 41.8 68.6 64.9 75.8 51.9 48.7 16.7 65.8 51.7 ... 55.3 45.5 54.4 6456.39 Fe ii 3.90 -2.08 56.4 65.0 ... 99.8 83.9 93.6 ... 92.8 64.2 37.7 113.3 ... 31.4 129.0 39.0 105.7 6320.43 La ii 0.17 -1.56 90.7 43.2 ...... 41.7 101.7 88.1 68.3 18.4 57.3 64.6 80.8 58.0 44.3 76.6 6390.46 La ii 0.32 -1.40 71.4 ... 57.1 ... 57.3 50.9 93.8 76.1 78.6 ... 57.7 65.1 ... 66.8 49.1 ... 5528.41 Mg i 4.35 -0.36 221.1 195.6 212.0 165.9 207.0 200.9 253.7 208.8 217.3 112.8 203.8 210.0 220.0 234.8 198.8 219.0 Continued on next page 6.A: Large tables 127

λ(Å) elem χ gf 038 045 052 065 076 077 079 081 084 085 091 092 096 097 100 104 6318.72 Mg i 5.11 -1.97 43.6 ...... 34.9 24.6 ...... 31.1 38.9 52.2 ... 17.7 34.0 6319.24 Mg i 5.11 -2.21 ... 27.0 26.9 ...... 23.3 36.2 24.9 ...... 30.6 42.1 ... 31.8 ... 6319.49 Mg i 5.11 -2.43 ...... 5420.36 Mn i 2.14 -1.46 217.4 160.5 166.5 67.8 188.6 198.1 204.4 203.5 202.4 ... 176.3 162.5 192.5 209.3 174.0 187.5 5432.55 Mn i 0.00 -3.80 287.6 247.7 275.0 120.2 280.9 272.6 ... 291.3 252.2 ... 215.8 251.6 289.0 271.4 249.7 264.5 5516.77 Mn i 2.18 -1.85 169.4 99.3 141.5 36.1 145.4 145.0 179.4 171.6 137.9 ... 117.2 132.7 141.2 141.7 111.8 138.6 6154.23 Na i 2.10 -1.56 38.8 ...... 23.5 ... 35.7 36.3 ...... 25.6 15.5 23.4 19.8 ...... 6160.75 Na i 2.10 -1.26 58.4 ... 23.3 ... 38.2 38.6 65.5 54.5 ...... 38.5 21.1 47.6 39.7 ... 28.5 5416.38 Nd ii 0.86 -0.98 17.2 ...... 17.1 27.5 20.6 18.5 34.1 29.3 ... 32.0 25.3 27.5 28.5 16.3 40.0 5431.54 Nd ii 1.12 -0.47 65.0 ... 43.7 ... 47.3 42.4 73.6 62.0 58.0 35.3 42.5 45.0 72.1 48.3 39.4 52.3 5485.71 Nd ii 1.26 -0.12 40.7 19.3 25.6 ... 27.4 25.2 48.2 34.7 30.1 ... 45.2 39.3 37.3 39.4 ... 35.9 5578.73 Ni i 1.68 -2.67 135.0 119.9 116.0 86.8 117.7 130.9 142.9 140.0 138.5 29.0 133.0 123.0 125.5 141.0 115.4 151.5 5587.87 Ni i 1.93 -2.37 147.1 119.4 121.6 83.3 147.5 159.3 150.8 183.1 157.3 30.6 158.0 129.2 146.5 176.0 135.9 158.7 5589.37 Ni i 3.90 -1.15 25.0 19.1 26.3 ... 32.4 24.6 29.2 34.7 29.1 ... 18.7 31.3 26.7 18.8 23.9 27.7 5593.75 Ni i 3.90 -0.79 41.1 31.2 33.4 21.8 34.5 47.1 43.5 57.6 31.9 ... 35.1 40.6 34.2 30.7 24.5 31.1 6128.97 Ni i 1.68 -3.39 94.8 61.2 93.5 48.1 87.6 76.9 99.4 90.0 82.1 ... 74.3 72.9 105.7 88.8 75.0 82.8 6130.14 Ni i 4.27 -0.98 17.1 ...... 17.9 ...... 17.1 ... 6177.25 Ni i 1.83 -3.60 72.6 40.2 51.7 ... 55.4 57.7 76.8 73.3 65.4 ... 69.5 ... 52.4 72.2 45.0 61.6 6186.72 Ni i 4.11 -0.90 40.7 28.3 22.8 ...... 41.8 32.9 43.6 31.0 ...... 30.2 26.9 ... 6204.61 Ni i 4.09 -1.15 ...... 18.7 20.8 22.3 24.5 22.4 ...... 28.9 ...... 23.0 22.4 ... 6223.99 Ni i 4.10 -0.97 24.2 ...... 24.5 ... 30.3 37.9 24.2 ...... 43.0 37.4 30.9 29.2 ... 6230.10 Ni i 4.11 -1.20 45.7 23.6 24.5 ... 26.5 ... 42.7 ... 26.7 ... 25.4 ...... 32.9 ...... 6322.17 Ni i 4.15 -1.21 21.0 ...... 27.0 ... 20.4 ...... 24.8 19.2 ...... 6327.60 Ni i 1.68 -3.09 110.9 94.9 100.3 70.4 109.7 106.5 127.7 116.7 107.2 ... 102.2 119.9 102.0 117.2 106.6 113.6 6378.26 Ni i 4.15 -0.82 29.4 ...... 25.9 30.9 ... 34.2 32.7 ... 24.2 19.5 34.5 ... 25.8 43.7 6384.67 Ni i 4.15 -1.00 22.9 ...... 33.3 ... 50.2 35.4 30.4 ...... 20.9 27.6 ... 36.2 6482.80 Ni i 1.94 -2.85 110.5 96.8 141.8 48.7 113.0 108.0 147.1 123.6 107.3 16.2 92.5 104.6 115.4 109.2 106.8 107.7 6586.32 Ni i 1.95 -2.79 137.0 135.4 44.1 123.9 66.9 89.8 ... 134.5 144.6 45.8 128.7 57.0 51.5 148.1 66.5 94.6 6598.61 Ni i 4.24 -0.93 ...... 17.7 ...... 34.9 24.2 ... 6635.14 Ni i 4.42 -0.75 ... 15.1 ...... 46.9 ...... 18.6 24.9 ...... 18.2 6300.31 O i 0.00 -9.75 ...... 6363.79 O i 0.02 -10.25 53.3 26.8 39.1 28.4 37.2 45.8 53.3 45.0 34.4 ... 24.6 33.5 57.3 39.5 30.7 46.3 5526.82 Sc ii 1.77 0.03 118.4 113.7 124.9 104.2 109.3 107.1 116.3 114.5 119.4 74.4 121.6 110.8 114.1 120.8 107.5 115.4 6245.62 Sc ii 1.51 -0.97 90.2 58.8 67.4 56.9 65.4 65.0 77.1 58.3 81.7 24.0 72.5 75.4 73.2 68.5 61.8 66.9 6309.90 Sc ii 1.50 -1.52 ... 33.4 ... 26.3 ...... 47.8 ... 38.3 ...... 27.7 42.5 ... 41.7 ... 6604.60 Sc ii 1.36 -1.31 87.1 84.3 109.3 61.2 98.5 80.1 125.6 83.5 105.3 38.9 76.4 90.6 89.4 108.5 82.1 80.4 6125.03 Si i 5.62 -1.57 25.0 24.7 ... 18.3 ...... 22.4 ...... 34.2 6142.48 Si i 5.62 -1.51 ...... 6145.02 Si i 5.61 -1.37 45.1 ...... 16.0 ...... 22.0 ...... 15.2 6155.14 Si i 5.62 -0.80 42.5 38.5 ... 32.4 35.6 39.2 57.8 31.5 27.3 17.0 35.0 25.7 48.1 38.3 23.9 40.1 6237.33 Si i 5.61 -1.02 26.6 ...... 34.7 ...... 21.3 25.5 26.1 24.6 ...... 6243.82 Si i 5.61 -1.27 ...... 23.2 ... 22.8 24.1 ...... 28.2 ...... 24.1 ... 33.2 ...... 5490.16 Ti i 1.46 -0.93 121.8 83.7 104.4 34.0 109.1 102.3 142.0 125.3 112.4 ... 84.4 103.8 109.5 98.8 84.1 116.1 5503.90 Ti i 2.58 -0.19 ... 41.0 ... 19.8 47.0 47.1 83.9 83.4 55.3 ... 43.9 74.4 56.7 71.7 33.9 58.4 6126.22 Ti i 1.07 -1.42 149.2 110.7 119.7 48.9 131.6 130.4 167.1 150.6 142.1 20.0 113.5 107.1 158.2 135.5 113.3 140.1 6220.50 Ti i 2.68 -0.14 78.1 45.9 ...... 48.8 79.3 78.6 ...... 58.6 ...... 6258.10 Ti i 1.44 -0.35 208.0 139.2 ... 77.3 176.3 159.6 231.6 191.6 195.5 ... 143.9 192.0 ... 182.0 ... 194.3 6303.77 Ti i 1.44 -1.57 ... 61.1 66.2 26.0 74.9 72.7 111.6 101.7 84.2 ... 50.4 88.8 102.0 80.1 55.4 83.2 6312.24 Ti i 1.46 -1.55 ...... 66.8 ... 63.2 66.9 115.9 97.8 79.9 ... 51.4 88.8 ... 80.6 54.5 88.2 6336.10 Ti i 1.44 -1.74 94.4 48.9 63.7 16.9 74.7 65.8 111.4 84.7 73.2 ... 41.0 72.2 90.8 70.4 47.0 71.6 6508.12 Ti i 1.43 -2.05 64.6 27.1 ...... 62.0 40.6 105.0 61.9 50.3 ... 28.1 58.9 59.3 37.8 30.2 58.8 6556.08 Ti i 1.46 -1.07 ... 104.7 ...... 164.6 ...... 6599.13 Ti i 0.90 -2.09 151.6 81.6 128.0 35.6 141.1 128.7 179.9 146.9 149.8 ... 87.8 109.4 141.6 130.8 105.4 120.1 6666.53 Ti i 1.46 -1.62 25.9 ... 15.1 ...... 18.7 28.2 25.8 ...... 18.0 ...... 5418.77 Ti ii 1.58 -2.11 111.6 107.9 73.6 89.0 94.6 98.7 97.3 91.1 103.5 69.1 96.9 94.9 108.1 108.4 104.0 103.0 6219.94 Ti ii 2.06 -2.82 27.2 15.0 ...... 6559.58 Ti ii 2.05 -2.02 ... 48.3 ... 45.2 ...... 16.4 ...... 81.2 ...... 6606.95 Ti ii 2.06 -2.79 42.2 30.2 ... 25.2 42.1 31.9 70.3 36.3 47.5 20.9 44.5 36.2 46.3 43.1 44.9 35.8 6680.13 Ti ii 3.09 -1.86 ...... 23.9 18.9 22.7 ...... 19.0 ... 16.6 32.6 27.0 ... 16.8 ... 16.3 6119.53 V i 1.06 -0.32 ...... 71.8 6128.33 V i 1.05 -2.30 ...... 30.3 ...... 6135.37 V i 1.05 -0.75 126.5 48.6 95.7 18.2 87.3 87.7 122.7 107.7 98.3 ... 73.6 101.1 100.9 87.5 62.6 ... 6150.15 V i 0.30 -1.79 156.7 98.6 118.7 18.7 125.6 129.4 186.0 154.2 146.0 17.3 99.4 124.1 156.0 135.7 91.4 150.5 6199.19 V i 0.29 -1.29 185.0 133.3 152.9 40.1 163.3 179.1 210.6 186.5 171.9 ... 130.0 180.5 192.4 162.0 131.0 179.2 6216.36 V i 0.28 -0.81 217.5 141.6 169.3 41.8 176.7 173.2 211.0 184.0 184.0 21.2 136.6 168.0 198.2 165.7 142.7 185.6 6224.51 V i 0.29 -2.01 127.1 77.8 98.9 24.1 115.9 125.0 138.4 136.0 124.2 ... 81.0 123.9 142.2 115.3 85.3 133.1 6233.20 V i 0.28 -2.07 102.1 64.1 92.4 ... 95.4 89.7 128.6 96.4 90.2 ... 61.2 106.5 108.4 85.4 53.3 97.8 6243.11 V i 0.30 -0.98 ...... 77.3 ...... 298.9 ...... 6251.82 V i 0.29 -1.30 163.8 100.2 140.9 36.0 137.2 137.9 174.9 157.7 158.0 ... 115.9 137.2 158.8 139.1 109.5 149.0 6274.66 V i 0.27 -1.67 ...... 91.9 ... 115.1 116.4 155.2 141.2 ...... 92.2 131.1 ... 94.0 90.8 141.8 6357.29 V i 1.85 -0.91 38.3 ...... 17.8 43.7 25.7 ... 15.5 ... 23.8 50.3 ... 15.4 21.2 6452.32 V i 1.19 -1.21 89.2 35.8 32.1 48.4 65.7 72.3 79.2 122.1 85.8 ... 48.5 68.1 59.3 73.6 35.8 74.0 6504.19 V i 1.18 -1.23 60.1 16.5 53.9 ... 41.6 33.3 104.7 62.9 38.6 ... 32.9 64.6 59.4 49.1 36.6 59.1 6531.41 V i 1.22 -0.84 74.8 ...... 70.8 ...... 5402.78 Y ii 1.84 -0.51 52.0 ... 30.7 ...... 67.2 35.5 26.3 ... 25.0 52.1 31.9 39.3 20.5 43.3 6362.35 Zn i 5.80 0.14 ... 17.1 ...... 36.6 ...... 92.9 25.2 ...... 6127.48 Zr i 0.15 -1.06 101.9 33.8 57.2 ... 69.5 67.8 112.9 99.4 79.8 18.2 32.0 95.0 91.7 77.6 43.3 84.5 6140.46 Zr i 0.52 -1.41 36.1 ...... 15.7 30.1 22.6 ...... 21.5 ... 18.5 ... 6143.18 Zr i 0.07 -1.10 120.7 40.0 62.8 ... 74.5 65.4 129.8 101.7 70.6 ... 40.7 70.9 87.5 73.9 ... 96.0 6192.95 Zr i 0.54 -2.07 31.8 ...... 128 chapter 6: HR spectroscopic study of Fornax Field Stars

Table 6.A2: Complete line list with parameters and associated EW s (in mÅ, measured by DAOSPEC) for all the stars. Part 2/5.

Equivalent witdh, one star per column, BLxxx λ(Å) elem χ gf 113 115 123 125 132 135 138 140 141 146 147 148 149 150 151 155 6141.73 Ba ii 0.70 -0.08 276.9 182.4 241.5 256.1 218.8 240.3 255.0 241.1 209.2 240.3 236.5 ... 236.0 230.1 243.6 265.2 6496.91 Ba ii 0.60 -0.38 ... 171.9 235.9 277.9 248.3 255.0 227.5 257.0 227.5 266.2 244.7 227.2 257.8 236.3 253.6 293.6 6122.23 Ca i 1.89 -0.32 249.6 192.4 242.8 254.1 250.0 244.7 255.0 234.6 234.5 230.0 162.3 293.5 228.1 256.4 240.9 ... 6156.03 Ca i 2.52 -2.39 43.4 ...... 40.5 28.5 ... 24.2 34.5 ...... 70.9 18.7 36.6 22.4 27.6 6161.30 Ca i 2.52 -1.27 156.4 52.9 131.7 145.8 126.1 113.7 148.3 131.4 119.2 116.1 66.5 170.4 111.5 125.8 119.4 155.3 6162.17 Ca i 1.90 -0.32 264.2 204.5 262.9 284.9 261.6 266.1 279.5 254.3 247.1 265.4 194.6 ... 244.5 268.6 265.7 277.2 6166.44 Ca i 2.52 -1.14 127.7 60.6 111.8 124.8 121.9 119.2 145.6 133.1 105.3 117.6 52.9 145.8 112.6 116.7 124.8 138.2 6169.04 Ca i 2.52 -0.80 145.9 87.2 141.7 139.1 135.1 129.7 153.1 148.0 149.2 139.0 59.7 143.1 116.5 155.5 148.3 147.9 6169.56 Ca i 2.52 -0.48 165.6 111.5 157.5 178.6 155.6 175.7 146.1 154.7 156.3 167.7 90.4 174.5 144.9 167.7 158.0 187.9 6439.08 Ca i 2.52 0.39 232.7 168.5 225.1 235.0 226.6 235.2 238.2 214.9 219.1 227.5 181.8 253.0 230.3 240.7 220.4 251.1 6455.60 Ca i 2.52 -1.29 155.8 85.6 106.2 92.6 172.9 192.2 134.1 135.9 159.0 155.3 ... 178.6 ... 194.4 177.9 149.0 6471.67 Ca i 2.52 -0.76 ...... 6493.79 Ca i 2.52 -0.32 195.3 139.0 193.3 213.1 187.8 174.8 196.5 176.7 192.2 186.3 134.8 205.0 188.9 199.2 183.4 193.1 6499.65 Ca i 2.52 -0.82 129.7 80.5 132.1 147.0 129.4 122.0 130.1 136.8 128.4 118.6 80.8 139.6 122.0 120.3 134.2 133.4 6508.84 Ca i 2.52 -2.41 42.4 ... 26.8 43.2 27.7 17.6 33.6 19.8 21.9 20.8 ... 46.0 22.3 33.2 29.5 51.5 6330.09 Cr i 0.94 -2.92 127.0 63.7 141.3 127.1 146.4 127.9 141.2 141.4 127.0 132.2 36.7 169.3 116.1 143.9 140.8 146.2 6645.13 Eu ii 1.37 0.20 71.3 15.9 53.5 78.2 49.1 55.0 69.0 79.7 56.2 52.8 99.1 73.9 53.1 68.6 59.5 61.5 5369.96 Fe i 4.37 0.54 167.3 136.9 218.7 145.2 142.3 143.1 173.4 160.3 166.5 165.7 128.4 163.3 132.5 155.5 160.1 119.1 5383.37 Fe i 4.31 0.50 185.9 150.1 150.5 171.1 157.9 160.1 188.8 162.0 157.8 161.6 114.1 182.2 161.0 174.9 159.2 170.3 5386.34 Fe i 4.16 -1.74 53.4 ...... 38.8 57.5 62.9 ... 44.6 51.8 ... 61.2 27.4 33.5 45.2 53.9 5393.17 Fe i 3.24 -0.92 199.6 180.7 198.8 217.9 209.6 233.1 216.2 200.3 218.7 212.3 176.0 239.2 187.7 207.6 221.0 214.6 5395.22 Fe i 4.45 -1.73 29.4 ... 29.8 30.2 23.3 30.1 31.4 29.3 23.1 ...... 34.6 34.9 ... 30.1 31.4 5405.79 Fe i 0.99 -1.85 ...... 262.0 ...... 5415.19 Fe i 4.39 0.51 155.8 145.2 151.4 174.2 167.5 186.8 152.5 155.3 176.2 167.1 108.5 168.1 162.7 166.5 174.0 184.9 5417.04 Fe i 4.42 -1.42 58.4 26.7 38.4 48.0 39.8 51.5 37.4 40.4 34.9 66.0 ... 53.7 28.8 52.4 35.7 71.3 5434.53 Fe i 1.01 -2.12 ... 299.9 ...... 245.4 ...... 5436.30 Fe i 4.39 -1.35 ...... 5464.29 Fe i 4.14 -1.62 82.1 ... 59.5 90.4 53.5 ... 51.8 70.7 52.6 60.1 27.0 ... 49.2 52.6 62.9 73.5 5470.09 Fe i 4.45 -1.60 15.1 ... 28.2 24.7 25.2 17.5 25.4 24.2 33.7 21.2 16.8 ... 26.5 22.1 27.1 27.7 5501.48 Fe i 0.96 -3.05 251.8 238.7 266.5 267.8 275.4 254.1 274.3 260.6 257.5 270.8 185.7 274.6 242.7 273.0 251.4 286.4 5506.79 Fe i 0.99 -2.79 ... 234.7 ...... 214.3 ... 279.6 ...... 5539.29 Fe i 3.64 -2.59 68.3 33.8 49.5 67.2 59.1 ... 57.5 58.8 58.1 59.5 ... 78.3 59.9 63.0 55.9 76.0 5586.77 Fe i 3.37 -0.10 235.9 177.8 206.8 244.4 239.6 245.6 243.9 225.7 228.8 224.0 143.4 248.2 227.3 245.4 228.6 250.2 6120.26 Fe i 0.91 -5.94 79.1 49.5 76.0 92.1 109.1 75.3 88.9 83.4 95.8 85.6 21.9 92.5 83.9 87.7 97.7 89.8 6136.62 Fe i 2.45 -1.50 ... 206.5 ...... 217.6 ...... 6137.00 Fe i 2.20 -2.95 ... 129.5 ...... 6151.62 Fe i 2.18 -3.37 140.1 95.0 138.2 142.5 147.0 125.3 137.1 138.8 125.4 134.2 80.9 147.5 135.1 128.9 134.2 137.7 6157.75 Fe i 4.07 -1.26 133.4 50.1 105.8 125.1 100.1 104.2 113.4 110.4 104.8 115.1 82.2 151.9 86.0 104.4 104.6 126.4 6159.38 Fe i 4.61 -1.97 ...... 23.5 26.4 ...... 18.7 19.1 ...... 65.3 ...... 16.9 34.5 6165.36 Fe i 4.14 -1.47 83.7 32.6 67.6 64.3 64.9 67.5 58.8 65.4 74.7 68.5 24.2 92.0 63.1 72.8 71.0 75.5 6173.34 Fe i 2.22 -2.85 162.7 113.0 156.6 171.8 169.0 156.1 171.4 167.6 157.9 158.9 121.6 176.5 156.0 158.8 157.1 154.7 6180.20 Fe i 2.73 -2.78 133.2 101.8 138.6 119.5 149.9 116.9 123.1 139.7 154.7 132.3 59.3 85.7 132.5 142.0 143.8 147.3 6187.99 Fe i 3.94 -1.58 84.5 43.5 75.3 79.9 84.2 75.2 67.6 81.8 72.9 80.1 40.8 ... 63.7 86.5 76.2 69.4 6200.31 Fe i 2.61 -2.44 158.0 120.6 143.2 152.2 148.0 159.7 153.8 152.7 151.5 149.3 98.8 145.4 154.9 163.4 155.3 158.0 6213.43 Fe i 2.22 -2.66 210.8 147.0 194.4 200.8 184.6 189.0 192.7 183.8 181.6 211.8 143.0 199.2 194.4 183.9 187.4 187.5 6219.29 Fe i 2.20 -2.44 206.1 160.9 189.8 197.5 207.2 ... 198.5 184.7 204.1 202.2 133.7 212.0 191.6 202.5 189.9 201.2 6226.74 Fe i 3.88 -2.20 68.4 32.8 52.3 65.2 55.6 57.7 52.5 57.2 64.5 53.7 29.7 110.4 46.9 70.8 60.1 67.6 6252.57 Fe i 2.40 -1.76 205.2 180.6 207.3 199.2 199.8 198.6 210.1 205.0 194.4 205.9 153.3 223.5 194.1 203.8 199.8 202.1 6265.13 Fe i 2.18 -2.55 178.4 137.1 186.1 173.1 196.7 169.8 190.6 185.0 167.5 195.1 125.9 195.1 176.7 189.2 182.5 199.8 6271.28 Fe i 3.32 -2.96 65.9 17.8 52.9 57.7 64.5 54.3 52.2 68.3 71.0 51.7 21.1 57.0 64.9 53.0 67.1 77.7 6297.80 Fe i 2.22 -2.74 ...... 6301.50 Fe i 3.65 -0.72 ...... 6307.85 Fe i 3.64 -3.27 16.1 ...... 24.0 ...... 23.3 6322.69 Fe i 2.59 -2.43 162.5 127.2 153.4 164.7 164.6 162.2 164.7 160.1 160.6 172.0 111.3 178.3 158.5 153.4 160.5 179.3 6330.85 Fe i 4.73 -1.22 42.9 ... 31.3 48.0 34.2 41.5 37.2 43.4 31.8 27.0 18.2 48.2 39.8 42.7 39.1 37.6 6335.33 Fe i 2.20 -2.23 206.3 173.4 208.8 210.1 218.4 209.7 222.1 221.7 210.4 211.5 141.0 231.0 216.4 216.5 211.3 221.3 6336.82 Fe i 3.69 -1.05 152.0 126.2 141.9 143.0 139.4 138.7 144.1 144.5 170.1 141.9 98.9 163.1 135.3 141.4 136.1 163.9 6344.15 Fe i 2.43 -2.92 184.5 117.9 174.1 185.6 165.2 166.3 173.1 177.7 167.8 163.8 118.9 206.3 160.1 181.2 163.1 187.8 6355.04 Fe i 2.84 -2.29 165.2 116.7 170.7 164.0 159.6 157.0 165.8 163.6 154.9 168.7 87.9 172.1 143.3 172.9 166.7 174.9 6380.75 Fe i 4.19 -1.50 93.9 33.1 73.6 ... 77.9 80.3 77.1 79.7 73.4 65.7 28.2 121.3 99.8 56.4 82.7 86.6 6392.54 Fe i 2.28 -3.95 88.2 51.5 97.3 93.7 90.2 83.9 86.1 82.7 94.8 92.9 31.1 108.3 89.6 80.9 90.9 ... 6393.61 Fe i 2.43 -1.63 251.5 205.5 239.4 231.0 255.6 245.3 244.3 235.1 244.5 235.4 169.2 261.1 228.0 255.6 236.7 259.4 6408.03 Fe i 3.69 -1.00 172.0 122.9 152.2 143.3 141.6 153.1 158.7 137.8 143.7 146.3 94.8 157.1 123.4 166.6 154.6 ... 6419.96 Fe i 4.73 -0.24 112.5 73.9 98.2 104.6 110.7 94.8 113.4 121.4 110.9 108.3 54.0 135.0 99.1 113.0 102.4 106.6 6421.36 Fe i 2.28 -2.01 234.4 182.0 214.5 244.3 233.0 214.8 244.0 221.2 211.1 223.1 175.3 243.3 224.9 230.0 216.7 228.1 6430.86 Fe i 2.18 -1.95 261.9 201.4 250.8 269.5 272.7 264.8 269.7 244.4 251.7 245.5 180.9 297.6 240.2 251.8 245.5 262.0 6481.87 Fe i 2.27 -2.98 154.2 126.2 167.0 181.5 168.9 158.4 195.0 159.7 163.0 162.9 136.6 176.4 172.4 167.1 158.1 153.1 6498.94 Fe i 0.96 -4.69 182.7 131.5 190.8 204.4 196.2 189.8 186.7 188.1 186.6 185.1 134.8 212.6 185.1 183.0 180.5 218.1 6533.93 Fe i 4.55 -1.46 47.3 24.4 39.8 60.5 55.5 41.5 46.8 36.7 55.2 45.5 25.9 55.6 50.5 46.1 46.4 49.6 6556.81 Fe i 4.79 -1.72 18.6 ...... 24.9 21.0 ... 21.1 ... 20.8 ... 20.7 ... 27.2 18.4 20.6 20.4 6569.22 Fe i 4.73 -0.42 134.3 61.1 98.3 118.4 112.2 84.9 106.9 109.0 103.9 110.9 89.4 136.6 117.0 105.7 103.6 110.4 6574.23 Fe i 0.99 -5.02 175.0 142.2 174.0 173.2 ...... 177.1 ... 160.1 171.1 98.8 222.9 144.9 169.5 ...... 6593.88 Fe i 2.43 -2.39 231.2 180.8 195.1 192.9 225.9 205.3 193.8 210.1 182.7 192.4 90.7 246.4 151.8 211.4 227.0 203.7 6597.56 Fe i 4.79 -1.07 93.4 48.1 50.0 63.7 79.3 67.6 71.8 66.2 57.4 60.8 30.7 96.8 49.7 60.1 72.0 73.4 6608.03 Fe i 2.28 -3.94 94.3 58.4 107.6 116.4 104.8 85.5 99.0 91.1 96.0 97.5 60.1 101.3 98.9 97.7 112.8 97.5 6609.12 Fe i 2.56 -2.66 162.7 143.0 168.4 185.4 169.7 163.5 158.3 160.6 164.5 165.9 137.5 161.4 170.6 165.5 167.5 159.9 6627.54 Fe i 4.54 -1.68 39.8 ... 27.9 31.2 25.7 22.6 25.2 22.2 31.8 20.9 34.8 27.9 47.2 39.8 18.6 22.6 6633.76 Fe i 4.56 -0.82 122.6 53.0 103.2 123.2 95.0 109.5 94.9 112.9 112.6 ... 48.1 128.5 118.7 116.7 100.8 111.7 6646.93 Fe i 2.60 -3.99 56.4 34.6 63.6 69.7 55.1 49.0 52.6 56.0 59.1 50.2 ... 72.2 65.1 70.3 47.5 66.6 6653.85 Fe i 4.15 -2.52 33.6 ...... 28.2 24.2 44.9 ... 30.2 34.7 39.6 34.3 37.1 ... 42.9 23.5 40.6 5414.08 Fe ii 3.22 -3.61 29.1 ...... 19.6 19.0 39.2 21.9 27.3 43.1 25.4 ... 24.4 ...... 25.5 5425.25 Fe ii 3.20 -3.36 52.3 37.0 41.5 52.5 40.0 49.2 38.1 44.9 69.4 54.5 50.3 36.7 51.9 45.2 36.7 70.5 6149.25 Fe ii 3.89 -2.72 44.7 25.4 31.8 ... 25.2 ... 40.8 37.5 36.1 33.3 33.6 49.8 24.7 28.9 29.1 ... 6432.68 Fe ii 2.89 -3.71 59.3 57.8 54.5 55.9 50.9 60.5 55.6 59.2 53.4 53.1 59.5 49.9 54.3 64.5 50.9 51.3 6456.39 Fe ii 3.90 -2.08 78.7 74.6 51.0 50.9 ... 44.3 46.7 123.6 78.7 65.5 ... 68.0 27.9 84.7 ...... 6320.43 La ii 0.17 -1.56 86.9 23.0 74.1 119.7 56.6 72.0 80.4 83.5 65.2 71.8 87.8 119.3 61.4 73.2 71.9 100.7 6390.46 La ii 0.32 -1.40 74.4 31.0 74.9 80.7 56.3 56.1 65.2 76.9 62.8 84.1 76.2 85.3 61.3 81.2 74.7 69.7 5528.41 Mg i 4.35 -0.36 229.7 168.1 218.5 234.1 214.5 219.7 211.1 203.8 201.3 213.3 179.4 221.2 203.4 211.4 215.0 205.7 Continued on next page 6.A: Large tables 129

λ(Å) elem χ gf 113 115 123 125 132 135 138 140 141 146 147 148 149 150 151 155 6318.72 Mg i 5.11 -1.97 ... 16.2 34.9 49.1 34.8 ... 35.6 46.7 39.3 30.9 18.4 49.4 ... 48.0 35.6 ... 6319.24 Mg i 5.11 -2.21 ... 16.2 ...... 20.5 ...... 16.8 ... 31.6 29.6 15.2 25.0 43.2 6319.49 Mg i 5.11 -2.43 ...... 17.4 ...... 29.5 ...... 30.2 ...... 5420.36 Mn i 2.14 -1.46 200.1 95.8 203.6 197.0 220.8 196.5 222.4 195.4 174.3 207.8 63.7 205.3 181.0 213.1 211.6 210.3 5432.55 Mn i 0.00 -3.80 263.5 179.6 254.6 266.5 291.7 262.1 ... 262.5 255.0 255.6 124.9 ... 241.8 293.8 261.2 294.8 5516.77 Mn i 2.18 -1.85 ... 47.3 147.1 174.4 162.9 148.4 175.0 136.1 144.6 138.2 32.4 162.0 ... 155.6 148.7 170.4 6154.23 Na i 2.10 -1.56 41.3 ...... 20.5 20.4 20.9 ...... 21.9 20.2 ... 46.1 ... 33.3 24.7 27.3 6160.75 Na i 2.10 -1.26 59.5 ... 31.9 66.9 22.8 49.0 ... 33.4 19.6 26.1 18.9 81.9 41.0 42.1 43.0 58.0 5416.38 Nd ii 0.86 -0.98 27.2 19.7 19.3 41.2 34.4 28.6 16.4 ... 27.1 38.4 16.3 50.2 ... 35.8 38.9 59.2 5431.54 Nd ii 1.12 -0.47 72.2 ... 48.1 72.8 58.9 63.0 54.4 58.8 51.9 57.6 62.1 103.5 ... 66.2 51.0 83.4 5485.71 Nd ii 1.26 -0.12 44.5 ... 37.1 47.6 33.0 37.2 36.3 33.1 32.1 41.4 58.7 50.5 29.4 32.4 36.1 26.9 5578.73 Ni i 1.68 -2.67 143.8 101.6 125.0 142.4 131.0 144.8 178.3 131.1 141.7 139.2 77.9 157.8 123.1 130.2 141.3 159.8 5587.87 Ni i 1.93 -2.37 178.7 110.3 160.3 167.9 174.8 168.4 160.2 154.8 155.5 164.5 77.3 168.8 146.9 165.8 164.0 161.1 5589.37 Ni i 3.90 -1.15 36.9 16.5 17.4 27.1 26.4 30.7 18.4 16.8 27.5 28.4 23.6 42.3 23.3 26.4 33.0 ... 5593.75 Ni i 3.90 -0.79 39.5 24.1 32.6 32.6 30.3 36.5 50.2 40.1 38.7 47.1 21.4 56.9 22.9 30.6 34.2 57.4 6128.97 Ni i 1.68 -3.39 94.9 60.5 87.9 90.5 86.8 88.6 94.2 85.4 81.7 90.6 38.2 106.1 80.3 80.6 80.3 98.7 6130.14 Ni i 4.27 -0.98 ...... 15.2 16.4 ...... 27.6 19.6 ...... 28.5 ...... 18.3 6177.25 Ni i 1.83 -3.60 60.9 30.1 62.8 59.5 71.7 57.9 52.1 69.8 59.9 61.2 15.5 56.6 52.3 59.5 55.4 68.9 6186.72 Ni i 4.11 -0.90 33.4 ... 30.8 39.1 ... 18.6 28.9 20.7 20.9 18.1 ... 63.2 22.5 ... 24.1 46.1 6204.61 Ni i 4.09 -1.15 ... 17.6 ... 21.5 24.8 ...... 23.7 ...... 27.2 26.1 6223.99 Ni i 4.10 -0.97 26.4 ... 28.9 33.5 34.3 39.7 37.0 ... 32.9 26.1 ... 61.9 ... 36.7 37.2 34.1 6230.10 Ni i 4.11 -1.20 36.7 ... 29.3 ...... 29.4 ...... 30.0 ... 41.0 6322.17 Ni i 4.15 -1.21 ...... 15.2 23.0 ...... 15.4 ...... 23.1 ...... 40.0 6327.60 Ni i 1.68 -3.09 119.6 83.7 118.4 111.7 128.7 102.5 123.2 133.8 95.5 119.3 54.5 106.5 105.5 96.0 113.4 134.1 6378.26 Ni i 4.15 -0.82 33.2 ... 27.8 44.3 33.8 ... 32.9 35.4 32.5 23.6 ... 66.8 ...... 27.1 ... 6384.67 Ni i 4.15 -1.00 33.9 21.6 23.3 ... 18.5 33.5 31.6 23.5 33.3 16.9 16.8 46.2 22.6 36.5 ... 46.1 6482.80 Ni i 1.94 -2.85 101.1 66.4 107.9 114.1 126.7 119.3 111.3 116.9 98.8 108.5 62.4 97.0 95.9 105.9 114.7 115.1 6586.32 Ni i 1.95 -2.79 194.5 133.9 128.3 77.4 184.8 163.8 155.5 150.6 148.5 144.5 ... 244.2 45.6 171.5 190.1 198.3 6598.61 Ni i 4.24 -0.93 ...... 32.1 21.3 30.7 22.5 ...... 28.0 27.6 ...... 6635.14 Ni i 4.42 -0.75 ...... 18.4 21.2 ...... 20.3 6300.31 O i 0.00 -9.75 ...... 60.9 ...... 49.7 ... 6363.79 O i 0.02 -10.25 25.3 ... 43.7 27.9 35.3 48.6 56.9 35.5 27.7 38.8 ... 76.2 29.8 30.0 42.6 46.1 5526.82 Sc ii 1.77 0.03 110.3 114.6 109.6 119.6 111.5 113.9 111.5 103.7 107.1 115.0 104.1 106.8 111.5 113.6 105.3 106.3 6245.62 Sc ii 1.51 -0.97 79.5 48.8 86.4 77.6 69.4 69.4 60.4 62.7 59.3 67.5 41.8 95.9 52.3 80.3 67.5 64.6 6309.90 Sc ii 1.50 -1.52 55.7 ... 42.8 46.8 ...... 44.5 ... 27.2 ...... 68.8 ... 41.6 ...... 6604.60 Sc ii 1.36 -1.31 105.8 63.6 99.6 97.4 82.0 93.5 78.6 83.5 73.7 90.6 80.8 109.0 90.8 77.5 92.9 98.3 6125.03 Si i 5.62 -1.57 ...... 20.1 20.8 ...... 31.7 ...... 6142.48 Si i 5.62 -1.51 ...... 6145.02 Si i 5.61 -1.37 25.9 ... 19.3 15.4 ...... 21.0 16.3 ...... 21.8 25.1 ...... 6155.14 Si i 5.62 -0.80 58.8 ... 42.5 46.5 38.9 40.0 50.4 35.3 30.6 52.0 24.6 59.3 47.5 52.5 31.9 49.2 6237.33 Si i 5.61 -1.02 38.7 16.0 20.6 31.0 21.7 18.8 ... 20.6 ... 24.1 16.8 31.6 21.1 25.6 ... 24.1 6243.82 Si i 5.61 -1.27 ...... 37.4 ...... 22.8 ... 22.0 18.5 59.8 ...... 29.3 ... 5490.16 Ti i 1.46 -0.93 110.1 58.0 110.6 116.7 120.2 91.1 122.7 109.7 102.3 114.6 27.2 130.8 82.4 111.7 118.5 128.8 5503.90 Ti i 2.58 -0.19 67.9 37.5 58.1 69.5 68.6 61.1 78.3 64.8 48.9 52.4 17.1 101.7 31.1 60.6 65.1 98.4 6126.22 Ti i 1.07 -1.42 116.7 68.4 139.4 137.9 148.0 144.5 150.6 146.2 118.2 132.2 51.4 192.2 112.1 145.7 139.4 153.2 6220.50 Ti i 2.68 -0.14 ...... 64.9 ... 76.8 ...... 27.1 ...... 91.7 6258.10 Ti i 1.44 -0.35 ... 98.9 190.7 ... 225.1 187.2 206.2 187.1 147.7 182.3 79.2 229.9 152.7 188.5 196.3 219.2 6303.77 Ti i 1.44 -1.57 93.2 42.4 80.8 107.9 91.6 91.4 104.1 86.3 79.1 85.1 ... 137.0 61.7 102.3 91.0 106.3 6312.24 Ti i 1.46 -1.55 ...... 92.2 88.8 89.8 95.7 83.1 ... 71.3 15.5 ... 67.4 ... 86.2 93.4 6336.10 Ti i 1.44 -1.74 68.1 26.4 74.0 76.7 79.4 76.6 90.5 68.3 65.0 73.7 ... 104.9 ... 84.8 78.0 86.6 6508.12 Ti i 1.43 -2.05 38.7 ... 46.6 68.8 ... 61.0 61.7 52.8 30.8 44.5 ... 100.9 40.0 52.2 54.7 76.2 6556.08 Ti i 1.46 -1.07 ... 63.2 143.9 ...... 111.7 ...... 139.5 ...... 6599.13 Ti i 0.90 -2.09 131.4 67.1 134.4 152.6 164.2 126.2 147.0 141.4 94.4 127.9 48.6 208.9 99.1 132.2 146.6 157.7 6666.53 Ti i 1.46 -1.62 ...... 16.4 34.4 ... 27.9 27.4 16.3 ...... 39.6 ... 15.5 28.1 32.9 5418.77 Ti ii 1.58 -2.11 106.8 93.4 88.2 102.3 109.2 97.9 108.2 109.0 104.4 99.2 82.3 97.1 122.5 112.2 115.8 107.1 6219.94 Ti ii 2.06 -2.82 ... 27.4 ...... 32.2 ...... 60.9 ...... 29.2 32.8 6559.58 Ti ii 2.05 -2.02 ... 36.1 61.0 ...... 59.4 ...... 57.0 ... 66.1 ...... 6606.95 Ti ii 2.06 -2.79 50.9 28.4 44.2 54.1 32.3 31.4 31.2 36.8 29.3 ... 61.6 37.6 50.0 33.7 35.9 28.7 6680.13 Ti ii 3.09 -1.86 33.2 18.7 ... 29.1 ... 21.6 22.9 ... 22.5 ... 24.3 ... 21.5 35.0 32.5 29.3 6119.53 V i 1.06 -0.32 82.3 ...... 6128.33 V i 1.05 -2.30 ...... 23.1 ... 17.2 ...... 43.4 ... 22.1 ... 19.4 6135.37 V i 1.05 -0.75 100.8 25.1 100.6 108.0 123.9 100.8 105.6 95.6 93.4 88.5 ... 153.8 75.9 90.9 97.9 122.0 6150.15 V i 0.30 -1.79 119.9 59.5 143.2 142.7 168.2 130.2 166.0 141.7 108.0 136.4 ... 177.8 102.1 157.5 147.7 169.2 6199.19 V i 0.29 -1.29 171.2 95.8 171.8 180.1 194.0 165.2 199.6 182.2 152.7 161.8 36.4 207.4 139.5 181.2 187.1 213.4 6216.36 V i 0.28 -0.81 160.3 78.5 175.3 189.1 187.8 175.4 199.9 184.7 144.7 170.8 52.3 256.3 154.2 187.2 187.2 207.0 6224.51 V i 0.29 -2.01 120.1 ... 121.9 128.5 137.2 113.9 145.1 123.8 103.5 122.7 21.9 177.8 103.8 119.4 132.3 153.6 6233.20 V i 0.28 -2.07 79.7 22.3 99.8 111.8 105.5 76.6 96.4 83.2 72.7 80.9 ... 123.7 53.8 90.8 87.9 121.3 6243.11 V i 0.30 -0.98 ... 124.7 ...... 58.6 ...... 6251.82 V i 0.29 -1.30 130.4 58.6 144.4 149.5 161.7 135.8 164.0 149.2 115.8 128.4 25.6 188.7 124.9 145.8 143.9 153.2 6274.66 V i 0.27 -1.67 ...... 133.2 126.7 142.6 114.0 ... 113.2 19.0 ... 77.9 ... 122.1 148.4 6357.29 V i 1.85 -0.91 21.3 ...... 36.8 30.5 24.3 ... 23.6 16.9 26.3 36.0 65.0 ... 28.3 ... 48.9 6452.32 V i 1.19 -1.21 95.2 32.5 73.5 96.5 103.1 86.9 90.8 79.7 61.2 67.2 ... 151.8 22.2 88.8 72.7 120.2 6504.19 V i 1.18 -1.23 55.7 ... 44.8 71.0 53.9 21.6 51.6 47.1 49.9 38.2 ... 77.6 34.0 57.2 41.9 47.9 6531.41 V i 1.22 -0.84 ... 29.1 66.0 ...... 71.7 ...... 100.9 ... 61.7 ...... 5402.78 Y ii 1.84 -0.51 40.7 ... 36.9 50.8 ... 34.7 32.2 48.4 22.9 37.6 22.9 45.7 43.3 ...... 59.2 6362.35 Zn i 5.80 0.14 ...... 31.9 ...... 46.5 ...... 6127.48 Zr i 0.15 -1.06 68.5 19.4 76.8 86.5 89.3 80.7 87.0 95.8 52.4 68.5 18.5 109.5 64.6 75.7 79.8 104.1 6140.46 Zr i 0.52 -1.41 ...... 18.8 21.7 ...... 25.1 22.7 ...... 53.4 ...... 18.9 6143.18 Zr i 0.07 -1.10 83.7 ... 77.0 108.7 78.8 70.2 107.7 78.6 60.7 72.7 23.9 145.5 57.6 73.9 65.8 111.0 6192.95 Zr i 0.54 -2.07 ...... 17.6 23.2 ...... 18.1 130 chapter 6: HR spectroscopic study of Fornax Field Stars

Table 6.A2: Complete line list with parameters and associated EW s (in mÅ, measured by DAOSPEC) for all the stars. Part 3/5.

Equivalent witdh, one star per column, BLxxx λ(Å) elem χ gf 156 158 160 163 166 168 171 173 180 185 190 195 196 197 203 204 6141.73 Ba ii 0.70 -0.08 225.1 ... 240.2 286.5 266.8 229.1 238.0 254.3 ... 271.5 221.8 198.9 233.8 248.2 235.4 226.6 6496.91 Ba ii 0.60 -0.38 244.0 250.3 248.0 ... 254.1 238.6 221.7 265.0 290.5 279.9 230.2 229.5 231.6 294.5 245.7 293.4 6122.23 Ca i 1.89 -0.32 215.2 219.7 242.5 287.1 251.3 245.5 247.3 258.7 210.9 254.2 249.5 205.9 225.2 236.6 245.0 234.6 6156.03 Ca i 2.52 -2.39 ... 22.1 32.8 40.5 ...... 29.0 30.0 44.3 20.5 ... 20.3 37.2 28.5 ... 6161.30 Ca i 2.52 -1.27 82.4 117.5 119.1 152.0 138.7 132.7 115.0 143.1 128.1 146.0 122.7 93.3 111.4 140.4 143.2 114.8 6162.17 Ca i 1.90 -0.32 236.3 240.9 261.9 299.6 263.2 271.3 268.3 285.0 256.5 294.3 254.4 238.4 251.0 253.5 251.8 270.0 6166.44 Ca i 2.52 -1.14 89.8 102.1 122.8 137.7 117.9 110.2 121.7 132.9 108.8 132.6 134.0 93.0 115.0 127.3 128.5 112.0 6169.04 Ca i 2.52 -0.80 128.6 116.8 136.9 160.1 143.0 154.3 147.0 145.1 101.7 154.0 145.7 108.5 128.6 124.5 148.2 140.9 6169.56 Ca i 2.52 -0.48 131.8 148.2 162.7 181.1 174.7 173.7 177.3 156.9 127.1 165.2 155.1 130.8 156.6 145.7 161.1 156.5 6439.08 Ca i 2.52 0.39 200.2 228.1 212.2 251.3 229.0 232.1 243.3 232.7 ... 236.3 246.8 202.8 216.5 ... 237.1 ... 6455.60 Ca i 2.52 -1.29 81.4 127.6 154.4 225.1 178.2 184.1 163.1 198.1 ... 171.9 166.4 124.8 110.0 ... 106.9 ... 6471.67 Ca i 2.52 -0.76 ...... 6493.79 Ca i 2.52 -0.32 168.5 163.9 179.8 198.6 192.5 186.1 199.3 193.7 185.8 182.5 192.0 152.5 182.4 207.2 192.7 208.7 6499.65 Ca i 2.52 -0.82 109.3 117.4 123.9 150.4 124.5 133.7 131.7 139.9 132.3 142.3 130.2 106.8 124.4 151.7 137.3 132.1 6508.84 Ca i 2.52 -2.41 ... 16.5 32.7 42.6 23.9 24.2 26.2 20.6 24.7 45.6 25.6 18.7 21.1 43.6 38.5 ... 6330.09 Cr i 0.94 -2.92 104.5 117.7 122.9 152.9 117.2 131.3 150.5 142.5 120.8 158.1 133.9 92.1 137.3 136.1 132.3 126.3 6645.13 Eu ii 1.37 0.20 60.6 73.1 60.1 79.0 56.8 67.5 57.4 61.5 73.6 69.8 61.4 50.0 58.8 76.5 68.6 103.2 5369.96 Fe i 4.37 0.54 162.0 166.7 206.6 171.3 151.9 168.6 178.3 143.9 155.9 ... 160.8 192.0 199.2 167.9 147.9 153.2 5383.37 Fe i 4.31 0.50 169.9 166.2 139.3 175.2 185.8 173.5 171.8 169.9 150.5 163.7 171.9 157.3 163.1 172.5 140.9 156.9 5386.34 Fe i 4.16 -1.74 53.6 53.1 54.8 47.3 38.5 ... 54.9 56.4 54.0 67.3 62.0 25.5 44.7 37.7 48.4 58.9 5393.17 Fe i 3.24 -0.92 183.9 210.6 219.9 230.0 229.4 216.6 211.2 251.7 193.1 218.9 204.4 181.8 194.5 179.7 220.3 197.7 5395.22 Fe i 4.45 -1.73 20.7 ...... 34.8 35.1 22.3 16.7 24.3 ... 26.4 42.7 ... 19.6 21.3 ...... 5405.79 Fe i 0.99 -1.85 ...... 5415.19 Fe i 4.39 0.51 136.8 151.1 159.0 193.5 181.7 175.9 195.7 191.2 137.7 163.0 176.3 163.9 169.2 150.8 174.7 153.3 5417.04 Fe i 4.42 -1.42 47.1 40.0 70.3 49.3 52.3 53.5 43.0 57.7 ... 53.1 48.2 46.1 46.2 ... 19.9 33.8 5434.53 Fe i 1.01 -2.12 ...... 5436.30 Fe i 4.39 -1.35 ... 101.2 ...... 5464.29 Fe i 4.14 -1.62 27.8 57.0 76.3 78.1 67.3 56.4 50.9 60.2 72.8 67.2 57.1 61.1 38.4 72.1 ... 54.1 5470.09 Fe i 4.45 -1.60 ... 25.4 28.3 42.6 22.5 16.5 ... 40.6 38.9 32.0 21.5 21.8 19.7 30.1 18.4 17.6 5501.48 Fe i 0.96 -3.05 234.0 233.3 257.0 278.4 266.8 264.8 282.2 272.6 232.1 255.2 280.6 243.2 262.5 274.4 261.9 230.6 5506.79 Fe i 0.99 -2.79 262.4 295.6 ...... 290.0 ...... 274.6 ...... 295.9 5539.29 Fe i 3.64 -2.59 32.4 64.9 65.7 76.2 70.3 57.9 57.3 85.4 43.4 75.9 57.1 53.5 46.2 39.2 57.9 40.8 5586.77 Fe i 3.37 -0.10 227.9 246.8 223.8 250.6 246.9 237.1 230.5 256.1 168.4 232.4 238.7 221.8 220.4 202.2 243.4 195.1 6120.26 Fe i 0.91 -5.94 76.0 79.1 96.3 110.2 103.0 109.5 94.5 82.0 63.8 83.3 107.0 66.7 88.7 90.0 98.9 82.2 6136.62 Fe i 2.45 -1.50 ...... 227.1 261.7 281.1 ...... 6137.00 Fe i 2.20 -2.95 ...... 171.3 158.0 170.5 ...... 6151.62 Fe i 2.18 -3.37 113.5 140.1 130.2 133.1 138.0 122.9 129.2 143.7 122.6 144.0 143.0 126.9 128.2 129.0 133.4 126.1 6157.75 Fe i 4.07 -1.26 94.5 101.4 116.5 120.5 112.6 120.9 109.4 112.1 141.1 118.0 108.9 95.3 105.5 121.0 115.0 121.7 6159.38 Fe i 4.61 -1.97 20.1 ...... 28.6 ...... 20.8 ... 38.7 17.0 ...... 24.3 20.3 18.5 6165.36 Fe i 4.14 -1.47 55.7 67.6 71.0 76.0 72.6 67.6 64.9 69.3 66.4 87.0 80.6 66.7 67.5 69.9 73.3 54.5 6173.34 Fe i 2.22 -2.85 142.5 142.6 173.1 169.8 155.6 157.9 153.4 165.3 160.6 162.6 182.1 155.7 154.8 163.1 163.9 162.8 6180.20 Fe i 2.73 -2.78 111.3 119.3 126.8 ... 154.7 170.7 147.2 144.3 74.0 126.0 151.5 118.7 ... 102.6 137.8 133.2 6187.99 Fe i 3.94 -1.58 64.3 64.9 82.2 85.5 83.3 79.1 68.8 82.8 73.2 81.0 79.5 64.7 77.8 81.6 74.0 71.1 6200.31 Fe i 2.61 -2.44 145.8 135.5 155.6 168.8 148.7 150.1 155.4 157.8 147.7 151.6 159.6 132.3 148.6 130.0 156.1 145.8 6213.43 Fe i 2.22 -2.66 171.7 193.8 185.5 195.0 192.3 181.1 201.0 191.1 187.4 203.5 205.5 164.8 190.2 193.6 184.3 190.5 6219.29 Fe i 2.20 -2.44 187.0 207.9 201.6 214.2 204.7 187.9 193.5 211.4 192.1 223.9 205.5 183.7 193.6 202.9 198.7 201.7 6226.74 Fe i 3.88 -2.20 33.5 58.5 59.9 71.2 59.1 59.8 54.7 45.7 58.4 60.0 72.2 68.3 55.5 60.3 72.0 68.6 6252.57 Fe i 2.40 -1.76 188.6 190.2 210.8 222.0 217.8 200.5 202.5 209.8 198.5 199.5 223.8 194.3 211.5 214.7 190.6 202.6 6265.13 Fe i 2.18 -2.55 174.0 168.5 192.2 193.1 192.4 185.9 188.1 170.4 173.1 185.7 196.2 162.4 184.3 187.6 176.2 182.3 6271.28 Fe i 3.32 -2.96 47.9 52.8 55.0 66.9 60.2 61.6 53.1 72.0 61.2 62.3 61.9 44.7 51.7 61.7 55.3 70.6 6297.80 Fe i 2.22 -2.74 ...... 6301.50 Fe i 3.65 -0.72 ...... 6307.85 Fe i 3.64 -3.27 ...... 23.2 ...... 33.5 ...... 18.3 ...... 6322.69 Fe i 2.59 -2.43 151.4 148.5 160.4 165.2 147.5 166.0 169.9 171.5 137.3 162.4 173.7 140.8 156.4 158.7 140.2 153.5 6330.85 Fe i 4.73 -1.22 22.6 31.5 39.7 38.0 38.9 36.2 26.2 38.3 35.7 43.1 37.8 25.8 25.2 34.7 28.5 50.5 6335.33 Fe i 2.20 -2.23 198.4 198.9 207.1 227.9 206.1 194.0 209.6 229.7 201.3 223.0 226.4 188.2 220.2 221.9 213.3 212.8 6336.82 Fe i 3.69 -1.05 153.7 138.2 142.5 153.3 138.0 154.9 144.5 164.3 144.8 144.1 154.6 137.6 149.2 144.0 134.1 140.9 6344.15 Fe i 2.43 -2.92 161.7 167.4 163.1 201.9 177.3 165.8 169.7 188.9 175.8 198.2 183.3 126.5 ... 163.6 168.5 160.7 6355.04 Fe i 2.84 -2.29 131.6 154.5 151.7 158.8 158.5 147.6 153.7 165.2 165.5 174.2 160.0 139.9 163.1 161.3 155.1 138.5 6380.75 Fe i 4.19 -1.50 50.2 88.4 74.7 93.3 67.6 80.9 68.7 96.5 72.3 99.7 90.6 77.2 80.1 75.6 84.5 99.0 6392.54 Fe i 2.28 -3.95 82.1 91.4 87.2 90.5 ... 76.7 88.9 116.1 75.4 93.4 97.6 66.6 88.9 91.1 98.1 94.1 6393.61 Fe i 2.43 -1.63 225.5 226.0 241.8 257.2 227.9 243.4 251.5 249.0 236.4 258.3 256.2 225.2 236.9 220.9 242.8 235.4 6408.03 Fe i 3.69 -1.00 124.1 126.1 145.9 177.4 151.3 144.5 165.7 152.3 ... 169.4 150.4 138.7 159.9 ... 128.7 ... 6419.96 Fe i 4.73 -0.24 91.4 112.4 106.4 123.4 89.2 96.0 112.2 120.5 ... 113.2 115.1 104.2 103.8 ... 107.7 ... 6421.36 Fe i 2.28 -2.01 209.8 226.0 215.6 242.9 231.2 194.9 223.8 255.2 ... 232.8 241.4 185.8 220.4 ... 222.5 ... 6430.86 Fe i 2.18 -1.95 230.1 248.4 236.8 275.6 260.4 255.1 265.1 282.6 ... 261.4 279.5 227.1 232.2 ... 256.6 ... 6481.87 Fe i 2.27 -2.98 147.4 148.1 176.7 169.4 183.8 160.2 175.7 161.7 172.7 185.0 171.9 151.2 158.9 178.4 176.9 197.5 6498.94 Fe i 0.96 -4.69 170.8 182.5 181.7 205.0 205.6 168.4 196.6 196.1 188.4 207.3 199.9 160.8 175.8 210.1 185.3 209.3 6533.93 Fe i 4.55 -1.46 29.3 48.7 61.8 56.6 41.8 38.0 44.7 50.9 ... 47.6 52.4 28.8 51.7 ... 42.2 ... 6556.81 Fe i 4.79 -1.72 ...... 15.9 24.1 ...... 15.1 ... 19.7 22.0 ... 27.0 ...... 6569.22 Fe i 4.73 -0.42 90.2 107.9 112.2 136.1 109.8 104.0 107.7 106.7 117.5 126.6 108.0 93.8 98.3 121.4 118.0 154.9 6574.23 Fe i 0.99 -5.02 ...... 167.7 ... 188.2 182.6 180.5 ...... 187.1 197.2 ... 187.4 ... 168.0 ... 6593.88 Fe i 2.43 -2.39 145.7 152.4 195.7 250.6 196.8 228.2 219.0 224.9 ... 208.3 228.4 180.0 196.7 ... 152.3 ... 6597.56 Fe i 4.79 -1.07 32.5 51.8 55.0 87.4 70.8 74.9 81.6 65.7 ... 57.5 77.0 52.0 57.1 ... 50.1 ... 6608.03 Fe i 2.28 -3.94 85.8 100.6 89.7 115.7 96.7 95.5 106.8 106.3 123.3 99.0 98.6 79.3 108.7 134.5 109.1 164.2 6609.12 Fe i 2.56 -2.66 165.1 162.7 150.2 190.8 166.6 150.6 171.3 184.7 184.2 161.9 170.8 159.1 159.7 185.0 173.3 225.5 6627.54 Fe i 4.54 -1.68 20.0 33.5 26.7 25.1 24.2 33.4 32.3 29.9 60.6 ... 33.8 20.8 23.3 49.4 41.7 110.8 6633.76 Fe i 4.56 -0.82 85.7 111.2 115.3 118.6 111.0 94.5 98.3 122.9 133.9 121.4 112.9 97.3 89.6 137.6 119.3 187.1 6646.93 Fe i 2.60 -3.99 38.6 59.8 ... 65.3 59.4 73.3 61.4 67.3 ... 64.6 65.2 36.6 62.4 72.8 55.3 115.7 6653.85 Fe i 4.15 -2.52 ... 40.3 ... 38.4 31.6 ... 34.8 29.7 36.2 32.1 27.3 21.9 ... 33.3 33.4 ... 5414.08 Fe ii 3.22 -3.61 23.4 39.7 32.5 25.7 51.9 ... 36.1 27.0 ...... 31.1 40.8 30.7 ... 29.3 ... 5425.25 Fe ii 3.20 -3.36 50.8 45.4 42.3 49.5 38.5 61.4 56.9 34.6 53.4 44.7 48.6 48.0 28.5 35.0 57.8 41.7 6149.25 Fe ii 3.89 -2.72 41.4 35.9 29.4 25.8 ... 27.5 ... 25.8 ... 42.8 21.1 ... 36.5 29.5 ... 43.9 6432.68 Fe ii 2.89 -3.71 59.9 70.8 55.4 66.9 45.2 35.2 51.1 48.7 ... 54.0 54.5 66.5 45.2 ... 53.5 ... 6456.39 Fe ii 3.90 -2.08 113.0 ... 69.2 ... 112.6 91.3 99.5 67.9 ... 43.3 ... 60.7 70.7 ...... 6320.43 La ii 0.17 -1.56 66.6 95.5 80.2 99.9 88.8 82.1 70.1 90.0 108.7 111.8 41.4 ... 75.6 79.3 73.3 53.4 6390.46 La ii 0.32 -1.40 73.9 70.9 ... 92.7 61.5 76.2 63.4 84.2 93.7 81.7 59.7 37.8 58.1 98.7 76.9 77.3 5528.41 Mg i 4.35 -0.36 202.6 214.5 206.5 221.8 216.6 205.1 214.3 225.0 209.3 228.0 209.7 196.9 208.5 209.9 212.6 217.3 Continued on next page 6.A: Large tables 131

λ(Å) elem χ gf 156 158 160 163 166 168 171 173 180 185 190 195 196 197 203 204 6318.72 Mg i 5.11 -1.97 ... 37.4 29.8 41.1 37.8 32.0 38.7 36.9 33.1 47.8 29.6 32.7 ... 40.3 37.3 ... 6319.24 Mg i 5.11 -2.21 20.3 35.0 ...... 25.3 ... 30.1 25.5 ... 25.5 ...... 28.9 ... 6319.49 Mg i 5.11 -2.43 ...... 31.2 ...... 5420.36 Mn i 2.14 -1.46 150.2 206.1 203.1 221.2 215.5 189.3 222.0 210.7 181.6 211.2 208.7 169.8 168.1 198.9 217.9 180.9 5432.55 Mn i 0.00 -3.80 214.4 255.9 264.5 ... 269.7 274.6 ...... 241.8 272.8 287.4 191.7 258.7 297.8 297.4 254.9 5516.77 Mn i 2.18 -1.85 95.4 ... 151.6 166.5 151.8 133.5 138.1 ... 130.4 182.3 145.0 108.8 133.6 159.5 152.5 142.7 6154.23 Na i 2.10 -1.56 ... 22.3 ... 41.9 19.6 ...... 25.5 19.3 49.6 ... 18.2 ... 23.6 24.3 ... 6160.75 Na i 2.10 -1.26 ... 42.3 40.7 53.0 47.3 47.9 ... 59.0 45.5 54.0 ... 23.7 19.0 42.0 53.5 37.6 5416.38 Nd ii 0.86 -0.98 35.7 35.1 43.2 54.3 38.9 29.1 30.3 28.6 ...... 16.6 28.8 ...... 5431.54 Nd ii 1.12 -0.47 ... 80.1 52.8 76.6 60.0 47.6 45.2 77.4 66.9 68.0 54.2 31.0 34.9 49.7 65.3 46.7 5485.71 Nd ii 1.26 -0.12 19.8 33.2 42.0 53.0 43.5 35.4 17.7 30.8 61.9 51.3 26.4 18.0 40.2 43.7 40.8 ... 5578.73 Ni i 1.68 -2.67 112.7 147.7 142.8 145.1 147.2 142.7 128.4 156.1 130.1 150.4 143.3 115.8 132.9 135.1 136.1 121.8 5587.87 Ni i 1.93 -2.37 136.7 147.5 163.4 191.0 175.0 167.2 154.0 169.1 146.5 176.9 161.7 153.4 150.3 143.2 153.9 137.3 5589.37 Ni i 3.90 -1.15 ...... 26.9 ... 19.9 23.5 31.0 36.2 25.2 45.0 19.1 22.3 21.7 35.6 23.8 32.3 5593.75 Ni i 3.90 -0.79 23.7 49.4 47.4 45.0 43.7 34.6 41.7 54.7 34.6 40.8 40.3 39.4 27.0 32.1 40.5 36.3 6128.97 Ni i 1.68 -3.39 68.1 79.7 76.2 84.7 82.6 89.9 84.1 83.6 79.3 95.2 87.8 69.8 72.5 86.7 98.0 76.8 6130.14 Ni i 4.27 -0.98 ...... 16.4 23.1 ...... 29.0 ...... 6177.25 Ni i 1.83 -3.60 54.1 53.2 69.2 79.8 51.3 62.1 69.6 55.4 44.1 64.2 67.1 38.7 55.5 57.8 56.6 60.8 6186.72 Ni i 4.11 -0.90 18.1 20.5 35.0 39.3 28.1 42.4 24.0 33.1 23.7 32.2 35.7 23.5 16.5 23.8 32.3 20.9 6204.61 Ni i 4.09 -1.15 ... 20.4 36.3 38.5 15.0 21.8 ... 25.7 ... 16.3 30.5 ...... 6223.99 Ni i 4.10 -0.97 ...... 20.6 47.2 ... 22.3 ... 33.9 18.6 41.4 ...... 21.7 36.9 38.3 ... 6230.10 Ni i 4.11 -1.20 ...... 29.2 ...... 27.6 29.1 37.6 47.3 ...... 20.5 ... 6322.17 Ni i 4.15 -1.21 ... 30.6 ...... 32.4 20.6 46.5 ... 26.7 ...... 6327.60 Ni i 1.68 -3.09 106.8 117.3 114.2 132.6 112.0 106.7 105.5 130.8 109.1 123.6 114.1 99.2 116.8 122.5 93.3 107.2 6378.26 Ni i 4.15 -0.82 20.6 33.7 28.8 33.4 44.1 27.9 38.4 ... 32.6 23.7 15.5 ... 27.7 ... 34.5 ... 6384.67 Ni i 4.15 -1.00 ...... 26.3 41.4 31.7 26.6 28.4 ...... 42.2 ...... 22.3 ...... 6482.80 Ni i 1.94 -2.85 105.1 116.7 109.1 120.5 106.3 97.9 101.4 113.7 104.0 113.9 106.0 96.5 102.1 131.1 111.4 109.4 6586.32 Ni i 1.95 -2.79 57.5 84.8 133.6 210.7 143.3 217.8 183.8 192.6 ... 176.2 185.4 109.3 179.5 ... 92.3 ... 6598.61 Ni i 4.24 -0.93 ...... 17.0 ... 30.3 ... 34.6 ...... 6635.14 Ni i 4.42 -0.75 29.3 ...... 25.1 ...... 38.8 25.7 ... 6300.31 O i 0.00 -9.75 ... 66.7 ...... 6363.79 O i 0.02 -10.25 ... 43.1 ... 46.5 32.7 33.3 41.0 82.8 33.3 43.3 35.4 22.6 ... 42.8 34.7 38.5 5526.82 Sc ii 1.77 0.03 116.2 104.1 119.5 105.9 110.4 99.2 130.1 95.1 123.1 118.7 99.5 99.6 124.8 115.6 129.0 112.1 6245.62 Sc ii 1.51 -0.97 74.3 73.9 75.9 61.7 65.2 65.3 78.0 72.4 82.5 64.3 ... 62.0 71.6 57.8 82.2 65.5 6309.90 Sc ii 1.50 -1.52 ...... 38.9 ... 41.7 41.5 39.1 ... 40.4 ... 29.6 ... 30.4 ... 47.4 46.8 6604.60 Sc ii 1.36 -1.31 79.4 96.7 80.9 105.7 58.9 84.3 102.2 88.2 101.6 74.8 76.6 89.8 81.3 112.3 101.7 130.0 6125.03 Si i 5.62 -1.57 ...... 6142.48 Si i 5.62 -1.51 ...... 18.4 ...... 6145.02 Si i 5.61 -1.37 20.6 17.7 21.8 ...... 22.4 ...... 6155.14 Si i 5.62 -0.80 ... 38.4 36.0 45.6 29.4 32.6 ... 55.7 54.0 53.8 35.9 46.7 38.5 44.0 34.2 47.1 6237.33 Si i 5.61 -1.02 18.4 ... 31.0 26.0 ... 17.4 ...... 31.4 ... 15.2 15.8 ... 24.4 19.9 ... 6243.82 Si i 5.61 -1.27 15.7 31.6 33.4 ... 36.7 33.6 ...... 43.3 ... 23.6 ...... 43.2 ... 5490.16 Ti i 1.46 -0.93 70.3 99.8 124.8 133.6 109.2 111.7 104.3 130.2 84.2 114.5 118.2 68.1 88.1 108.5 126.7 111.7 5503.90 Ti i 2.58 -0.19 23.3 60.9 71.0 87.9 64.4 66.8 51.1 76.5 45.1 73.1 68.1 25.3 47.4 58.7 62.7 41.7 6126.22 Ti i 1.07 -1.42 113.8 119.7 138.8 153.9 132.6 132.0 132.4 152.3 124.4 149.9 147.5 101.3 120.6 145.1 154.4 126.1 6220.50 Ti i 2.68 -0.14 ... 80.5 ...... 69.0 ...... 72.4 ... 90.4 ...... 6258.10 Ti i 1.44 -0.35 138.9 157.8 186.4 219.4 176.7 206.5 ... 207.7 149.3 197.1 191.4 117.3 158.3 205.0 185.6 178.4 6303.77 Ti i 1.44 -1.57 33.8 60.0 ... 86.3 87.1 98.3 71.8 105.4 78.6 105.8 79.9 44.6 79.1 90.3 80.7 78.9 6312.24 Ti i 1.46 -1.55 37.9 60.4 81.1 97.8 72.4 ... 68.2 88.5 ... 87.9 84.4 44.5 ... 89.6 89.7 75.1 6336.10 Ti i 1.44 -1.74 36.1 54.6 73.5 102.5 75.8 68.5 83.5 100.7 54.6 97.4 64.3 36.0 62.0 79.5 73.6 82.2 6508.12 Ti i 1.43 -2.05 ... 24.8 62.3 ... 56.1 42.3 42.8 59.6 33.9 64.9 45.2 22.9 33.8 65.3 54.9 57.5 6556.08 Ti i 1.46 -1.07 ...... 133.3 ...... 126.7 ...... 123.4 ...... 6599.13 Ti i 0.90 -2.09 72.4 93.8 141.2 166.2 117.3 136.8 144.5 178.2 ... 154.9 138.3 75.9 119.3 ... 115.4 ... 6666.53 Ti i 1.46 -1.62 ... 29.9 ... 19.5 ... 18.8 ... 22.5 24.3 42.7 ... 16.2 ... 30.3 20.7 ... 5418.77 Ti ii 1.58 -2.11 95.7 108.8 110.7 113.7 102.1 112.0 124.6 113.1 89.6 101.4 107.3 94.0 94.2 104.3 100.3 97.9 6219.94 Ti ii 2.06 -2.82 ... 36.6 ...... 24.0 ... 29.5 ... 31.7 29.9 19.0 20.8 ...... 6559.58 Ti ii 2.05 -2.02 ...... 66.8 ...... 50.8 ...... 6606.95 Ti ii 2.06 -2.79 36.6 43.4 43.2 59.3 41.4 21.7 30.1 49.3 62.8 41.2 ... 41.6 34.3 61.9 ... 116.7 6680.13 Ti ii 3.09 -1.86 ... 30.1 ... 41.2 35.1 ... 33.8 46.0 27.6 18.9 24.0 27.8 23.6 42.1 29.4 41.7 6119.53 V i 1.06 -0.32 ...... 95.2 ... 85.0 83.8 ...... 71.8 ...... 6128.33 V i 1.05 -2.30 ...... 28.4 ...... 6135.37 V i 1.05 -0.75 59.0 74.2 98.4 125.6 94.6 92.1 90.6 120.3 88.8 120.2 98.7 44.8 76.0 116.9 107.6 79.7 6150.15 V i 0.30 -1.79 96.1 122.6 143.3 167.4 136.1 150.2 144.5 172.1 95.1 165.9 131.1 69.0 123.5 153.3 143.5 140.1 6199.19 V i 0.29 -1.29 111.7 150.8 204.2 202.9 166.0 178.2 177.3 214.4 144.4 195.4 186.9 101.1 155.0 180.9 193.7 ... 6216.36 V i 0.28 -0.81 128.8 168.3 182.5 194.6 176.4 177.0 164.9 199.8 166.2 203.9 184.0 119.0 150.0 196.7 182.8 178.4 6224.51 V i 0.29 -2.01 73.1 101.4 134.9 147.3 121.0 142.6 130.0 147.7 101.6 150.4 120.1 72.8 99.1 136.0 136.8 108.5 6233.20 V i 0.28 -2.07 39.5 73.3 100.9 96.7 81.9 93.7 88.3 100.8 66.8 108.1 91.8 29.3 72.9 104.0 100.5 72.9 6243.11 V i 0.30 -0.98 ... 234.4 ...... 284.4 ...... 177.1 ...... 6251.82 V i 0.29 -1.30 97.9 116.9 144.0 149.8 144.7 144.5 135.5 158.8 128.1 172.4 147.0 83.6 130.8 161.6 151.6 134.5 6274.66 V i 0.27 -1.67 65.1 108.5 133.7 132.8 ...... 134.1 ... 127.4 ... 57.2 ... 128.9 119.2 ... 6357.29 V i 1.85 -0.91 ... 23.3 25.5 30.9 31.4 ...... 35.0 ... 35.2 20.4 19.6 ...... 26.4 ... 6452.32 V i 1.19 -1.21 18.9 50.7 102.7 124.9 86.3 99.2 97.0 115.4 ... 110.9 105.5 32.6 64.3 ... 66.7 ... 6504.19 V i 1.18 -1.23 18.6 35.6 56.3 68.4 48.3 47.8 58.9 53.9 44.1 42.7 47.6 15.5 33.7 64.6 54.6 46.0 6531.41 V i 1.22 -0.84 ...... 80.0 ...... 52.5 ...... 5402.78 Y ii 1.84 -0.51 ... 34.8 40.5 51.6 43.8 27.1 28.8 35.0 58.2 47.6 21.6 ... 36.3 53.9 30.6 26.6 6362.35 Zn i 5.80 0.14 ...... 22.5 ...... 17.2 ...... 6127.48 Zr i 0.15 -1.06 ... 75.5 77.0 104.5 77.4 81.4 76.3 101.7 65.5 88.5 72.5 ... 69.4 87.8 73.8 ... 6140.46 Zr i 0.52 -1.41 ...... 16.7 ... 19.2 ...... 19.8 27.5 ...... 26.2 27.3 ... 6143.18 Zr i 0.07 -1.10 24.9 66.9 81.3 90.2 79.4 84.7 71.4 108.8 81.7 105.4 70.4 26.9 55.3 102.3 85.2 54.1 6192.95 Zr i 0.54 -2.07 ...... 23.7 23.8 ...... 19.6 ...... 132 chapter 6: HR spectroscopic study of Fornax Field Stars

Table 6.A2: Complete line list with parameters and associated EW s (in mÅ, measured by DAOSPEC) for all the stars. Part 4/5.

Equivalent witdh, one star per column, BLxxx λ(Å) elem χ gf 205 208 210 211 213 216 218 221 227 228 229 233 239 242 247 250 6141.73 Ba ii 0.70 -0.08 267.8 259.2 265.1 271.8 252.9 282.6 ... 251.0 244.0 215.3 275.9 261.5 228.5 226.7 250.1 ... 6496.91 Ba ii 0.60 -0.38 243.1 270.3 284.0 297.9 235.1 281.4 280.7 256.7 268.8 232.7 284.4 247.9 258.8 250.3 246.8 ... 6122.23 Ca i 1.89 -0.32 244.8 250.7 233.3 279.0 246.7 276.6 ... 247.1 230.7 248.7 267.5 242.6 228.3 213.3 250.5 ... 6156.03 Ca i 2.52 -2.39 ... 26.0 25.6 36.6 24.6 42.3 39.8 25.7 26.1 35.5 42.9 26.5 20.7 28.5 36.9 38.6 6161.30 Ca i 2.52 -1.27 150.3 142.6 145.4 155.4 137.5 160.3 163.3 127.7 136.1 129.6 143.4 141.2 108.0 106.0 145.6 167.0 6162.17 Ca i 1.90 -0.32 261.5 273.6 274.5 ... 268.1 ...... 255.5 266.5 269.6 289.8 285.5 249.2 237.1 279.7 ... 6166.44 Ca i 2.52 -1.14 133.0 132.0 119.9 146.1 117.2 136.1 148.2 117.9 117.6 123.8 158.7 121.9 118.0 96.5 133.1 174.5 6169.04 Ca i 2.52 -0.80 144.1 155.5 133.2 156.5 152.7 149.2 148.9 144.0 126.5 150.0 146.1 147.2 150.0 95.7 147.2 170.9 6169.56 Ca i 2.52 -0.48 164.3 170.2 176.8 167.0 159.8 171.8 183.3 167.8 169.8 153.2 176.3 164.3 146.0 117.5 169.5 210.3 6439.08 Ca i 2.52 0.39 225.4 228.8 231.7 234.7 227.2 ... 243.2 205.2 210.7 231.2 235.4 235.6 224.7 200.1 237.1 243.9 6455.60 Ca i 2.52 -1.29 192.1 161.5 184.7 113.3 117.5 ...... 57.1 102.5 133.8 175.2 185.3 175.1 123.0 163.5 222.9 6471.67 Ca i 2.52 -0.76 ...... 6493.79 Ca i 2.52 -0.32 195.2 181.3 191.7 191.3 182.7 222.4 223.7 179.6 178.8 189.8 208.7 196.6 180.4 160.4 163.1 190.3 6499.65 Ca i 2.52 -0.82 111.5 122.6 135.9 140.6 124.9 153.6 145.6 115.7 124.0 116.1 137.1 126.5 126.6 125.1 121.6 159.1 6508.84 Ca i 2.52 -2.41 31.7 36.6 32.1 50.9 25.0 67.1 58.0 36.9 24.0 18.6 41.6 33.1 22.5 16.0 25.5 67.4 6330.09 Cr i 0.94 -2.92 122.0 122.3 132.8 160.3 130.2 161.8 160.6 133.3 138.7 150.4 142.5 132.2 116.1 120.3 147.0 178.8 6645.13 Eu ii 1.37 0.20 57.2 52.8 72.2 74.0 68.7 86.3 68.4 83.5 76.2 50.7 80.2 65.1 40.1 47.2 66.4 62.5 5369.96 Fe i 4.37 0.54 167.4 163.5 164.9 168.3 147.9 175.4 146.3 201.7 148.9 217.2 ... 152.5 148.5 134.5 194.3 199.0 5383.37 Fe i 4.31 0.50 181.5 162.1 170.0 164.5 169.7 172.1 204.0 171.3 166.6 166.8 165.7 169.5 145.3 155.0 182.6 177.1 5386.34 Fe i 4.16 -1.74 60.5 52.5 53.0 ... 53.7 55.1 59.5 ... 22.1 46.8 51.8 61.5 72.8 25.3 60.5 ... 5393.17 Fe i 3.24 -0.92 221.4 227.6 200.5 209.7 222.4 226.2 211.3 211.4 210.9 209.1 231.5 230.0 224.4 201.2 222.3 246.8 5395.22 Fe i 4.45 -1.73 25.6 24.3 24.7 37.6 ... 31.9 36.1 ... 44.0 35.6 19.7 40.2 ...... 36.0 5405.79 Fe i 0.99 -1.85 ...... 5415.19 Fe i 4.39 0.51 188.5 163.3 179.9 185.6 175.6 172.9 163.6 144.5 153.8 173.8 172.0 164.7 179.3 133.7 189.9 170.1 5417.04 Fe i 4.42 -1.42 47.8 44.0 45.5 36.5 28.8 40.2 49.6 32.7 49.8 44.8 34.4 46.3 67.8 27.4 58.3 42.6 5434.53 Fe i 1.01 -2.12 ...... 5436.30 Fe i 4.39 -1.35 ...... 5464.29 Fe i 4.14 -1.62 78.2 81.2 59.0 ... 56.1 88.6 89.9 62.3 61.6 54.0 81.9 85.5 55.0 61.4 52.1 ... 5470.09 Fe i 4.45 -1.60 24.2 28.7 ... 51.0 29.4 55.1 32.4 20.7 39.1 16.2 ... 30.9 22.1 ... 26.7 57.7 5501.48 Fe i 0.96 -3.05 250.9 257.7 261.3 265.2 254.2 274.4 297.6 242.7 230.1 273.8 279.0 263.5 260.7 238.4 274.6 ... 5506.79 Fe i 0.99 -2.79 ...... 275.1 ...... 297.8 282.4 ...... 5539.29 Fe i 3.64 -2.59 75.1 72.4 73.9 82.0 61.6 74.2 95.8 58.1 60.1 54.3 78.0 71.8 56.2 36.9 78.9 86.2 5586.77 Fe i 3.37 -0.10 239.6 237.1 247.2 236.5 236.5 222.0 249.5 214.1 192.1 235.0 242.8 238.3 231.8 192.0 279.5 288.0 6120.26 Fe i 0.91 -5.94 78.6 78.7 89.7 107.2 105.7 96.8 113.2 88.6 75.8 100.8 97.2 82.1 82.5 63.2 87.8 112.1 6136.62 Fe i 2.45 -1.50 ...... 288.4 ...... 255.6 ...... 6137.00 Fe i 2.20 -2.95 ...... 164.4 ...... 174.1 ...... 164.9 ...... 6151.62 Fe i 2.18 -3.37 134.8 134.6 126.9 147.0 136.5 142.0 141.4 124.7 137.7 139.2 133.4 150.7 125.6 115.7 147.1 146.2 6157.75 Fe i 4.07 -1.26 122.8 123.1 115.8 127.0 118.0 124.3 132.6 108.9 116.6 102.5 125.4 134.5 109.0 103.9 121.7 109.4 6159.38 Fe i 4.61 -1.97 16.6 ... 16.8 34.0 ...... 40.8 19.3 ... 26.7 ... 16.9 ...... 6165.36 Fe i 4.14 -1.47 65.1 72.9 73.8 76.1 68.3 70.3 74.1 68.4 64.8 78.8 77.9 80.4 58.5 52.3 68.7 97.4 6173.34 Fe i 2.22 -2.85 167.9 171.2 159.8 156.2 160.9 166.2 179.1 166.7 159.6 165.3 179.0 179.5 150.8 156.8 177.6 175.4 6180.20 Fe i 2.73 -2.78 142.3 135.5 158.9 133.2 137.1 131.0 141.5 122.1 107.0 133.3 143.2 139.8 138.4 87.1 141.0 144.5 6187.99 Fe i 3.94 -1.58 80.3 86.8 97.2 85.8 78.7 80.5 85.4 77.2 66.0 78.5 77.8 70.5 78.7 60.8 92.8 94.0 6200.31 Fe i 2.61 -2.44 157.3 150.8 157.3 150.1 148.1 152.8 147.3 154.7 129.6 156.9 159.5 155.7 159.0 141.7 175.0 138.3 6213.43 Fe i 2.22 -2.66 179.4 184.1 192.5 182.8 191.9 187.1 183.4 196.9 200.1 189.3 198.5 187.0 168.8 177.8 200.1 184.1 6219.29 Fe i 2.20 -2.44 201.7 204.1 198.7 203.1 192.4 204.3 190.6 194.0 198.8 208.5 200.7 198.6 195.0 191.2 223.6 198.7 6226.74 Fe i 3.88 -2.20 71.7 72.4 49.0 70.7 75.1 69.2 70.5 49.4 53.9 59.6 76.6 69.8 53.1 47.5 51.8 87.6 6252.57 Fe i 2.40 -1.76 202.2 196.9 204.6 200.4 200.6 217.1 189.9 193.4 207.3 206.2 203.2 200.5 208.4 189.7 205.0 196.3 6265.13 Fe i 2.18 -2.55 181.0 179.9 190.6 183.7 167.1 189.0 201.1 177.6 181.4 192.7 182.8 173.4 175.5 171.9 188.3 213.4 6271.28 Fe i 3.32 -2.96 71.3 59.8 84.7 81.2 68.9 73.8 71.0 ... 56.2 58.1 55.0 63.8 60.7 50.9 51.9 65.4 6297.80 Fe i 2.22 -2.74 ...... 6301.50 Fe i 3.65 -0.72 ...... 6307.85 Fe i 3.64 -3.27 ...... 26.1 26.6 ...... 18.2 ... 21.8 ...... 6322.69 Fe i 2.59 -2.43 152.4 155.6 162.8 151.3 156.4 173.8 168.2 154.4 153.9 179.7 156.8 162.6 153.0 142.6 172.5 170.5 6330.85 Fe i 4.73 -1.22 51.9 63.2 45.0 54.3 37.3 49.6 43.7 35.7 34.3 33.5 52.2 ... 39.5 29.9 52.5 26.0 6335.33 Fe i 2.20 -2.23 196.9 216.0 218.7 213.1 216.5 233.3 228.9 206.0 211.5 208.0 226.1 201.3 203.5 202.1 209.1 233.0 6336.82 Fe i 3.69 -1.05 161.6 143.0 160.9 147.3 143.2 155.8 149.3 141.8 143.2 149.9 155.6 142.5 142.7 128.4 158.8 145.1 6344.15 Fe i 2.43 -2.92 187.9 191.9 179.6 195.9 185.9 191.7 200.8 170.1 170.1 174.4 189.4 196.4 159.6 145.3 191.2 204.8 6355.04 Fe i 2.84 -2.29 162.9 164.4 151.7 171.9 170.8 178.9 171.9 152.1 150.7 163.7 139.8 169.0 146.2 136.6 158.0 168.8 6380.75 Fe i 4.19 -1.50 97.3 102.2 89.6 77.1 87.3 90.2 103.2 85.2 83.9 89.3 75.3 93.9 77.4 68.0 87.3 85.8 6392.54 Fe i 2.28 -3.95 103.5 92.9 96.2 94.8 95.1 106.7 110.1 85.6 83.0 93.5 96.7 107.3 80.2 75.7 102.6 97.7 6393.61 Fe i 2.43 -1.63 228.9 245.7 256.1 259.2 238.7 263.7 256.5 239.2 233.3 253.0 248.8 238.4 236.2 231.7 264.9 280.4 6408.03 Fe i 3.69 -1.00 159.6 144.2 161.6 160.2 152.4 ... 173.0 129.1 125.9 160.8 158.2 156.8 146.4 140.1 167.5 175.0 6419.96 Fe i 4.73 -0.24 118.5 116.0 130.2 122.3 102.3 ... 118.8 109.5 111.5 92.4 135.7 117.2 116.6 108.2 120.5 124.7 6421.36 Fe i 2.28 -2.01 224.1 208.2 230.2 230.3 226.3 ... 228.7 222.5 209.4 229.7 235.3 226.7 221.9 212.6 233.4 272.4 6430.86 Fe i 2.18 -1.95 236.7 247.0 276.1 274.0 247.2 ... 293.5 245.6 256.8 260.1 252.5 264.6 248.3 228.3 235.1 ... 6481.87 Fe i 2.27 -2.98 163.5 167.5 181.6 189.0 155.5 207.5 180.3 173.3 157.6 172.0 159.7 174.4 171.4 161.4 164.2 180.8 6498.94 Fe i 0.96 -4.69 177.3 190.6 201.9 216.9 187.1 219.8 220.2 183.3 181.2 197.4 192.1 191.6 163.4 173.1 183.1 235.5 6533.93 Fe i 4.55 -1.46 46.9 54.7 54.1 42.5 43.7 ... 47.6 46.8 48.5 54.7 61.4 51.9 50.7 41.8 49.1 37.1 6556.81 Fe i 4.79 -1.72 ... 16.9 23.8 24.3 ... 25.3 ... 22.8 25.6 19.6 ... 16.5 ...... 6569.22 Fe i 4.73 -0.42 126.2 122.2 124.0 117.1 109.5 139.8 124.8 98.3 111.5 96.3 128.5 130.1 100.2 88.2 101.6 128.6 6574.23 Fe i 0.99 -5.02 161.1 184.0 205.3 194.7 168.4 ... 200.7 159.5 153.3 190.8 196.9 180.7 171.2 ...... 6593.88 Fe i 2.43 -2.39 206.4 203.9 248.3 195.4 197.4 ... 238.5 162.8 128.0 215.0 220.9 202.0 196.8 171.6 231.3 249.6 6597.56 Fe i 4.79 -1.07 74.0 75.2 75.1 68.6 57.8 ... 73.9 55.4 35.0 66.5 72.9 74.8 67.0 43.8 69.7 78.2 6608.03 Fe i 2.28 -3.94 98.4 104.0 97.6 114.9 88.4 133.3 112.3 103.0 92.2 103.9 98.3 106.2 105.2 76.0 94.4 115.6 6609.12 Fe i 2.56 -2.66 175.7 175.4 164.2 169.7 167.0 198.7 177.0 166.1 195.2 172.0 156.1 161.2 162.4 165.2 158.6 174.1 6627.54 Fe i 4.54 -1.68 26.7 37.4 33.3 35.7 30.9 48.2 30.0 34.7 37.8 22.6 36.9 36.5 15.9 20.4 27.0 23.8 6633.76 Fe i 4.56 -0.82 125.1 120.0 126.9 118.1 113.7 148.6 115.4 110.8 118.0 108.7 134.8 122.3 97.5 93.4 126.1 105.0 6646.93 Fe i 2.60 -3.99 54.5 61.8 61.7 64.9 48.6 80.3 71.3 55.6 69.8 54.7 70.4 70.2 45.7 42.9 63.6 53.7 6653.85 Fe i 4.15 -2.52 23.7 22.9 39.1 35.4 26.0 41.2 44.8 34.8 41.6 ... 35.6 30.7 39.7 35.2 27.5 58.5 5414.08 Fe ii 3.22 -3.61 34.7 ... 39.0 ... 22.5 ...... 25.8 30.9 ... 46.9 ... 44.2 ... 44.3 ... 5425.25 Fe ii 3.20 -3.36 64.5 47.6 46.0 19.3 40.5 27.1 36.6 54.5 40.7 32.9 31.1 42.0 36.1 56.6 69.4 37.9 6149.25 Fe ii 3.89 -2.72 35.3 ... 30.5 29.9 36.7 30.4 ... 19.7 ... 29.8 ... 33.1 ...... 31.3 38.3 6432.68 Fe ii 2.89 -3.71 62.8 48.8 58.5 37.8 51.7 ... 36.2 50.3 48.3 47.9 52.3 59.5 53.8 56.8 39.8 39.5 6456.39 Fe ii 3.90 -2.08 86.2 63.5 124.5 45.8 63.8 ... 78.3 22.8 ... 84.1 70.7 80.2 76.2 39.9 76.7 127.7 6320.43 La ii 0.17 -1.56 80.1 94.6 90.4 96.5 81.2 107.2 106.3 83.2 65.3 51.3 102.2 87.0 41.0 47.9 88.7 92.9 6390.46 La ii 0.32 -1.40 70.7 77.9 88.4 85.7 79.7 99.8 93.8 98.9 69.8 56.5 87.1 67.6 49.2 66.5 78.1 96.8 5528.41 Mg i 4.35 -0.36 217.6 216.7 209.4 211.1 200.0 231.6 234.9 215.0 205.8 215.0 234.1 214.4 204.7 209.8 220.6 230.1 Continued on next page 6.A: Large tables 133

λ(Å) elem χ gf 205 208 210 211 213 216 218 221 227 228 229 233 239 242 247 250 6318.72 Mg i 5.11 -1.97 46.5 41.8 55.4 46.9 39.1 51.1 47.3 ...... 31.8 35.8 44.6 ... 41.6 ... 58.6 6319.24 Mg i 5.11 -2.21 ... 36.2 ... 23.7 ... 25.3 33.9 19.2 29.8 ...... 19.4 ... 43.4 34.6 6319.49 Mg i 5.11 -2.43 ...... 35.7 ...... 24.0 ... 36.0 ...... 5420.36 Mn i 2.14 -1.46 222.1 201.1 213.4 230.0 191.8 206.6 221.1 185.9 155.9 217.5 222.2 204.7 224.6 180.7 226.2 211.8 5432.55 Mn i 0.00 -3.80 248.6 272.1 276.2 294.5 270.2 297.9 ... 237.3 266.0 278.9 274.8 271.8 257.9 239.2 281.5 ... 5516.77 Mn i 2.18 -1.85 158.4 ... 175.1 172.6 141.3 191.9 182.3 140.3 160.1 ... 178.1 160.3 126.4 131.3 ... 157.8 6154.23 Na i 2.10 -1.56 37.4 38.0 36.5 40.5 ... 42.5 27.2 ... 33.8 16.6 29.5 37.6 16.8 ... 24.6 44.5 6160.75 Na i 2.10 -1.26 50.6 51.6 51.1 57.3 40.1 65.8 60.3 27.5 48.6 29.5 51.5 63.1 ... 28.9 54.0 88.7 5416.38 Nd ii 0.86 -0.98 ... 38.9 50.4 43.3 36.4 40.5 40.0 ... 34.3 ... 21.4 22.1 22.8 ... 57.5 65.4 5431.54 Nd ii 1.12 -0.47 65.2 60.6 57.3 68.0 ... 66.6 73.5 69.4 90.6 47.5 64.9 52.6 30.8 66.1 86.5 66.9 5485.71 Nd ii 1.26 -0.12 37.2 40.6 45.4 54.2 41.9 58.1 54.4 50.7 36.4 ... 53.4 51.7 25.4 34.8 ... 36.7 5578.73 Ni i 1.68 -2.67 142.9 143.8 160.6 142.6 142.5 148.6 164.2 121.6 123.3 150.8 145.0 142.7 129.9 129.3 135.9 184.6 5587.87 Ni i 1.93 -2.37 171.5 168.7 191.9 154.0 154.7 177.2 172.2 131.3 166.7 165.8 184.2 168.3 163.9 121.0 180.6 177.0 5589.37 Ni i 3.90 -1.15 27.6 35.4 31.8 32.0 26.8 28.3 37.3 ... 26.3 20.2 20.1 26.7 19.7 17.8 ...... 5593.75 Ni i 3.90 -0.79 51.5 56.2 59.7 43.0 35.4 39.2 43.3 32.4 45.5 35.0 38.1 51.7 29.4 38.9 45.0 49.4 6128.97 Ni i 1.68 -3.39 78.6 85.2 93.6 97.5 93.8 106.0 104.8 90.8 71.8 92.9 90.9 99.0 78.8 79.3 90.2 93.8 6130.14 Ni i 4.27 -0.98 17.0 19.7 ... 19.2 ...... 33.2 ...... 25.6 ... 6177.25 Ni i 1.83 -3.60 57.7 68.3 77.5 73.0 58.1 79.0 61.1 54.6 55.8 61.7 76.6 66.2 60.1 45.9 78.2 63.9 6186.72 Ni i 4.11 -0.90 33.5 33.6 40.0 39.3 38.9 50.3 52.2 29.5 22.0 30.9 32.3 30.2 25.1 25.5 41.4 86.0 6204.61 Ni i 4.09 -1.15 28.4 23.7 ... 20.2 25.2 ... 22.5 16.8 23.9 ... 25.1 27.2 27.4 ... 30.5 ... 6223.99 Ni i 4.10 -0.97 45.0 44.0 41.2 44.4 34.7 28.8 44.0 39.0 ... 25.2 40.3 32.3 26.2 31.1 48.3 61.9 6230.10 Ni i 4.11 -1.20 26.8 30.5 39.3 ... 26.7 45.1 ... 15.2 ... 28.3 ... 32.0 ... 24.3 27.2 46.6 6322.17 Ni i 4.15 -1.21 23.4 26.4 16.4 29.0 21.4 ...... 15.2 ...... 6327.60 Ni i 1.68 -3.09 121.5 118.8 124.0 125.6 103.0 131.3 129.8 116.6 109.6 112.7 116.5 109.8 110.1 97.3 129.5 130.5 6378.26 Ni i 4.15 -0.82 33.2 25.4 23.3 40.5 33.4 32.9 50.0 30.1 31.3 30.8 32.5 34.2 22.8 ... 38.5 47.2 6384.67 Ni i 4.15 -1.00 ...... 29.8 47.9 ... 38.7 45.6 ... 38.8 ... 37.7 47.0 ... 33.3 58.1 43.2 6482.80 Ni i 1.94 -2.85 103.8 99.7 121.0 116.2 86.2 127.1 117.2 104.6 101.6 105.6 97.7 113.8 103.5 100.4 110.0 125.8 6586.32 Ni i 1.95 -2.79 142.5 165.3 179.6 139.6 154.3 ... 157.2 80.0 58.3 150.4 174.5 155.2 154.8 120.3 185.7 185.8 6598.61 Ni i 4.24 -0.93 ... 30.9 ...... 33.8 17.0 ... 17.8 ...... 42.1 ... 6635.14 Ni i 4.42 -0.75 ... 17.0 ...... 36.7 ... 31.8 23.1 ...... 20.1 6300.31 O i 0.00 -9.75 ...... 64.4 ... 65.4 6363.79 O i 0.02 -10.25 ... 39.7 39.4 47.0 44.8 51.8 51.7 34.5 39.3 42.8 50.7 31.8 27.7 37.2 41.5 54.8 5526.82 Sc ii 1.77 0.03 112.9 110.3 103.8 104.6 115.0 123.4 99.0 108.6 113.0 100.1 117.8 102.3 118.4 101.7 114.5 103.8 6245.62 Sc ii 1.51 -0.97 80.1 72.5 64.2 69.6 77.7 ... 69.7 67.1 76.2 64.0 67.8 65.6 61.8 65.4 74.5 69.5 6309.90 Sc ii 1.50 -1.52 33.8 ...... 46.4 58.7 57.0 ... 57.1 22.3 44.7 47.9 ...... 6604.60 Sc ii 1.36 -1.31 87.2 98.9 97.4 100.1 93.4 111.0 98.2 109.9 84.6 92.1 85.0 89.8 87.5 78.4 75.7 95.7 6125.03 Si i 5.62 -1.57 41.7 ...... 48.4 ...... 16.4 ...... 6142.48 Si i 5.62 -1.51 ...... 18.5 ... 21.8 ...... 6145.02 Si i 5.61 -1.37 ...... 27.6 19.2 ...... 23.1 34.5 ...... 6155.14 Si i 5.62 -0.80 57.4 45.8 40.5 45.3 43.8 51.5 50.1 43.2 54.9 31.3 42.5 45.8 32.7 45.0 42.3 33.1 6237.33 Si i 5.61 -1.02 32.5 28.2 18.4 23.9 21.0 24.3 18.1 23.3 25.5 ... 27.3 19.1 23.2 23.2 ... 37.4 6243.82 Si i 5.61 -1.27 41.7 47.4 35.3 56.3 29.1 ...... 29.2 26.3 46.1 33.8 23.5 ... 37.5 40.2 5490.16 Ti i 1.46 -0.93 100.1 107.3 122.6 134.3 112.7 135.9 148.3 109.9 112.4 105.2 123.3 134.1 93.7 95.3 133.0 140.3 5503.90 Ti i 2.58 -0.19 67.3 70.1 71.8 83.0 68.6 88.7 89.3 57.4 44.5 58.7 83.7 79.9 43.2 39.5 71.4 109.5 6126.22 Ti i 1.07 -1.42 122.8 150.5 134.6 164.4 145.3 169.4 170.1 134.9 130.7 143.2 151.8 140.1 118.9 117.5 159.2 189.2 6220.50 Ti i 2.68 -0.14 ...... 87.8 ...... 86.3 ...... 52.1 ...... 65.2 6258.10 Ti i 1.44 -0.35 151.1 182.8 187.5 207.9 184.1 228.0 232.0 170.1 199.9 212.6 206.9 178.7 150.3 142.5 153.9 297.9 6303.77 Ti i 1.44 -1.57 77.5 89.4 88.6 118.9 93.1 118.0 115.2 76.7 78.8 86.2 112.3 91.1 80.3 67.8 91.0 122.5 6312.24 Ti i 1.46 -1.55 ... 87.6 93.7 109.2 ...... 73.4 76.1 92.9 ... 91.8 61.6 66.7 74.4 130.0 6336.10 Ti i 1.44 -1.74 67.5 80.4 76.9 94.4 80.4 96.4 98.4 69.0 79.1 77.2 84.1 84.7 55.2 40.0 84.6 128.3 6508.12 Ti i 1.43 -2.05 31.5 45.7 59.1 81.0 57.5 86.4 84.7 45.7 66.1 52.5 66.3 65.7 26.5 ... 46.9 110.6 6556.08 Ti i 1.46 -1.07 ...... 133.0 ...... 6599.13 Ti i 0.90 -2.09 111.3 120.8 153.8 181.9 142.2 ... 186.1 101.0 113.8 148.0 156.7 132.8 110.5 99.0 149.9 221.2 6666.53 Ti i 1.46 -1.62 19.2 34.5 34.7 36.1 ... 41.7 38.9 ...... 35.8 18.4 15.1 ... 37.9 59.0 5418.77 Ti ii 1.58 -2.11 111.5 96.6 114.0 97.2 111.6 101.5 97.4 103.3 95.0 100.9 109.3 104.3 82.5 80.7 98.5 99.3 6219.94 Ti ii 2.06 -2.82 36.1 ...... 41.7 37.2 ... 28.1 40.3 ... 6559.58 Ti ii 2.05 -2.02 ...... 49.2 72.9 54.2 ...... 6606.95 Ti ii 2.06 -2.79 45.7 49.4 43.0 52.7 40.4 67.7 35.3 55.5 45.0 40.3 25.8 51.0 34.4 48.0 27.7 23.8 6680.13 Ti ii 3.09 -1.86 36.3 ... 29.2 19.9 22.4 25.4 18.2 26.7 ... 28.0 21.7 27.1 ...... 31.5 ... 6119.53 V i 1.06 -0.32 ...... 96.7 ...... 6128.33 V i 1.05 -2.30 ...... 17.8 ...... 41.6 6135.37 V i 1.05 -0.75 80.7 100.2 104.6 133.1 102.6 136.0 140.1 89.2 106.0 114.3 122.1 104.1 ... 87.5 89.4 150.5 6150.15 V i 0.30 -1.79 120.7 126.9 138.9 158.2 148.7 173.3 182.5 122.2 151.2 162.3 152.9 149.4 112.1 106.3 147.3 198.5 6199.19 V i 0.29 -1.29 155.0 166.1 195.5 221.2 177.2 217.9 212.1 154.2 177.1 186.7 215.9 196.5 144.0 141.5 186.4 232.6 6216.36 V i 0.28 -0.81 150.6 174.7 188.8 200.7 183.0 219.5 213.9 166.7 204.3 195.9 198.9 195.9 152.6 162.2 180.1 225.7 6224.51 V i 0.29 -2.01 107.7 122.9 132.0 159.3 142.3 154.2 161.2 120.2 ... 130.9 148.4 140.2 91.9 111.0 141.8 196.7 6233.20 V i 0.28 -2.07 66.4 91.8 95.3 108.2 88.2 128.0 134.0 94.6 99.7 101.9 112.8 94.1 63.2 69.7 81.0 130.6 6243.11 V i 0.30 -0.98 ...... 281.0 ...... 6251.82 V i 0.29 -1.30 112.2 139.4 140.5 169.3 140.3 177.0 160.9 137.1 155.4 156.3 161.9 141.1 116.6 117.8 135.5 164.7 6274.66 V i 0.27 -1.67 ... 111.8 126.9 142.9 ...... 109.8 118.2 130.0 ...... 92.2 79.0 107.5 156.5 6357.29 V i 1.85 -0.91 17.3 27.6 33.7 51.5 27.1 45.6 49.2 15.4 16.9 26.8 36.8 39.5 ... 20.6 34.0 69.3 6452.32 V i 1.19 -1.21 68.2 80.4 114.3 123.4 90.4 ... 137.6 62.4 64.5 81.1 103.8 98.1 69.6 45.7 ... 167.8 6504.19 V i 1.18 -1.23 38.1 42.8 50.1 73.6 60.8 87.5 86.4 49.6 50.1 44.5 72.3 50.7 45.2 41.3 39.0 73.2 6531.41 V i 1.22 -0.84 ...... 70.3 ... 97.2 ...... 84.6 ...... 5402.78 Y ii 1.84 -0.51 44.6 24.4 45.0 58.4 53.7 55.9 55.2 40.5 50.7 29.3 26.5 60.0 ... 32.0 74.5 49.2 6362.35 Zn i 5.80 0.14 ...... 6127.48 Zr i 0.15 -1.06 63.0 81.6 92.7 122.1 84.2 114.8 132.5 93.4 78.8 68.7 104.0 87.2 45.2 56.2 80.3 132.5 6140.46 Zr i 0.52 -1.41 ... 16.4 ... 27.9 ... 31.8 35.0 15.9 ...... 28.7 ...... 50.9 6143.18 Zr i 0.07 -1.10 59.9 77.9 75.0 121.3 79.9 127.9 113.2 83.8 98.1 84.3 110.9 85.8 52.4 59.6 67.4 136.3 6192.95 Zr i 0.54 -2.07 ... 17.4 20.1 21.9 15.5 21.4 19.6 ...... 34.6 134 chapter 6: HR spectroscopic study of Fornax Field Stars

Table 6.A2: Complete line list with parameters and associated EW s (in mÅ, measured by DAOSPEC) for all the stars. Part 5/5.

Equivalent witdh, one star per column, BLxxx λ(Å) elem χ gf 253 257 258 260 261 262 266 267 269 278 279 295 300 304 311 315 323 6141.73 Ba ii 0.70 -0.08 287.8 ... 297.3 232.5 218.4 230.4 173.6 248.7 246.2 ... 167.2 ... 293.7 261.3 253.0 215.4 260.3 6496.91 Ba ii 0.60 -0.38 282.5 290.1 ... 228.3 202.2 266.1 185.8 237.1 273.8 ... 160.6 ... 265.7 244.0 ... 193.9 264.9 6122.23 Ca i 1.89 -0.32 279.3 291.0 281.2 242.9 236.3 ... 182.2 223.9 258.7 ... 174.8 ... 275.1 271.8 247.5 201.6 252.6 6156.03 Ca i 2.52 -2.39 38.4 46.8 43.7 32.1 20.5 ...... 32.0 28.5 45.4 ...... 17.5 15.6 36.5 ...... 6161.30 Ca i 2.52 -1.27 163.2 166.7 180.6 127.1 127.9 113.3 50.6 132.8 132.2 177.4 47.0 161.4 145.4 143.2 138.3 87.7 148.1 6162.17 Ca i 1.90 -0.32 ...... 278.8 266.5 256.1 196.0 270.4 282.6 ... 184.9 ... 291.6 288.4 277.4 221.8 282.1 6166.44 Ca i 2.52 -1.14 148.6 150.4 150.5 136.3 125.4 121.0 64.0 146.7 124.0 147.9 37.8 157.0 135.9 130.7 128.6 103.2 113.8 6169.04 Ca i 2.52 -0.80 163.3 169.7 161.6 145.9 135.6 129.1 94.4 148.5 144.6 159.9 64.5 154.9 152.0 149.4 133.3 123.6 105.8 6169.56 Ca i 2.52 -0.48 170.4 196.1 182.1 165.5 140.1 154.1 112.4 168.4 151.7 177.3 86.8 196.2 176.1 171.7 171.3 156.6 146.5 6439.08 Ca i 2.52 0.39 236.5 223.4 247.4 213.6 209.2 230.1 183.4 225.8 215.5 238.4 155.1 260.0 224.0 228.4 ... 207.2 235.2 6455.60 Ca i 2.52 -1.29 84.2 113.4 187.4 152.1 ...... 29.3 153.2 75.5 170.9 ... 192.0 ... 130.8 ... 155.9 60.6 6471.67 Ca i 2.52 -0.76 ...... 6493.79 Ca i 2.52 -0.32 209.7 223.2 189.5 187.3 182.2 181.5 139.8 196.4 194.6 211.4 126.6 199.2 210.3 205.2 204.5 156.9 195.8 6499.65 Ca i 2.52 -0.82 144.6 159.2 138.6 118.5 121.0 133.3 77.6 106.5 138.1 151.2 45.9 141.8 127.3 131.3 152.9 105.2 126.8 6508.84 Ca i 2.52 -2.41 45.2 57.3 55.8 38.7 32.4 43.9 ... 37.1 46.5 61.5 ... 57.0 29.9 28.7 38.6 ... 41.5 6330.09 Cr i 0.94 -2.92 152.8 162.8 163.2 142.5 123.5 130.7 64.1 132.6 130.4 172.1 35.2 159.7 150.2 140.8 137.0 106.9 158.7 6645.13 Eu ii 1.37 0.20 75.8 74.3 90.6 51.4 62.8 54.1 30.6 64.9 82.4 78.2 34.6 82.9 61.2 59.4 113.1 75.5 70.1 5369.96 Fe i 4.37 0.54 171.7 171.0 180.1 182.4 172.0 179.4 112.4 164.9 141.9 247.5 101.7 208.0 164.4 158.8 238.5 145.2 165.1 5383.37 Fe i 4.31 0.50 153.4 169.7 181.0 171.1 145.9 149.5 130.6 135.1 153.1 182.5 93.5 173.8 187.3 160.0 139.8 129.9 162.6 5386.34 Fe i 4.16 -1.74 ... 65.9 ...... 45.7 33.1 ... 56.6 55.8 ... 17.7 57.3 ... 35.6 39.6 58.9 35.8 5393.17 Fe i 3.24 -0.92 226.5 248.5 235.0 210.7 213.9 192.2 169.3 229.2 204.2 235.2 125.3 246.4 218.6 224.3 198.2 210.6 178.5 5395.22 Fe i 4.45 -1.73 37.3 39.0 43.7 ...... 17.1 ... 31.1 27.5 31.5 ... 36.2 ... 31.6 29.1 31.2 22.2 5405.79 Fe i 0.99 -1.85 ...... 283.9 ...... 284.6 ...... 5415.19 Fe i 4.39 0.51 177.9 181.1 190.1 182.6 159.5 166.7 145.1 161.7 146.3 173.3 110.4 178.1 163.2 166.8 141.8 150.5 151.3 5417.04 Fe i 4.42 -1.42 65.8 47.7 43.5 36.1 39.1 43.0 19.2 77.1 23.1 33.3 ... 46.4 35.3 32.4 ... 40.8 23.0 5434.53 Fe i 1.01 -2.12 ...... 271.4 ...... 259.8 ...... 5436.30 Fe i 4.39 -1.35 ...... 5464.29 Fe i 4.14 -1.62 92.2 ... 90.0 56.4 59.3 56.0 ...... 74.2 ...... 108.1 71.5 58.9 62.9 19.7 67.0 5470.09 Fe i 4.45 -1.60 34.5 46.1 ... 28.5 33.1 23.5 ... 29.9 31.6 43.6 ... 35.5 33.3 29.8 40.3 ... 39.2 5501.48 Fe i 0.96 -3.05 277.6 268.8 274.5 250.1 264.8 245.2 207.2 250.1 234.0 279.6 185.4 278.9 277.9 276.9 241.0 237.5 285.3 5506.79 Fe i 0.99 -2.79 ...... 290.9 212.0 ...... 204.9 ...... 294.6 271.7 ... 5539.29 Fe i 3.64 -2.59 78.5 95.8 83.4 71.9 61.5 55.6 26.6 87.2 52.8 86.6 21.8 84.0 ... 57.8 48.2 48.9 61.5 5586.77 Fe i 3.37 -0.10 252.5 248.6 278.2 246.4 225.6 213.5 171.9 224.0 199.3 243.8 149.2 274.9 225.7 221.1 190.1 212.5 211.9 6120.26 Fe i 0.91 -5.94 101.4 111.6 106.3 102.5 97.4 72.3 28.4 93.0 83.3 96.8 36.1 102.3 94.0 107.8 106.2 76.8 82.3 6136.62 Fe i 2.45 -1.50 ...... 254.6 ... 237.6 274.3 282.0 ... 159.6 ...... 222.3 ... 6137.00 Fe i 2.20 -2.95 ...... 172.1 ...... 158.3 163.6 195.8 85.9 ...... 139.1 ... 6151.62 Fe i 2.18 -3.37 148.8 136.1 144.6 122.5 127.2 128.9 90.8 124.8 122.8 137.6 76.5 149.9 124.0 140.5 128.3 119.4 134.3 6157.75 Fe i 4.07 -1.26 129.7 142.4 139.5 102.0 100.3 125.4 59.8 125.9 107.2 143.3 43.0 146.4 127.3 109.5 119.9 107.2 130.0 6159.38 Fe i 4.61 -1.97 27.0 41.7 37.7 20.1 ... 29.4 ... 30.5 25.2 52.9 ... 45.3 ...... 22.7 ...... 6165.36 Fe i 4.14 -1.47 82.7 87.5 86.5 59.3 65.0 67.8 39.1 75.4 60.6 84.0 28.0 76.3 65.3 68.5 72.3 80.4 72.8 6173.34 Fe i 2.22 -2.85 183.2 186.7 180.4 163.0 161.3 156.7 100.5 170.7 148.2 171.0 99.7 184.4 179.6 176.4 169.7 143.7 190.2 6180.20 Fe i 2.73 -2.78 ... 148.7 168.4 147.2 165.3 100.1 98.9 146.3 125.3 134.8 84.8 157.8 140.4 138.1 131.9 125.8 76.9 6187.99 Fe i 3.94 -1.58 87.3 108.2 97.1 96.2 67.7 81.1 42.9 85.6 75.3 94.7 36.8 78.5 78.1 88.2 80.3 78.4 77.8 6200.31 Fe i 2.61 -2.44 171.8 164.5 170.2 165.1 151.3 147.9 114.6 166.8 145.1 149.0 104.1 165.9 163.2 161.1 157.6 139.1 146.9 6213.43 Fe i 2.22 -2.66 196.8 195.7 192.3 193.3 161.9 200.9 140.1 184.5 185.9 204.3 134.1 197.2 188.4 201.4 183.5 151.7 206.2 6219.29 Fe i 2.20 -2.44 215.8 211.3 206.0 214.4 173.8 206.8 154.0 197.0 189.1 201.4 132.3 200.5 189.8 202.1 192.9 179.9 213.1 6226.74 Fe i 3.88 -2.20 67.3 78.6 72.2 63.4 ... 61.1 22.7 63.5 46.8 ...... 80.6 52.6 60.6 54.8 57.7 69.8 6252.57 Fe i 2.40 -1.76 217.6 211.9 205.2 197.7 197.9 202.1 168.2 189.9 202.8 213.3 135.4 221.5 196.8 218.8 205.8 173.9 217.1 6265.13 Fe i 2.18 -2.55 195.3 207.8 209.4 188.2 163.6 180.2 107.0 180.3 187.9 208.3 120.1 206.1 191.0 195.1 193.5 160.3 175.7 6271.28 Fe i 3.32 -2.96 66.5 82.4 59.1 64.5 52.9 55.9 29.5 59.5 57.6 70.0 37.0 75.4 44.6 57.2 65.8 47.7 51.9 6297.80 Fe i 2.22 -2.74 ...... 6301.50 Fe i 3.65 -0.72 ...... 6307.85 Fe i 3.64 -3.27 ... 25.0 ...... 19.5 ... 21.4 ... 27.9 ...... 6322.69 Fe i 2.59 -2.43 176.2 160.4 163.1 176.9 148.3 153.4 124.0 153.4 155.5 167.8 99.6 169.0 144.1 158.5 165.5 142.7 166.6 6330.85 Fe i 4.73 -1.22 35.3 ... 50.8 42.8 ... 49.4 ... 41.6 36.5 54.1 ... 50.0 37.7 28.8 ... 27.6 40.7 6335.33 Fe i 2.20 -2.23 235.2 224.2 214.1 221.1 207.0 213.7 161.9 201.8 201.1 225.6 151.8 230.9 213.8 237.8 206.4 180.8 234.2 6336.82 Fe i 3.69 -1.05 157.5 150.6 146.2 153.3 144.4 134.0 122.5 162.7 140.0 156.0 80.9 161.6 152.4 155.2 146.1 92.0 148.4 6344.15 Fe i 2.43 -2.92 206.9 188.4 191.4 183.1 161.7 208.7 109.9 165.4 163.5 209.6 82.6 217.8 189.7 184.2 175.9 153.5 176.2 6355.04 Fe i 2.84 -2.29 189.5 177.7 174.5 181.2 136.5 149.6 96.0 159.7 152.6 186.8 100.9 183.4 144.1 166.6 154.4 144.1 162.0 6380.75 Fe i 4.19 -1.50 106.5 99.1 89.6 81.8 94.7 77.4 38.4 99.5 76.0 98.8 34.6 93.0 78.3 83.4 93.6 59.7 87.2 6392.54 Fe i 2.28 -3.95 93.9 98.2 104.3 108.8 100.9 87.8 30.4 99.3 85.8 101.4 36.9 113.5 87.1 106.7 96.7 67.5 86.1 6393.61 Fe i 2.43 -1.63 263.2 261.0 261.7 246.0 246.5 249.8 169.9 240.0 240.2 250.0 161.7 273.4 256.6 252.2 253.5 198.7 243.8 6408.03 Fe i 3.69 -1.00 147.4 168.0 168.1 153.0 161.0 150.2 102.9 144.7 123.7 175.4 77.8 173.2 153.4 158.4 ... 154.9 147.6 6419.96 Fe i 4.73 -0.24 125.2 148.8 137.2 124.9 99.9 116.3 59.9 131.9 112.6 133.6 51.8 127.7 98.9 126.0 ... 99.8 110.0 6421.36 Fe i 2.28 -2.01 251.2 235.6 232.1 235.2 221.2 229.9 178.8 205.2 225.5 236.9 152.3 248.4 222.6 236.1 ... 204.5 240.8 6430.86 Fe i 2.18 -1.95 278.9 292.2 274.9 270.0 255.0 274.3 195.8 250.8 263.4 294.1 165.1 269.1 240.9 279.6 ... 202.8 263.7 6481.87 Fe i 2.27 -2.98 184.2 195.7 161.7 172.9 160.0 178.4 142.7 163.6 172.2 197.0 97.8 179.3 175.8 163.6 197.4 151.7 181.3 6498.94 Fe i 0.96 -4.69 215.0 211.0 213.0 200.6 186.2 192.6 137.8 180.5 203.4 239.2 97.4 219.5 202.4 190.9 228.8 152.8 211.5 6533.93 Fe i 4.55 -1.46 55.0 45.7 61.3 52.6 52.5 55.8 ... 50.3 42.8 50.6 ... 57.1 45.8 41.9 ... 35.9 50.1 6556.81 Fe i 4.79 -1.72 24.7 26.9 28.1 ... 26.3 37.4 ... 19.1 31.8 23.4 ...... 17.2 15.5 58.1 ... 20.5 6569.22 Fe i 4.73 -0.42 123.7 136.9 131.4 109.9 96.8 133.7 70.0 106.8 116.7 135.7 51.0 139.2 100.6 109.3 144.0 97.8 119.6 6574.23 Fe i 0.99 -5.02 188.7 190.3 208.8 185.8 165.0 ... 117.7 176.4 ... 189.1 83.8 ... 183.9 197.3 ... 143.8 193.6 6593.88 Fe i 2.43 -2.39 190.6 210.9 225.2 208.0 174.9 160.8 124.7 193.5 138.1 207.2 90.9 241.2 213.2 200.8 ... 172.9 196.0 6597.56 Fe i 4.79 -1.07 51.5 73.6 85.0 62.9 46.7 62.5 19.9 66.6 75.0 56.3 19.4 74.2 64.4 55.1 ... 66.1 55.2 6608.03 Fe i 2.28 -3.94 117.4 112.9 115.0 111.9 87.4 105.7 54.9 78.3 131.0 120.0 59.3 119.9 97.5 100.9 155.1 73.8 110.7 6609.12 Fe i 2.56 -2.66 157.1 176.1 163.4 166.0 128.7 166.6 120.0 152.2 196.8 176.8 76.7 166.4 158.0 162.5 223.2 122.1 190.2 6627.54 Fe i 4.54 -1.68 53.4 32.8 41.7 36.0 30.4 43.2 18.5 29.8 56.9 42.9 ... 25.8 33.0 20.6 105.3 24.3 29.6 6633.76 Fe i 4.56 -0.82 128.0 130.5 131.2 111.8 107.8 127.5 49.9 136.3 141.5 134.0 44.8 133.6 98.7 103.5 149.8 98.2 114.2 6646.93 Fe i 2.60 -3.99 78.9 74.4 86.7 54.1 65.0 60.6 17.9 66.5 53.2 76.2 ... 79.9 69.6 71.1 86.1 52.4 51.2 6653.85 Fe i 4.15 -2.52 ... 44.4 46.5 49.4 29.9 38.5 ... 31.8 42.7 40.8 16.1 50.9 21.0 37.8 ... 29.9 34.4 5414.08 Fe ii 3.22 -3.61 ...... 20.3 22.5 ... 30.9 ...... 5425.25 Fe ii 3.20 -3.36 35.6 ... 45.5 52.5 82.1 51.8 51.2 58.9 ... 30.6 56.7 30.4 40.7 29.2 20.7 70.4 22.6 6149.25 Fe ii 3.89 -2.72 35.2 28.9 35.5 35.8 ... 34.0 25.9 27.7 20.0 ...... 43.4 28.2 34.0 38.6 24.7 ... 6432.68 Fe ii 2.89 -3.71 49.8 26.0 46.2 65.1 55.5 74.1 56.4 50.3 62.1 41.8 33.7 ... 47.1 45.9 ... 40.4 47.1 6456.39 Fe ii 3.90 -2.08 40.4 31.9 64.7 ... 60.6 79.2 29.5 86.6 ... 71.4 49.2 86.6 71.8 67.2 ... 66.2 37.7 6320.43 La ii 0.17 -1.56 94.8 128.7 111.5 63.0 70.7 53.0 20.2 91.6 73.2 109.7 25.4 130.3 118.1 82.0 79.3 63.4 68.7 6390.46 La ii 0.32 -1.40 108.4 109.9 98.9 68.4 64.9 69.4 30.2 72.0 80.7 95.8 24.6 108.7 93.6 ... 74.5 79.4 70.4 5528.41 Mg i 4.35 -0.36 227.2 223.4 220.8 205.5 219.7 217.4 169.5 216.3 221.3 234.8 164.9 225.7 240.7 208.9 208.1 207.0 220.0 Continued on next page 6.A: Large tables 135

λ(Å) elem χ gf 253 257 258 260 261 262 266 267 269 278 279 295 300 304 311 315 323 6318.72 Mg i 5.11 -1.97 44.0 64.6 47.2 32.4 25.6 50.3 27.4 43.6 28.3 ...... 49.7 42.8 ... 41.7 ... 30.7 6319.24 Mg i 5.11 -2.21 26.0 25.8 ...... 20.5 ...... 26.4 38.8 ... 41.9 ...... 25.7 ...... 6319.49 Mg i 5.11 -2.43 ...... 28.2 ...... 31.0 18.5 ...... 15.2 5420.36 Mn i 2.14 -1.46 203.5 220.4 238.5 190.0 185.7 193.4 71.1 201.4 204.6 222.2 57.3 229.3 196.5 217.0 172.2 199.3 192.7 5432.55 Mn i 0.00 -3.80 292.1 ... 283.8 268.2 ... 271.9 156.3 236.9 254.0 ... 111.4 ... 281.0 270.9 284.1 220.5 291.9 5516.77 Mn i 2.18 -1.85 167.7 174.7 179.1 146.2 156.5 142.5 49.9 139.3 164.4 210.9 ... 211.2 158.8 161.1 154.3 108.0 144.2 6154.23 Na i 2.10 -1.56 41.9 37.6 45.3 ... 25.0 ...... 31.7 24.5 52.8 16.2 48.8 33.4 17.0 24.3 ...... 6160.75 Na i 2.10 -1.26 68.6 62.1 63.6 34.0 32.5 25.2 ... 58.8 51.6 96.0 ... 70.2 52.3 38.9 47.2 26.3 44.3 5416.38 Nd ii 0.86 -0.98 19.8 36.1 54.3 40.6 48.5 ...... 37.3 18.9 32.6 ... 69.6 35.0 33.4 ...... 5431.54 Nd ii 1.12 -0.47 70.6 75.9 121.6 50.1 40.4 59.3 ... 56.7 61.9 74.7 27.7 91.0 60.1 35.5 ... 38.3 58.7 5485.71 Nd ii 1.26 -0.12 55.3 60.8 59.8 31.8 46.5 37.0 ... 23.3 35.6 66.5 ... 50.6 59.8 43.1 34.8 ... 33.9 5578.73 Ni i 1.68 -2.67 147.4 165.4 161.2 125.8 121.5 121.0 96.1 148.4 141.2 174.5 81.1 161.9 143.0 152.3 131.7 128.5 140.8 5587.87 Ni i 1.93 -2.37 171.3 204.0 176.8 159.7 138.7 142.1 79.7 164.8 145.6 188.1 73.0 191.4 161.9 177.5 131.3 148.8 146.6 5589.37 Ni i 3.90 -1.15 37.4 25.3 25.2 38.7 28.7 22.6 ...... 40.2 39.9 ... 36.7 ... 24.5 31.2 28.5 25.9 5593.75 Ni i 3.90 -0.79 39.7 53.6 47.5 38.2 45.3 42.2 ... 28.6 30.1 54.8 39.7 43.5 35.9 35.2 19.7 49.5 24.3 6128.97 Ni i 1.68 -3.39 105.2 96.0 99.1 93.3 86.7 94.7 48.8 92.7 86.6 108.3 34.0 99.2 99.3 95.3 77.0 42.2 93.8 6130.14 Ni i 4.27 -0.98 16.7 27.5 ...... 25.4 ...... 26.6 ...... 22.4 ...... 25.1 17.2 6177.25 Ni i 1.83 -3.60 82.2 80.4 76.0 62.4 69.3 61.0 39.8 69.0 56.1 66.7 31.7 78.2 72.8 67.1 67.9 66.2 62.4 6186.72 Ni i 4.11 -0.90 37.1 52.5 52.6 37.1 ... 20.1 ... 27.4 25.2 44.9 23.9 37.3 39.8 ... 33.6 32.3 30.7 6204.61 Ni i 4.09 -1.15 40.7 ...... 15.4 ... 17.3 16.5 16.9 ... 26.0 15.8 20.8 ...... 21.7 31.3 15.0 6223.99 Ni i 4.10 -0.97 37.8 45.9 49.3 23.0 39.9 ...... 25.7 46.1 ... 31.2 28.4 30.1 42.3 42.0 ... 6230.10 Ni i 4.11 -1.20 35.0 46.7 43.8 ...... 36.6 29.1 70.0 ... 44.2 31.1 31.5 ... 21.9 47.8 6322.17 Ni i 4.15 -1.21 ... 32.3 ...... 23.1 ...... 20.4 24.4 ...... 6327.60 Ni i 1.68 -3.09 125.6 135.2 130.8 115.7 115.6 117.1 71.5 113.1 110.2 130.9 43.0 133.5 116.1 121.3 118.2 92.1 114.8 6378.26 Ni i 4.15 -0.82 40.4 46.7 35.4 30.4 18.4 32.6 21.6 23.9 29.3 46.1 ... 41.6 26.0 20.4 40.0 41.3 31.9 6384.67 Ni i 4.15 -1.00 38.5 49.3 60.2 ...... 20.0 18.8 46.5 47.6 50.0 30.6 54.0 44.0 28.1 38.4 ... 34.2 6482.80 Ni i 1.94 -2.85 113.7 115.3 110.1 98.6 93.3 118.0 66.3 110.8 136.0 134.6 57.3 110.1 102.2 96.2 136.9 93.3 105.5 6586.32 Ni i 1.95 -2.79 122.2 133.4 148.7 158.1 136.6 46.1 59.8 164.4 41.4 116.8 41.1 162.2 164.2 157.8 ... 150.0 135.1 6598.61 Ni i 4.24 -0.93 27.1 ...... 26.3 6635.14 Ni i 4.42 -0.75 25.1 ... 23.9 ...... 36.9 ...... 45.6 23.9 21.6 ...... 68.7 ... 18.4 6300.31 O i 0.00 -9.75 ...... 6363.79 O i 0.02 -10.25 45.8 54.1 40.5 42.5 27.0 33.2 23.4 41.1 48.6 46.6 ... 31.9 37.8 46.1 38.9 41.9 55.7 5526.82 Sc ii 1.77 0.03 119.7 110.3 110.1 91.8 111.6 119.5 96.0 107.6 107.9 115.4 115.7 96.4 122.6 109.3 112.3 131.2 105.6 6245.62 Sc ii 1.51 -0.97 79.9 67.0 77.7 71.9 65.9 66.6 68.4 61.3 69.2 ... 52.8 78.5 54.1 61.2 82.8 54.5 60.6 6309.90 Sc ii 1.50 -1.52 58.0 54.7 ... 34.7 37.1 ... 18.7 39.9 ...... 30.6 ... 54.7 43.5 ...... 6604.60 Sc ii 1.36 -1.31 101.0 88.9 94.8 82.7 84.2 98.3 58.4 83.7 113.2 93.9 54.5 102.4 87.0 82.5 126.4 71.7 82.6 6125.03 Si i 5.62 -1.57 ...... 31.7 ...... 32.3 38.8 ...... 56.2 ...... 17.4 ... 6142.48 Si i 5.62 -1.51 ...... 26.0 ...... 16.5 ...... 16.9 ... 6145.02 Si i 5.61 -1.37 15.4 22.7 ...... 22.1 ... 21.2 ...... 19.3 ... 21.7 6155.14 Si i 5.62 -0.80 47.6 55.8 39.1 42.8 27.6 45.5 ... 41.0 38.9 51.2 34.6 49.0 40.0 39.0 43.5 47.9 45.0 6237.33 Si i 5.61 -1.02 26.6 23.7 32.1 16.0 ... 30.2 ... 33.6 23.4 43.1 22.3 31.0 26.8 ... 21.6 31.4 ... 6243.82 Si i 5.61 -1.27 ... 53.1 44.0 ...... 22.5 28.0 33.8 46.9 ...... 29.3 17.0 23.1 ... 5490.16 Ti i 1.46 -0.93 132.3 158.9 148.0 98.6 116.9 97.2 50.8 87.2 113.2 143.2 32.3 152.3 123.0 132.5 106.5 92.5 117.3 5503.90 Ti i 2.58 -0.19 76.8 94.2 120.5 66.3 65.9 56.5 23.4 53.6 52.8 107.6 ... 103.1 86.6 67.2 53.2 40.6 71.7 6126.22 Ti i 1.07 -1.42 167.5 167.6 172.5 130.1 130.7 127.2 56.8 138.8 148.4 182.3 32.4 172.4 162.0 159.6 138.8 119.0 160.5 6220.50 Ti i 2.68 -0.14 ...... 24.1 ... 56.6 94.0 ... 81.3 80.0 ...... 6258.10 Ti i 1.44 -0.35 223.9 231.1 232.3 195.2 160.5 181.7 83.5 165.8 209.4 258.2 72.4 243.1 216.5 223.1 181.3 138.1 204.8 6303.77 Ti i 1.44 -1.57 121.7 117.2 115.7 91.1 79.0 87.1 41.0 77.9 87.3 121.2 ... 115.6 96.2 100.3 87.9 68.3 97.9 6312.24 Ti i 1.46 -1.55 ... 114.7 107.9 ...... 79.2 20.6 78.5 91.3 113.1 ... 100.9 ...... 88.7 45.6 86.3 6336.10 Ti i 1.44 -1.74 103.0 101.2 115.0 90.8 66.1 63.6 26.0 77.7 90.6 120.9 25.3 97.7 90.7 89.7 73.6 68.5 103.8 6508.12 Ti i 1.43 -2.05 69.8 90.0 72.0 56.9 45.3 53.3 ... 38.1 71.5 94.2 ... 90.2 58.0 54.3 63.9 31.0 63.5 6556.08 Ti i 1.46 -1.07 ...... 138.0 ...... 139.8 ...... 6599.13 Ti i 0.90 -2.09 153.9 186.8 181.3 147.5 111.3 136.9 46.9 109.4 167.9 173.1 24.5 194.5 159.1 174.3 ... 78.9 164.2 6666.53 Ti i 1.46 -1.62 41.8 50.1 59.5 ...... 39.7 39.2 ...... 29.8 33.9 46.2 ...... 5418.77 Ti ii 1.58 -2.11 111.2 97.0 96.3 99.6 115.3 93.2 97.6 116.2 83.3 109.4 122.1 92.7 109.8 102.2 83.0 110.0 96.5 6219.94 Ti ii 2.06 -2.82 ...... 28.0 31.3 28.1 ... 21.6 48.8 ... 33.9 27.4 ... 40.7 ...... 21.8 ... 6559.58 Ti ii 2.05 -2.02 67.8 ...... 72.9 ...... 57.7 61.8 ...... 6606.95 Ti ii 2.06 -2.79 43.8 40.7 46.2 24.6 34.4 47.6 39.3 34.1 59.6 49.2 31.7 49.9 33.7 ... 88.0 27.1 31.1 6680.13 Ti ii 3.09 -1.86 23.1 29.5 33.0 17.5 27.9 ... 26.5 41.5 29.1 21.4 26.1 36.4 25.1 17.9 36.6 30.7 19.9 6119.53 V i 1.06 -0.32 ...... 97.4 ...... 74.5 ... 79.1 6128.33 V i 1.05 -2.30 ...... 18.8 ...... 30.5 ...... 6135.37 V i 1.05 -0.75 139.1 151.3 ... 88.9 103.1 102.4 23.7 97.8 113.4 145.4 18.9 145.0 114.8 116.7 109.2 53.5 116.8 6150.15 V i 0.30 -1.79 166.8 178.6 152.7 148.5 140.0 128.8 39.7 123.6 147.8 181.6 23.9 181.2 170.7 172.2 148.8 95.5 150.0 6199.19 V i 0.29 -1.29 222.5 220.2 219.8 188.6 170.3 182.9 80.9 185.5 188.0 224.1 44.0 216.4 206.0 200.3 177.1 122.5 176.8 6216.36 V i 0.28 -0.81 198.5 213.6 215.0 183.8 172.5 178.3 76.9 161.4 197.1 220.1 46.5 199.8 186.6 200.0 187.5 164.8 221.4 6224.51 V i 0.29 -2.01 146.9 157.1 166.9 131.9 119.3 125.7 31.5 117.9 141.9 170.8 ... 156.3 142.7 150.1 119.8 85.1 153.5 6233.20 V i 0.28 -2.07 106.0 107.5 117.6 111.4 80.0 87.4 ... 73.4 107.9 130.9 ... 130.1 109.1 119.3 95.1 54.2 113.5 6243.11 V i 0.30 -0.98 ...... 261.0 ... 110.6 244.7 ...... 47.0 ...... 177.2 ... 6251.82 V i 0.29 -1.30 171.7 166.0 160.4 128.5 137.9 143.1 73.3 139.6 163.9 182.3 29.4 172.7 168.1 155.8 141.1 94.5 174.5 6274.66 V i 0.27 -1.67 ...... 138.9 ...... 101.9 47.2 ... 135.5 151.8 ... 164.7 ...... 120.0 77.7 128.9 6357.29 V i 1.85 -0.91 27.9 60.1 30.6 21.5 ...... 30.3 27.8 55.9 ... 50.7 35.7 18.3 40.1 24.1 38.0 6452.32 V i 1.19 -1.21 97.3 128.1 140.8 99.8 57.9 38.1 15.2 100.2 51.9 119.0 ... 144.9 93.7 96.7 ... 48.1 97.5 6504.19 V i 1.18 -1.23 77.0 94.2 69.1 46.0 52.0 41.9 19.2 ... 62.7 94.4 ... 75.4 68.1 53.1 86.1 27.5 54.1 6531.41 V i 1.22 -0.84 84.0 ...... 56.8 ...... 62.7 79.8 ...... 5402.78 Y ii 1.84 -0.51 43.1 62.4 64.5 33.3 44.3 ... 27.9 50.5 32.5 63.1 ... 61.5 41.8 27.8 64.1 ... 51.8 6362.35 Zn i 5.80 0.14 ... 30.6 ...... 21.2 ...... 21.5 ... 19.8 6127.48 Zr i 0.15 -1.06 103.2 112.5 110.0 82.8 58.2 71.0 15.3 72.1 99.1 125.1 ... 121.8 106.2 95.2 87.5 53.4 108.8 6140.46 Zr i 0.52 -1.41 26.0 32.5 39.3 ... 19.7 22.9 ...... 19.6 47.1 ... 45.9 27.2 28.0 ... 23.2 ... 6143.18 Zr i 0.07 -1.10 110.6 124.4 114.9 82.2 77.8 55.6 18.4 62.9 95.0 150.3 18.9 110.1 114.0 101.3 89.8 45.5 115.5 6192.95 Zr i 0.54 -2.07 15.5 35.9 21.1 ...... 24.0 29.0 ... 18.2 ... 16.9 ...... 21.3 136 chapter 6: HR spectroscopic study of Fornax Field Stars (n) σ 0.13 (2) 0.09 (3) 0.31 (2) 0.13 (2) 0.19 (1) 0.12 (2) 0.15 (2) 0.21 (1) 0.12 (2) 0.12 (3) 0.21 (1) 0.10 (3) 0.15 (2) 0.09 (3) 0.19 (3) 0.18 (2) 0.10 (3) 0.31 (2) 0.13 (2) 0.11 (2) 0.14 (2) 0.13 (3) 0.13 (2) 0.11 (3) 0.12 (3) 0.24 (2) 0.13 (3) 0.16 (2) 0.18 (2) 0.24 (2) 0.18 (1) 0.15 (1) 0.13 (3) 0.14 (3) 0.17 (1) 0.17 (3) 0.13 (3) 0.18 (2) 0.15 (3) 0.24 (3) 0.16 (2) 0.13 (3) /Fe] ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± i [Mg 0.03 0.02 0.07 0.00 0.35 0.04 0.02 0.06 0.00 0.01 0.02 0.04 0.43 0.01 0.06 -0.15 -0.27 -0.06 -0.32 -0.12 -0.04 -0.01 -0.03 -0.04 -0.13 -0.02 -0.08 -0.11 -0.14 -0.13 -0.06 -0.06 -0.03 -0.01 -0.20 -0.16 -0.07 -0.11 -0.03 -0.12 -0.08 -0.07 n σ Continued on next page ... (0) 0.16 (2) 0.16 (1) 0.34 (1) 0.19 (1) 0.31 (2) 0.13 (2) 0.13 (2) 0.15 (1) 0.12 (2) 0.16 (2) 0.25 (1) 0.12 (2) 0.13 (2) 0.21 (1) 0.15 (2) 0.11 (2) 0.12 (2) 0.25 (2) 0.14 (2) 0.15 (2) 0.12 (2) 0.13 (2) 0.15 (2) 0.11 (2) 0.23 (2) 0.23 (2) 0.18 (2) 0.16 (2) 0.13 (2) 0.16 (2) 0.13 (2) 0.16 (2) 0.15 (1) 0.14 (2) 0.16 (2) 0.13 (2) 0.13 (2) 0.15 (2) 0.24 (2) 0.18 (2) 0.11 (2) /Fe] ± ii ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ... [La 0.43 0.17 0.25 0.21 0.44 0.36 0.24 0.75 0.32 0.30 0.36 0.25 0.15 0.45 0.40 0.11 0.38 0.58 0.01 0.29 0.40 0.45 0.28 0.49 1.09 0.54 0.31 0.37 0.42 0.34 0.54 0.58 0.46 0.62 0.36 0.53 0.33 0.47 0.76 0.43 -0.01 (n) σ ... (0) ... (0) 0.23 (1) 0.16 (1) 0.34 (1) 0.18 (1) 0.19 (1) 0.15 (1) 0.44 (1) 0.19 (1) 0.19 (1) 0.17 (1) 0.23 (1) 0.25 (1) 0.17 (1) 0.21 (1) 0.21 (1) 0.16 (1) 0.17 (1) 0.25 (1) 0.20 (1) 0.21 (1) 0.17 (1) 0.18 (1) 0.21 (1) 0.16 (1) 0.32 (1) 0.33 (1) 0.25 (1) 0.22 (1) 0.18 (1) 0.22 (1) 0.19 (1) 0.23 (1) 0.15 (1) 0.20 (1) 0.22 (1) 0.18 (1) 0.18 (1) 0.21 (1) 0.34 (1) 0.22 (1) ± ± /Fe] ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ii ...... ) and the number of lines used (n). The quoted error σ 0.62 0.19 0.67 0.35 0.34 0.29 0.45 0.26 0.44 0.52 0.53 0.53 0.48 0.64 0.45 0.12 0.40 0.51 0.24 0.47 0.55 0.62 0.39 0.41 1.38 0.35 0.43 0.50 0.46 0.39 0.65 0.64 0.47 0.51 0.42 0.54 0.44 0.40 0.58 0.37 [Eu n σ ... (0) 0.23 (1) 0.16 (1) 0.34 (1) 0.18 (1) 0.19 (1) 0.15 (1) 0.44 (1) 0.19 (1) 0.19 (1) 0.17 (1) 0.23 (1) 0.25 (1) 0.17 (1) 0.19 (1) 0.21 (1) 0.21 (1) 0.16 (1) 0.17 (1) 0.25 (1) 0.20 (1) 0.21 (1) 0.17 (1) 0.18 (1) 0.21 (1) 0.16 (1) 0.32 (1) 0.33 (1) 0.25 (1) 0.22 (1) 0.18 (1) 0.22 (1) 0.19 (1) 0.23 (1) 0.15 (1) 0.20 (1) 0.22 (1) 0.18 (1) 0.18 (1) 0.21 (1) 0.34 (1) 0.22 (1) /Fe] ± i ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ... [Cr -0.31 -0.36 -0.51 -0.51 -0.42 -0.62 -0.30 -0.52 -0.57 -0.34 -0.45 -0.53 -0.54 -0.50 -0.75 -0.42 -0.55 -0.40 -0.57 -0.50 -0.41 -0.47 -0.39 -0.47 -0.40 -0.80 -0.23 -0.52 -0.38 -0.33 -0.40 -0.44 -0.49 -0.60 -0.23 -0.59 -0.30 -0.28 -0.54 -0.46 -0.15 (n) σ 0.11 (8) 0.06 (9) 0.06 (8) 0.12 (8) 0.12 (9) 0.06 (9) 0.06 (8) 0.17 (7) 0.07 (9) 0.06 (9) 0.07 (7) 0.08 (9) 0.10 (6) 0.09 (9) 0.08 (8) 0.07 (9) 0.09 (9) 0.06 (8) 0.06 (8) 0.09 (8) 0.07 (8) 0.08 (7) 0.09 (9) 0.06 (9) 0.08 (7) 0.10 (8) 0.11 (9) 0.14 (8) 0.09 (8) 0.08 (7) 0.07 (8) 0.08 (9) 0.07 (7) 0.08 (9) 0.08 (9) 0.07 (8) 0.08 (7) 0.07 (7) 0.07 (7) 0.08 (8) 0.12 (8) 0.10 (9) /Fe] ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± i [Ca 0.33 -0.25 -0.33 -0.42 -0.09 -0.21 -0.46 -0.26 -0.34 -0.30 -0.15 -0.23 -0.18 -0.18 -0.48 -0.20 -0.08 -0.28 -0.30 -0.21 -0.45 -0.30 -0.18 -0.30 -0.33 -0.36 -0.30 -0.29 -0.30 -0.18 -0.25 -0.17 -0.28 -0.38 -0.31 -0.08 -0.23 -0.12 -0.26 -0.36 -0.25 -0.06 n σ 0.16 (2) 0.39 (2) 0.13 (2) 0.16 (2) 0.13 (2) 0.44 (1) 0.13 (2) 0.21 (2) 0.11 (2) 0.12 (2) 0.16 (2) 0.18 (2) 0.20 (2) 0.25 (2) 0.15 (2) 0.21 (1) 0.11 (2) 0.12 (2) 0.18 (2) 0.21 (2) 0.15 (2) 0.21 (2) 0.13 (2) 0.15 (2) 0.20 (2) 0.23 (2) 0.33 (1) 0.18 (2) 0.16 (2) 0.13 (2) 0.20 (2) 0.13 (2) 0.23 (1) 0.12 (2) 0.20 (1) 0.16 (2) 0.13 (2) 0.13 (2) 0.15 (2) 0.34 (1) 0.16 (2) 0.21 (2) /Fe] ii ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± [Ba 0.80 0.57 0.19 0.49 0.29 0.98 0.91 0.66 0.12 0.56 0.66 0.70 0.56 0.20 0.58 0.94 0.07 0.40 0.87 0.23 0.62 0.47 0.69 0.28 0.58 1.27 0.09 0.57 0.36 0.74 0.76 0.51 0.95 0.55 0.95 0.74 0.55 0.18 0.54 1.22 0.96 -0.05 (n) σ 0.09 (4) 0.08 (5) 0.25 (2) 0.17 (5) 0.14 (4) 0.24 (4) 0.42 (2) 0.20 (4) 0.10 (4) 0.10 (4) 0.21 (5) 0.10 (3) 0.18 (4) 0.24 (5) 0.13 (5) 0.23 (4) 0.10 (5) 0.12 (4) 0.07 (4) 0.13 (3) 0.12 (4) 0.17 (4) 0.13 (5) 0.12 (5) 0.07 (5) 0.10 (4) 0.15 (4) 0.17 (5) 0.15 (4) 0.10 (3) 0.21 (3) 0.19 (5) 0.15 (4) 0.08 (5) 0.14 (4) 0.24 (4) 0.20 (4) 0.14 (4) 0.14 (5) 0.06 (1) 0.14 (4) 0.27 (5) /H] ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ii [Fe 0.22 -0.70 -0.78 -0.52 -1.06 -0.43 -0.38 -0.16 -0.44 -0.59 -1.99 -0.60 -0.57 -0.95 -0.53 -0.67 -0.33 -0.65 -0.85 -0.78 -0.61 -0.61 -0.69 -0.58 -0.42 -0.64 -0.87 -0.56 -0.83 -0.41 -0.60 -0.52 -0.62 -0.40 -0.52 -0.52 -0.54 -0.57 -0.19 -0.55 -0.68 -0.78 n σ /H] i 0.07 (43) 0.05 (48) 0.06 (37) 0.07 (42) 0.06 (43) 0.06 (40) 0.08 (39) 0.06 (40) 0.05 (39) 0.05 (31) 0.06 (46) 0.08 (45) 0.08 (42) 0.05 (43) 0.06 (48) 0.06 (43) 0.06 (42) 0.06 (40) 0.05 (41) 0.07 (44) 0.06 (43) 0.07 (39) 0.06 (43) 0.07 (41) 0.06 (46) 0.06 (39) 0.07 (47) 0.11 (37) 0.07 (45) 0.06 (42) 0.07 (45) 0.07 (40) 0.07 (45) 0.07 (43) 0.06 (37) 0.07 (41) 0.07 (42) 0.08 (41) 0.06 (40) 0.07 (41) 0.07 (38) 0.07 (39) ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± [Fe Abundance ratio of the Fornax Field stars, where we list the abundance, its associated error ( -0.88 -1.09 -1.02 -1.43 -0.85 -0.79 -0.52 -0.62 -0.82 -2.58 -0.96 -0.95 -0.75 -0.92 -0.93 -0.96 -0.75 -1.44 -0.97 -0.73 -0.85 -0.95 -1.01 -0.86 -0.82 -0.92 -1.37 -0.63 -0.91 -0.83 -0.86 -0.75 -1.13 -0.87 -0.87 -0.74 -0.89 -0.88 -0.90 -0.78 -0.91 -0.71 Star BL038 BL045 BL052 BL065 BL076 BL077 BL079 BL081 BL084 BL085 BL091 BL092 BL096 BL097 BL100 BL104 BL113 BL115 BL123 BL125 BL132 BL135 BL138 BL140 BL141 BL146 BL147 BL148 BL149 BL150 BL151 BL155 BL156 BL158 BL160 BL163 BL166 BL168 BL171 BL173 BL180 BL185 Table 6.A3: is the error on [element/H], not [X/Fe]. Part 1 6.A: Large tables 137 (n) σ 0.13 (2) 0.13 (2) 0.18 (1) 0.16 (3) 0.34 (1) 0.13 (2) 0.13 (3) 0.13 (3) 0.16 (2) 0.14 (3) 0.15 (3) 0.15 (2) 0.18 (2) 0.11 (2) 0.17 (2) 0.13 (3) 0.16 (2) 0.15 (3) 0.12 (3) 0.14 (3) 0.16 (2) 0.13 (2) 0.15 (2) 0.16 (3) 0.16 (3) 0.14 (3) 0.16 (2) 0.20 (3) 0.21 (1) 0.15 (3) 0.16 (3) 0.18 (3) 0.17 (3) 0.23 (2) 0.23 (2) 0.12 (3) 0.18 (2) 0.21 (1) 0.11 (3) /Fe] ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± i [Mg 0.05 0.06 0.15 0.10 0.04 0.00 0.00 0.18 0.18 -0.32 -0.07 -0.22 -0.07 -0.10 -0.07 -0.08 -0.15 -0.19 -0.03 -0.09 -0.11 -0.05 -0.16 -0.16 -0.13 -0.15 -0.03 -0.20 -0.11 -0.29 -0.27 -0.13 -0.01 -0.06 -0.04 -0.09 -0.17 -0.01 -0.18 n σ 0.15 (2) 0.19 (1) 0.13 (2) 0.17 (2) 0.19 (2) 0.24 (2) 0.13 (2) 0.16 (2) 0.15 (2) 0.16 (2) 0.15 (2) 0.18 (2) 0.18 (2) 0.15 (2) 0.18 (2) 0.11 (2) 0.17 (2) 0.16 (2) 0.16 (2) 0.16 (2) 0.11 (2) 0.18 (2) 0.14 (2) 0.17 (2) 0.16 (2) 0.13 (2) 0.15 (2) 0.19 (2) 0.14 (2) 0.14 (2) 0.20 (2) 0.18 (2) 0.15 (2) 0.18 (2) 0.13 (2) 0.20 (1) 0.24 (2) 0.15 (2) 0.18 (2) /Fe] ii ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± [La 0.01 0.10 0.40 0.37 0.36 0.53 0.41 0.44 0.52 0.45 0.47 0.62 0.48 0.61 0.47 0.08 0.46 0.33 0.14 0.27 0.50 0.43 0.53 0.62 0.56 0.26 0.30 0.25 0.18 0.47 0.40 0.60 0.23 0.67 0.74 0.36 0.36 0.49 0.18 (n) σ 0.18 (1) 0.19 (1) 0.18 (1) 0.24 (1) 0.27 (1) 0.34 (1) 0.19 (1) 0.22 (1) 0.21 (1) 0.22 (1) 0.21 (1) 0.25 (1) 0.26 (1) 0.21 (1) 0.25 (1) 0.16 (1) 0.24 (1) 0.22 (1) 0.22 (1) 0.21 (1) 0.15 (1) 0.26 (1) 0.20 (1) 0.24 (1) 0.23 (1) 0.19 (1) 0.21 (1) 0.27 (1) 0.20 (1) 0.20 (1) 0.28 (1) 0.25 (1) 0.21 (1) 0.25 (1) 0.19 (1) 0.20 (1) 0.34 (1) 0.21 (1) 0.26 (1) /Fe] ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ii 0.39 0.43 0.50 0.54 0.51 0.92 0.31 0.24 0.47 0.40 0.52 0.55 0.29 0.68 0.70 0.31 0.48 0.36 0.26 0.47 0.47 0.26 0.42 0.32 0.49 0.32 0.46 0.33 0.45 0.37 0.64 0.40 0.64 0.44 0.46 0.38 0.85 0.69 0.43 [Eu n σ 0.18 (1) 0.19 (1) 0.18 (1) 0.24 (1) 0.27 (1) 0.34 (1) 0.19 (1) 0.22 (1) 0.21 (1) 0.22 (1) 0.21 (1) 0.25 (1) 0.26 (1) 0.21 (1) 0.25 (1) 0.16 (1) 0.24 (1) 0.22 (1) 0.22 (1) 0.21 (1) 0.15 (1) 0.26 (1) 0.20 (1) 0.24 (1) 0.23 (1) 0.19 (1) 0.21 (1) 0.27 (1) 0.20 (1) 0.20 (1) 0.28 (1) 0.25 (1) 0.21 (1) 0.25 (1) 0.19 (1) 0.20 (1) 0.34 (1) 0.21 (1) 0.26 (1) /Fe] i ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± [Cr -0.67 -0.43 -0.40 -0.57 -0.47 -0.30 -0.42 -0.56 -0.53 -0.21 -0.51 -0.28 -0.38 -0.39 -0.21 -0.37 -0.58 -0.64 -0.47 -0.37 -0.38 -0.17 -0.50 -0.47 -0.35 -0.49 -0.55 -0.38 -0.36 -0.32 -0.54 -0.11 -0.51 -0.45 -0.28 -0.57 -0.52 -0.51 -0.48 (n) σ 0.06 (8) 0.11 (8) 0.08 (8) 0.06 (9) 0.10 (7) 0.09 (9) 0.15 (5) 0.12 (7) 0.10 (9) 0.07 (8) 0.07 (9) 0.07 (9) 0.10 (7) 0.10 (7) 0.10 (9) 0.08 (9) 0.08 (9) 0.11 (8) 0.08 (8) 0.08 (8) 0.09 (9) 0.11 (9) 0.11 (7) 0.11 (8) 0.08 (8) 0.13 (9) 0.08 (9) 0.08 (8) 0.10 (7) 0.10 (9) 0.09 (9) 0.11 (9) 0.13 (8) 0.09 (7) 0.08 (8) 0.13 (7) 0.09 (6) 0.14 (8) 0.09 (10) /Fe] ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± i ± [Ca 0.04 -0.43 -0.20 -0.34 -0.42 -0.19 -0.15 -0.21 -0.07 -0.32 -0.20 -0.32 -0.18 -0.26 -0.27 -0.21 -0.45 -0.25 -0.47 -0.21 -0.45 -0.30 -0.06 -0.32 -0.30 -0.16 -0.30 -0.36 -0.19 -0.16 -0.22 -0.03 -0.07 -0.16 -0.34 -0.25 -0.34 -0.56 -0.18 n σ ... (0) ... (0) ... (0) 0.13 (2) 0.25 (2) 0.13 (2) 0.30 (2) 0.19 (2) 0.47 (2) 0.16 (2) 0.16 (2) 0.15 (2) 0.16 (2) 0.15 (2) 0.18 (2) 0.26 (1) 0.15 (2) 0.18 (2) 0.17 (2) 0.16 (2) 0.21 (2) 0.16 (2) 0.11 (2) 0.14 (2) 0.24 (1) 0.23 (1) 0.13 (2) 0.15 (2) 0.25 (2) 0.14 (2) 0.14 (2) 0.20 (2) 0.15 (2) 0.13 (2) 0.14 (2) 0.34 (1) 0.18 (2) 0.18 (2) 0.13 (2) /Fe] ± ± ± ii ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ...... [Ba 0.06 0.17 0.22 0.79 0.58 0.46 0.76 0.86 0.90 0.98 0.65 0.96 1.02 0.80 0.94 0.79 0.66 0.67 0.64 0.49 0.88 0.91 1.01 0.07 0.40 0.59 0.18 0.65 0.89 0.46 1.08 0.56 0.63 0.66 0.53 -0.05 (n) σ 0.12 (4) 0.09 (4) 0.10 (5) 0.13 (2) 0.12 (3) 0.16 (2) 0.08 (5) 0.06 (3) 0.22 (5) 0.14 (4) 0.10 (5) 0.19 (2) 0.23 (3) 0.20 (5) 0.14 (3) 0.17 (4) 0.14 (4) 0.12 (4) 0.12 (4) 0.23 (3) 0.12 (5) 0.34 (4) 0.12 (4) 0.15 (3) 0.08 (4) 0.11 (3) 0.24 (4) 0.15 (5) 0.21 (4) 0.11 (5) 0.40 (2) 0.16 (3) 0.20 (3) 0.27 (3) 0.09 (4) 0.12 (4) 0.40 (2) 0.18 (4) 0.19 (3) /H] ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ii [Fe -0.60 -0.77 -0.76 -0.79 -0.55 -0.74 -0.63 -0.73 -0.37 -0.95 -0.63 -0.93 -0.60 -0.75 -0.65 -0.49 -0.48 -0.46 -0.67 -0.73 -0.61 -0.51 -0.61 -1.09 -0.53 -0.36 -0.37 -0.44 -1.16 -0.64 -0.47 -0.87 -1.31 -0.25 -0.60 -0.68 -1.00 -0.62 -0.69 n σ /H] i 0.05 (43) 0.06 (48) 0.06 (43) 0.07 (39) 0.07 (43) 0.07 (33) 0.07 (43) 0.06 (43) 0.07 (42) 0.07 (43) 0.06 (42) 0.07 (35) 0.09 (41) 0.06 (43) 0.08 (43) 0.05 (42) 0.06 (39) 0.07 (44) 0.07 (43) 0.07 (43) 0.09 (37) 0.09 (38) 0.06 (39) 0.07 (41) 0.08 (39) 0.06 (39) 0.09 (43) 0.07 (44) 0.06 (43) 0.08 (46) 0.08 (44) 0.07 (38) 0.08 (44) 0.08 (38) 0.07 (40) 0.06 (42) 0.06 (34) 0.10 (45) 0.06 (42) ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± [Fe -0.79 -0.97 -1.02 -0.89 -0.83 -1.00 -0.69 -0.66 -0.76 -0.67 -0.87 -0.72 -0.60 -0.86 -0.91 -0.88 -0.71 -0.68 -0.88 -1.04 -0.82 -0.68 -0.66 -0.58 -0.56 -0.86 -0.79 -0.78 -1.44 -0.72 -0.81 -0.73 -1.51 -0.69 -0.92 -0.89 -0.78 -0.82 -0.88 Star BL190 BL195 BL196 BL197 BL203 BL204 BL205 BL208 BL210 BL211 BL213 BL216 BL218 BL221 BL227 BL228 BL229 BL233 BL239 BL242 BL247 BL250 BL253 BL257 BL258 BL260 BL261 BL262 BL266 BL267 BL269 BL278 BL279 BL295 BL300 BL304 BL311 BL315 BL323 138 chapter 6: HR spectroscopic study of Fornax Field Stars (n) σ ... (0) ... (0) ... (0) ... (0) ... (0) ... (0) ... (0) ... (0) ... (0) ... (0) 0.34 (1) 0.19 (1) 0.19 (1) 0.17 (1) 0.25 (1) 0.17 (1) 0.19 (1) 0.21 (1) 0.17 (1) 0.25 (1) 0.21 (1) 0.17 (1) 0.21 (1) 0.16 (1) 0.32 (1) 0.33 (1) 0.23 (1) 0.15 (1) 0.18 (1) 0.18 (1) 0.21 (1) 0.22 (1) 0.23 (1) 0.44 (1) 0.23 (1) 0.21 (1) 0.18 (1) 0.25 (1) 0.22 (1) 0.20 (1) 0.22 (1) 0.34 (1) ± ± ± ± ± ± ± ± ± ± /Fe] ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ii ...... [Y 0.10 0.08 0.20 0.05 0.06 0.03 0.16 0.01 0.00 0.22 -0.16 -0.34 -0.40 -0.31 -0.32 -0.06 -0.41 -0.19 -0.09 -0.01 -0.10 -0.15 -0.45 -0.08 -0.04 -0.19 -0.13 -0.05 -0.32 -0.25 -0.21 -0.12 n σ Continued on next page 0.49 (2) 0.15 (2) 0.45 (2) 0.16 (3) 0.21 (3) 0.09 (5) 0.16 (1) 0.20 (4) 0.22 (4) 0.13 (3) 0.12 (4) 0.13 (4) 0.18 (3) 0.23 (2) 0.26 (4) 0.21 (3) 0.20 (3) 0.14 (3) 0.11 (3) 0.21 (2) 0.12 (3) 0.18 (3) 0.14 (5) 0.20 (3) 0.14 (2) 0.15 (3) 0.14 (3) 0.13 (2) 0.27 (3) 0.14 (3) 0.17 (4) 0.12 (4) 0.13 (4) 0.20 (2) 0.12 (4) 0.16 (2) 0.21 (3) 0.23 (3) 0.10 (3) 0.17 (3) 0.19 (4) 0.11 (4) /Fe] ii ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± [Ti 0.18 0.03 0.28 0.09 0.27 0.45 0.27 0.33 0.10 0.21 0.30 0.21 0.24 0.11 0.36 0.07 0.14 0.15 0.17 0.60 0.41 0.02 0.26 0.15 0.33 0.35 0.31 0.45 0.28 0.01 0.37 0.40 0.01 -0.12 -0.01 -0.11 -0.05 -0.10 -0.08 -0.15 -0.19 -0.01 (n) σ 0.15 (1) 0.13 (6) 0.06 (8) 0.14 (7) 0.13 (7) 0.08 (7) 0.06 (9) 0.16 (8) 0.06 (9) 0.06 (9) 0.09 (9) 0.09 (8) 0.07 (8) 0.07 (9) 0.08 (7) 0.08 (8) 0.06 (9) 0.08 (9) 0.08 (7) 0.07 (9) 0.06 (8) 0.06 (9) 0.07 (9) 0.05 (9) 0.12 (7) 0.09 (8) 0.08 (8) 0.06 (9) 0.07 (8) 0.06 (8) 0.15 (8) 0.07 (8) 0.07 (9) 0.11 (9) 0.05 (10) 0.06 (10) 0.07 (10) 0.05 (10) 0.10 (10) 0.08 (11) 0.07 (10) 0.08 (10) /Fe] ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± i ± ± ± ± ± ± ± ± ) and the number of lines used (n). The quoted error [Ti σ 0.56 0.06 0.00 -0.16 -0.37 -0.18 -0.21 -0.05 -0.38 -0.25 -0.23 -0.19 -0.41 -0.13 -0.18 -0.25 -0.19 -0.09 -0.38 -0.02 -0.10 -0.24 -0.38 -0.16 -0.32 -0.40 -0.15 -0.12 -0.43 -0.11 -0.05 -0.23 -0.19 -0.18 -0.23 -0.38 -0.19 -0.19 -0.07 -0.40 -0.11 -0.08 n σ ... (0) 0.17 (4) 0.24 (2) 0.34 (1) 0.21 (2) 0.15 (1) 0.23 (3) 0.10 (4) 0.16 (1) 0.10 (3) 0.15 (3) 0.18 (3) 0.17 (1) 0.08 (4) 0.18 (3) 0.23 (3) 0.17 (4) 0.13 (3) 0.12 (3) 0.12 (4) 0.09 (4) 0.21 (1) 0.20 (3) 0.11 (3) 0.11 (2) 0.31 (2) 0.13 (2) 0.24 (2) 0.18 (3) 0.12 (4) 0.18 (2) 0.13 (3) 0.16 (2) 0.12 (4) 0.10 (3) 0.16 (2) 0.20 (2) 0.13 (3) 0.14 (2) 0.18 (3) 0.20 (3) 0.11 (4) /Fe] ± i ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ... [Si 0.35 0.25 0.16 0.32 0.84 0.48 0.00 0.00 0.03 0.00 0.04 0.29 0.04 0.10 0.03 0.23 0.04 0.04 0.07 0.01 0.18 0.06 -0.16 -0.09 -0.18 -0.45 -0.10 -0.07 -0.08 -0.07 -0.08 -0.08 -0.15 -0.15 -0.21 -0.02 -0.06 -0.18 -0.08 -0.02 -0.02 (n) σ ... (0) ... (0) ... (0) ... (0) ... (0) 0.21 (1) 0.25 (1) 0.14 (2) 0.21 (2) 0.23 (1) 0.16 (1) 0.34 (1) 0.18 (1) 0.19 (1) 0.15 (1) 0.44 (1) 0.19 (1) 0.19 (1) 0.17 (1) 0.23 (1) 0.25 (1) 0.17 (1) 0.19 (1) 0.21 (1) 0.17 (1) 0.21 (1) 0.17 (1) 0.18 (1) 0.21 (1) 0.16 (1) 0.33 (1) 0.25 (1) 0.22 (1) 0.22 (1) 0.16 (2) 0.20 (1) 0.22 (1) 0.18 (1) 0.18 (1) 0.21 (1) 0.34 (1) 0.22 (1) ± ± ± ± ± /Fe] ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± i ...... [O 0.34 0.22 0.28 0.56 0.20 0.25 0.12 0.14 0.05 0.10 0.14 0.34 0.28 0.16 0.37 0.30 0.43 0.44 0.13 0.03 0.28 0.44 0.16 0.05 0.27 0.17 0.28 0.17 0.17 0.30 0.65 0.17 0.13 -0.05 -0.05 -0.11 -0.04 n σ 0.08 (8) 0.11 (9) 0.15 (4) /Fe] i 0.06 (15) 0.07 (11) 0.10 (11) 0.08 (13) 0.06 (14) 0.12 (14) 0.06 (14) 0.07 (13) 0.07 (13) 0.09 (11) 0.08 (12) 0.07 (16) 0.07 (15) 0.07 (12) 0.06 (12) 0.07 (10) 0.05 (16) 0.06 (15) 0.07 (12) 0.08 (14) 0.08 (14) 0.08 (14) 0.06 (15) 0.06 (13) 0.09 (14) 0.08 (13) 0.11 (12) 0.06 (11) 0.09 (15) 0.08 (11) 0.08 (12) 0.06 (16) 0.06 (12) 0.07 (13) 0.06 (13) 0.07 (14) 0.08 (13) 0.09 (13) 0.09 (16) ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± [Ni 0.29 -0.15 -0.34 -0.13 -0.22 -0.28 -0.26 -0.23 -0.24 -0.24 -0.22 -0.09 -0.20 -0.29 -0.10 -0.27 -0.15 -0.24 -0.12 -0.16 -0.33 -0.19 -0.05 -0.10 -0.15 -0.24 -0.19 -0.16 -0.26 -0.22 -0.20 -0.12 -0.11 -0.15 -0.19 -0.13 -0.18 -0.16 -0.20 -0.14 -0.27 -0.24 (n) σ 0.16 (1) 0.21 (3) 0.24 (2) 0.18 (1) 0.15 (3) 0.14 (3) 0.25 (3) 0.17 (3) 0.18 (3) 0.15 (1) 0.10 (3) 0.13 (3) 0.21 (3) 0.10 (3) 0.18 (2) 0.13 (3) 0.18 (3) 0.16 (1) 0.12 (3) 0.16 (3) 0.17 (3) 0.17 (3) 0.17 (3) 0.28 (2) 0.14 (3) 0.12 (3) 0.23 (3) 0.27 (3) 0.25 (1) 0.20 (3) 0.13 (3) 0.31 (3) 0.33 (2) 0.27 (3) 0.11 (3) 0.14 (3) 0.13 (3) 0.11 (3) 0.22 (3) 0.25 (3) 0.24 (2) 0.21 (2) /Fe] ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ii 0.22 0.22 0.49 0.20 0.23 0.36 0.21 0.21 1.53 0.42 0.30 0.27 0.30 0.30 0.34 0.25 0.48 0.34 0.39 0.14 0.37 0.23 0.32 0.25 0.27 0.88 0.73 0.05 0.41 0.31 0.64 0.35 0.51 0.46 0.50 0.49 0.29 0.30 0.46 0.65 0.41 [Nd -0.06 n σ ... (0) ... (0) ... (0) ... (0) ... (0) ... (0) ... (0) ... (0) ... (0) 0.16 (2) 0.34 (1) 0.13 (2) 0.15 (1) 0.31 (2) 0.13 (2) 0.12 (2) 0.16 (2) 0.18 (2) 0.12 (2) 0.21 (1) 0.15 (2) 0.17 (1) 0.21 (2) 0.14 (2) 0.15 (2) 0.18 (1) 0.18 (2) 0.11 (2) 0.32 (1) 0.23 (2) 0.25 (1) 0.16 (2) 0.13 (2) 0.16 (2) 0.16 (2) 0.15 (1) 0.14 (2) 0.16 (2) 0.18 (1) 0.15 (2) 0.24 (2) 0.16 (2) /Fe] ± ± ± ± ± ± ± ± ± i ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ...... [Na Abundance ratio of the Fornax Field stars, where we list the abundance, its associated error ( -0.46 -0.85 -0.67 -0.78 -0.70 -0.70 -0.47 -0.95 -0.77 -0.63 -0.79 -0.36 -0.74 -0.56 -0.98 -0.51 -0.82 -0.91 -0.72 -0.40 -0.47 -0.55 -0.60 -0.64 -0.57 -0.60 -0.67 -0.46 -0.57 -0.47 -0.61 -0.50 -0.49 Star BL038 BL045 BL052 BL065 BL076 BL077 BL079 BL081 BL084 BL085 BL091 BL092 BL096 BL097 BL100 BL104 BL113 BL115 BL123 BL125 BL132 BL135 BL138 BL140 BL141 BL146 BL147 BL148 BL149 BL150 BL151 BL155 BL156 BL158 BL160 BL163 BL166 BL168 BL171 BL173 BL180 BL185 Table 6.A3: is the error on [element/H], not [X/Fe]. Part 2 6.A: Large tables 139 (n) σ ... (0) ... (0) ... (0) ... (0) ... (0) 0.23 (1) 0.18 (1) 0.27 (1) 0.34 (1) 0.19 (1) 0.22 (1) 0.21 (1) 0.26 (1) 0.21 (1) 0.17 (1) 0.24 (1) 0.21 (1) 0.26 (1) 0.20 (1) 0.19 (1) 0.21 (1) 0.20 (1) 0.28 (1) 0.19 (1) 0.20 (1) 0.24 (1) 0.22 (1) 0.21 (1) 0.25 (1) 0.25 (1) 0.22 (1) 0.15 (1) 0.24 (1) 0.23 (1) 0.20 (1) 0.25 (1) 0.25 (1) 0.34 (1) 0.26 (1) ± ± ± ± ± /Fe] ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ii ...... [Y 0.13 0.04 0.15 0.06 0.19 0.11 0.43 0.04 0.10 0.07 0.10 0.08 0.24 0.06 -0.50 -0.05 -0.28 -0.20 -0.14 -0.54 -0.08 -0.05 -0.06 -0.30 -0.46 -0.07 -0.12 -0.17 -0.21 -0.03 -0.06 -0.25 -0.03 -0.33 n σ 0.14 (3) 0.14 (5) 0.17 (4) 0.20 (4) 0.13 (4) 0.27 (2) 0.18 (2) 0.12 (4) 0.10 (4) 0.14 (3) 0.15 (4) 0.28 (3) 0.32 (2) 0.44 (3) 0.12 (4) 0.29 (2) 0.13 (3) 0.19 (3) 0.18 (3) 0.26 (2) 0.20 (3) 0.12 (4) 0.13 (4) 0.24 (3) 0.17 (4) 0.23 (3) 0.18 (4) 0.09 (5) 0.30 (2) 0.12 (4) 0.13 (4) 0.31 (3) 0.12 (4) 0.14 (4) 0.29 (3) 0.12 (5) 0.42 (3) 0.14 (4) 0.17 (3) /Fe] ii ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± [Ti 0.11 0.23 0.32 0.23 0.36 0.17 0.05 0.16 0.08 0.47 0.23 0.16 0.05 0.19 0.16 0.18 0.03 0.16 0.11 0.11 0.50 0.31 0.05 0.10 0.67 0.21 0.09 0.47 0.33 0.07 -0.04 -0.03 -0.10 -0.38 -0.08 -0.25 -0.05 -0.06 -0.12 (n) σ 0.06 (9) 0.12 (8) 0.08 (9) 0.06 (9) 0.06 (9) 0.09 (7) 0.09 (9) 0.10 (9) 0.07 (9) 0.08 (8) 0.07 (9) 0.10 (6) 0.09 (8) 0.07 (9) 0.09 (8) 0.05 (9) 0.09 (7) 0.07 (9) 0.07 (9) 0.07 (8) 0.11 (9) 0.08 (7) 0.11 (9) 0.11 (9) 0.08 (8) 0.07 (9) 0.09 (9) 0.08 (9) 0.07 (8) 0.13 (5) 0.08 (9) 0.07 (8) 0.10 (9) 0.11 (9) 0.07 (9) 0.09 (8) 0.08 (10) 0.09 (10) 0.10 (10) /Fe] ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± i ± ± ± [Ti 0.01 0.16 -0.40 -0.17 -0.28 -0.30 -0.19 -0.20 -0.13 -0.03 -0.14 -0.02 -0.15 -0.26 -0.12 -0.19 -0.24 -0.21 -0.29 -0.23 -0.04 -0.21 -0.19 -0.22 -0.19 -0.27 -0.12 -0.13 -0.16 -0.23 -0.10 -0.04 -0.27 -0.33 -0.33 -0.34 -0.19 -0.23 -0.13 n σ 0.18 (1) 0.34 (1) 0.26 (4) 0.10 (5) 0.10 (6) 0.15 (2) 0.18 (2) 0.18 (3) 0.28 (3) 0.20 (1) 0.20 (4) 0.12 (3) 0.32 (3) 0.14 (2) 0.18 (2) 0.16 (2) 0.16 (3) 0.17 (2) 0.24 (3) 0.13 (5) 0.14 (4) 0.19 (3) 0.14 (3) 0.26 (2) 0.15 (2) 0.12 (3) 0.10 (6) 0.12 (4) 0.13 (3) 0.11 (4) 0.14 (5) 0.18 (3) 0.21 (1) 0.16 (3) 0.10 (7) 0.12 (4) 0.11 (3) 0.15 (5) 0.11 (6) /Fe] i ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± [Si 0.03 0.11 0.38 0.01 0.16 0.09 0.15 0.18 0.30 0.39 0.27 0.34 0.04 0.11 0.12 -0.22 -0.02 -0.03 -0.05 -0.02 -0.16 -0.20 -0.10 -0.08 -0.13 -0.05 -0.03 -0.18 -0.21 -0.13 -0.03 -0.20 -0.39 -0.13 -0.10 -0.04 -0.02 -0.12 -0.04 (n) σ ... (0) ... (0) ... (0) 0.18 (1) 0.19 (1) 0.24 (1) 0.27 (1) 0.34 (1) 0.22 (1) 0.21 (1) 0.22 (1) 0.21 (1) 0.25 (1) 0.26 (1) 0.21 (1) 0.25 (1) 0.16 (1) 0.24 (1) 0.22 (1) 0.15 (2) 0.15 (1) 0.21 (2) 0.20 (1) 0.24 (1) 0.23 (1) 0.19 (1) 0.27 (1) 0.20 (1) 0.20 (1) 0.28 (1) 0.25 (1) 0.19 (1) 0.20 (1) 0.34 (1) 0.21 (1) 0.26 (1) 0.22 (1) 0.21 (1) 0.25 (1) ± ± ± /Fe] ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± i ...... [O 0.09 0.10 0.19 0.13 0.40 0.15 0.16 0.15 0.29 0.24 0.09 0.16 0.29 0.22 0.24 0.11 0.19 0.23 0.05 0.15 0.17 0.03 0.25 0.11 0.38 0.18 0.27 0.16 0.18 0.23 0.15 0.29 0.26 -0.01 -0.02 -0.09 n σ 0.07 (8) 0.11 (9) /Fe] i 0.07 (11) 0.06 (11) 0.07 (13) 0.08 (14) 0.06 (14) 0.06 (16) 0.07 (13) 0.07 (16) 0.06 (15) 0.08 (14) 0.08 (16) 0.07 (15) 0.07 (12) 0.07 (14) 0.09 (14) 0.08 (15) 0.08 (14) 0.06 (13) 0.08 (15) 0.10 (14) 0.07 (17) 0.07 (15) 0.07 (15) 0.07 (12) 0.10 (12) 0.08 (13) 0.07 (10) 0.09 (14) 0.10 (15) 0.07 (14) 0.10 (12) 0.07 (15) 0.07 (14) 0.08 (14) 0.10 (12) 0.11 (15) 0.08 (15) ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± [Ni -0.21 -0.24 -0.31 -0.27 -0.16 -0.21 -0.23 -0.31 -0.12 -0.17 -0.25 -0.10 -0.19 -0.25 -0.19 -0.25 -0.27 -0.24 -0.22 -0.13 -0.07 -0.16 -0.19 -0.20 -0.20 -0.26 -0.28 -0.26 -0.18 -0.19 -0.22 -0.13 -0.03 -0.19 -0.22 -0.19 -0.16 -0.32 -0.23 (n) σ ... (0) 0.31 (2) 0.14 (3) 0.10 (3) 0.17 (2) 0.27 (2) 0.34 (1) 0.30 (2) 0.14 (3) 0.12 (3) 0.13 (3) 0.15 (2) 0.14 (3) 0.15 (3) 0.21 (2) 0.31 (3) 0.16 (1) 0.17 (3) 0.13 (3) 0.13 (3) 0.33 (2) 0.13 (2) 0.23 (3) 0.20 (3) 0.14 (3) 0.33 (3) 0.15 (3) 0.12 (3) 0.26 (2) 0.23 (3) 0.21 (3) 0.14 (3) 0.21 (1) 0.21 (3) 0.11 (3) 0.12 (3) 0.34 (1) 0.21 (1) 0.26 (2) ± /Fe] ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ii ... 0.27 0.03 0.31 0.24 0.30 0.60 0.42 0.36 0.50 0.32 0.36 0.36 0.26 0.51 0.44 0.41 0.40 0.25 0.14 0.32 0.94 0.39 0.36 0.31 0.50 0.36 0.45 0.26 0.12 0.27 0.50 0.66 0.32 0.49 0.21 0.02 0.34 0.16 [Nd n σ ... (0) ... (0) 0.21 (1) 0.13 (2) 0.18 (1) 0.17 (2) 0.19 (2) 0.34 (1) 0.13 (2) 0.16 (2) 0.15 (2) 0.16 (2) 0.21 (1) 0.18 (2) 0.18 (2) 0.21 (1) 0.18 (2) 0.11 (2) 0.17 (2) 0.16 (2) 0.22 (1) 0.21 (1) 0.11 (2) 0.18 (2) 0.14 (2) 0.17 (2) 0.16 (2) 0.19 (1) 0.15 (2) 0.27 (1) 0.14 (2) 0.20 (2) 0.18 (2) 0.18 (2) 0.13 (2) 0.14 (2) 0.24 (2) 0.21 (1) 0.26 (1) /Fe] ± ± i ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ...... [Na 0.04 -0.61 -0.94 -0.70 -0.56 -0.48 -0.48 -0.55 -0.60 -0.60 -0.67 -0.52 -0.81 -0.87 -0.46 -0.86 -0.71 -0.56 -0.74 -0.65 -0.56 -0.37 -0.58 -0.73 -0.66 -0.80 -0.78 -0.93 -0.43 -0.62 -0.17 -0.50 -0.47 -0.79 -0.68 -0.85 -0.76 Star BL190 BL195 BL196 BL197 BL203 BL204 BL205 BL208 BL210 BL211 BL213 BL216 BL218 BL221 BL227 BL228 BL229 BL233 BL239 BL242 BL247 BL250 BL253 BL257 BL258 BL260 BL261 BL262 BL266 BL267 BL269 BL278 BL279 BL295 BL300 BL304 BL311 BL315 BL323

Chapter 7 Conclusions

he object of this thesis has been to determine detailed chemical abundances of indi- T vidual stars in the nearby Fornax dwarf spheroidal galaxy, based on high resolution observations with the VLT spectrographs UVES and FLAMES. Fornax, is one of the few dwarf galaxies to have an extensive population of globular clusters (with 5), and it has also had a complex ﬁeld star formation history dominated by a star formation at intermediate ages. For this thesis, samples of individual stars were studied in both the globular clusters and in the ﬁeld star populations of Fornax.

7.1 New Data Reduction and Analysis Techniques

UVES is an Echelle spectrograph for which classical observations of single stars followed by careful data reduction and analysis star by star is possible. FLAMES is a powerful new multiplexing spectrograph that was more challenging to use. It required new techniques to be developed, from the preparation of the observing run to the last steps of getting the stellar abundances. In the case of FLAMES the 100+ ﬁbres that were typically allocated to scientiﬁc targets made a high degree of automating very important. Another important aspect was the relatively low resolution (of the HR mode) of GIRAFFE and also the limited wavelength coverage. This required careful adjustments and testing of the usual approach applied to UVES data.

An important achievement of this thesis is the method developed to analyse approximately one hundred stellar spectra in a consistent and statistically robust manner, using tools that are typically used on spectra with twice the resolution and larger wavelength coverage. This required bringing together several complex tasks, including accurate stellar atmospheric models, atomic data for the absorption lines, codes of line formation, EW measurements and signal extraction methods, all of which need to be properly included and treated in order to obtain accurate results. We developed a pipeline that delivers stellar parameters and abundances in a controlled manner. This involved developing error analysis and diagnostics to carefully test the robustness of the results. 142 chapter 7: Conclusions 7.2 The Fornax Globular Clusters

The Fornax dSph contains five globular clusters with a range of properties such as metallicity, central concentration and Horizontal Branch structure. Using VLT/UVES we have taken the first high resolution spectra of individual stars in the Fornax globular clusters. We obtained detailed chemical abundances for 9 individual RGB stars in 3 of the 5 For- nax globular clusters. This makes the Fornax globular clusters some of the very few extra-galactic globular clusters that have been studied in this detail. From my results it is clear that Clusters 1, 2 and 3 were formed promptly and early in the history of Fornax dSph. They are over abundant in α-elements (O, Mg, Ca) at a similar level to Galactic clusters at identical [Fe/H], and the heavy element abundances (Y, Ba, Eu) in the 3 clusters are compatible with dominant r-process enrichment. In addition, Cluster 1 is found to be the most metal-poor globular cluster known, although the difference in metallicity between Cluster 1 and M 92 or M 15 in the Milky Way is small.

Thus, despite their very diﬀerent mass, morphology and global star formation history, the abundance patterns of individual stars in the Fornax GCs are almost identical to those found in the Milky Way globular clusters. This extends to the ubiquitous deep- mixing patterns found among globular cluster members that were dected in two stars of Cluster 1 and Cluster 3, and the rare anomalies like the Eu-rich stars of Cluster 3 for which the only Galactic counterpart known to date is M15. This suggests that all globular clusters, regardless of their host galaxy, were formed with the same initial conditions at their epoch of formation, namely the same pre- or self-enriched processes and identical nucleosynthesis patterns.

7.3 Fornax Field stars

Thanks to the multi-fibre capability of FLAMES we have been able to make detailed abundance measurements of a large sample of 81 RGB stars in the central part of For- nax. This is a significant, even dramatic, improvement on the previous UVES sample of 3 individual field stars. Our abundance ratios provide detailed information as to what were the chemical enrichment processes in Fornax, and how they differ from the MW.

We find that Fornax field stars exhibit unusually low α-element ratios, as well as Ni and Na abundances. The [α/Fe] dependence on [Fe/H] is different from the Milky Way, meaning that there has been a different efficiency of chemical enrichment of the ISM. Fornax field stars are clearly predominantly enriched by s-process elements, showing the strong role of (metal poor) AGB stars. This is clearly seen from the high [Ba/Y] ratios, compared to the Milky Way.

Our sample, which was randomly chosen from the entire breadth of the RGB, is dominated by a relatively young, relatively metal rich population (see Figure 7.1). This means that we have obtained the most detailed picture of the chemical enrichment of Fornax during the last ∼4 Gyrs. There is only one ﬁeld star in our sample which appears to be old and metal poor, and its properties are almost indistinguishable from the globular clusters in Fornax, and also from Galactic halo stars at the same [Fe/H]. Figure 7.1 shows the main abundance results (alpha, Ba, and Eu) versus time. The ages were determined 7.3: Fornax Field stars 143

Figure 7.1: Here we show the result of our abundance analysis for alpha elements, and an s- and an r-process element compared to the ages of the individual stars observed in Fornax ﬁeld star population. Representative error bars are shown on the bottom left corner of each panel.

from a colour-magnitude diagram ﬁnding the appropriate isochrone using the detailed spectroscopic abundances (Fe, and alpha). This allows us to determine how the diﬀerent abundances vary with time.

The [α/Fe] ratios were higher in the past and have gradually decreased towards more recent times. This is a sign that SN Ia are becoming increasingly important with time, similar to what we were able to deduce from the [α/Fe] versus [Fe/H] plot of chapter 6. It is clear that the s-process abundances (e.g., [Ba/Fe]) show a slow increase in the contribution with increasing time (and [Fe/H]). This means that as the stellar population 144 chapter 7: Conclusions becomes more metal rich there has been a steady rise in the Ba abundance, and the ages of the stars show that this rise began about 2-4 Gyr ago. The s-process is a much stronger contributor to the chemical evolution of Fornax than it is to the MW, or the Sculptor dSph. This suggests that stellar winds (e.g., from AGB stars) have played a uniquely important role in the (recent, 2-4 Gyr ago) enrichment history of Fornax. There is no such trend in r-process, shown in the [Eu/Fe] panel of Figure 7.1. With the exception of the Eu-rich Cluster 3 points, all the observed stars in Fornax have a more or less constant [Eu/Fe].

Our detailed abundance studies conﬁrm and deepen the diﬃculties found in earlier more limited surveys in understanding the role (if any) of dwarf galaxies, such as Fornax in the build up of our Milky Way. These results also challenge our understanding of basic nucleosynthetic processes, with for example, ratios of [Ni/Fe] that are well below what was typically thought possible.

Further work on Fornax will be to investigate the diﬀerent regions of this surprisingly complex dwarf galaxy in more detail, and speciﬁcally to increase our sample of high resolution abundances for metal poor stars in Fnx. Bibliography

Alonso, A., Arribas, S., & Martínez-Roger, C. 1999a, A&AS, 139, 335 —. 1999b, A&AS, 140, 261 —. 2001, A&A, 376, 1039 Alvarez, R. & Plez, B. 1998, A&A, 330, 1109 Arlandini, C., Käppeler, F., Wisshak, K., Gallino, R., Lugaro, M., Busso, M., & Straniero, O. 1999, ApJ, 525, 886 Asplund, M. 2005, ARA&A, 43, 481 Baade, D., Meisenheimer, K., Iwert, O., Alonso, J., Augusteĳn, T., Beletic, J., Belle- mann, H., Benesch, W., Boehm, A., Boehnhardt, H., Brewer, J., Deiries, S., Delabre, B., Donaldson, R., Dupuy, C., Franke, P., Gerdes, R., Gilliotte, A., Grimm, B., Had- dad, N., Hess, G., Ihle, G., Klein, R., Lenzen, R., Lizon, J.-L., Mancini, D., Muench, N., Pizarro, A., Prado, P., Rahmer, G., Reyes, J., Richardson, F., Robledo, E., Sanchez, F., Silber, A., Sinclaire, P., Wackermann, R., & Zaggia, S. 1999, The Messenger, 95, 15 Barklem, P. S., Christlieb, N., Beers, T. C., Hill, V., Bessell, M. S., Holmberg, J., Marsteller, B., Rossi, S., Zickgraf, F.-J., & Reimers, D. 2005, A&A, 439, 129 Battaglia, G., Tolstoy, E., Helmi, A., Irwin, M. J., Letarte, B., Jablonka, P., Hill, V., Venn, K. A., Shetrone, M. D., Arimoto, N., Primas, F., Kaufer, A., Francois, P., Szeifert, T., Abel, T., & Sadakane, K. 2006, A&A, 459, 423 Belokurov, V., Zucker, D. B., Evans, N. W., Kleyna, J. T., Koposov, S., Hodgkin, S. T., Irwin, M. J., Gilmore, G., Wilkinson, M. I., Fellhauer, M., Bramich, D. M., Hewett, P. C., Vidrih, S., De Jong, J. T. A., Smith, J. A., Rix, H. ., Bell, E. F., Wyse, R. F. G., Newberg, H. J., Mayeur, P. A., Yanny, B., Rockosi, C. M., Gnedin, O. Y., Schneider, D. P., Beers, T. C., Barentine, J. C., Brewington, H., Brinkmann, J., Harvanek, M., Kleinman, S. J., Krzesinski, J., Long, D., Nitta, A., & Snedden, S. A. 2006a, ArXiv Astrophysics e-prints Belokurov, V., Zucker, D. B., Evans, N. W., Wilkinson, M. I., Irwin, M. J., Hodgkin, S., Bramich, D. M., Irwin, J. M., Gilmore, G., Willman, B., Vidrih, S., Newberg, H. J., Wyse, R. F. G., Fellhauer, M., Hewett, P. C., Cole, N., Bell, E. F., Beers, T. C., Rockosi, C. M., Yanny, B., Grebel, E. K., Schneider, D. P., Lupton, R., Barentine, J. C., Brewington, H., Brinkmann, J., Harvanek, M., Kleinman, S. J., Krzesinski, J., Long, D., Nitta, A., Smith, J. A., & Snedden, S. A. 2006b, ApJ, 647, L111 146 BIBLIOGRAPHY

Bensby, T., Feltzing, S., & Lundström, I. 2003, A&A, 410, 527 Bensby, T., Feltzing, S., Lundström, I., & Ilyin, I. 2005, A&A, 433, 185 Bersier, D. 2000, ApJ, 543, L23 Blackwell, D. E. & Lynas-Gray, A. E. 1998, A&AS, 129, 505 Bonifacio, P., Hill, V., Molaro, P., Pasquini, L., Di Marcantonio, P., & Santin, P. 2000, A&A, 359, 663 Buonanno, R., Corsi, C. E., Castellani, M., Marconi, G., Fusi Pecci, F., & Zinn, R. 1999, AJ, 118, 1671 Buonanno, R., Corsi, C. E., Fusi Pecci, F., Hardy, E., & Zinn, R. 1985, A&A, 152, 65 Buonanno, R., Corsi, C. E., Zinn, R., Fusi Pecci, F., Hardy, E., & Suntzeﬀ, N. B. 1998, ApJ, 501, L33+ Burbidge, E. M., Burbidge, G. R., Fowler, W. A., & Hoyle, F. 1957, Reviews of Modern Physics, 29, 547 Burris, D. L., Pilachowski, C. A., Armandroﬀ, T. E., Sneden, C., Cowan, J. J., & Roe, H. 2000, ApJ, 544, 302 Carretta, E., Gratton, R., Cohen, J. G., Beers, T. C., & Christlieb, N. 2002, AJ, 124, 481 Carroll, B. W. & Ostlie, D. A. 1996, An Introduction to Modern Astrophysics (An Introduction to Modern Astrophysics, by B.W. Carroll and D.A. Ostlie. Benjamin Cummings, 1996. ISBN 0-201-54730-9.) Cayrel, R., Depagne, E., Spite, M., Hill, V., Spite, F., François, P., Plez, B., Beers, T., Primas, F., Andersen, J., Barbuy, B., Bonifacio, P., Molaro, P., & Nordström, B. 2004, A&A, 416, 1117 Chamberlain, J. W. & Aller, L. H. 1951, ApJ, 114, 52 Chen, Y. Q., Nissen, P. E., Zhao, G., Zhang, H. W., & Benoni, T. 2000, A&AS, 141, 491 Clayton, D. D. 1983, Principles of stellar evolution and nucleosynthesis (Chicago: Uni- versity of Chicago Press, 1983) Coleman, M., Da Costa, G. S., Bland-Hawthorn, J., Martínez-Delgado, D., Freeman, K. C., & Malin, D. 2004, AJ, 127, 832 Coleman, M. G. & Da Costa, G. S. 2005, Publications of the Astronomical Society of Australia, 22, 162 Dekker, H., D’Odorico, S., Kaufer, A., Delabre, B., & Kotzlowski, H. 2000, in Proc. SPIE Vol. 4008, p. 534-545, Optical and IR Telescope Instrumentation and Detectors, Masanori Iye; Alan F. Moorwood; Eds., 534–545 Demers, S., Grondin, L., & Kunkel, W. E. 1990, PASP, 102, 632 Dinescu, D. I., Keeney, B. A., Majewski, S. R., & Girard, T. M. 2004, AJ, 128, 687 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 Einasto, J., Saar, E., Kaasik, A., & Chernin, A. D. 1974, Nature, 252, 111 Fasano, G. & Vio, R. 1988, Bulletin d’Information du Centre de Donnees Stellaires, 35, 191 Fulbright, J. P. 2002, AJ, 123, 404 Fulbright, J. P., McWilliam, A., & Rich, R. M. 2006, ArXiv Astrophysics e-prints BIBLIOGRAPHY 147

Gallart, C., Aparicio, A., Zinn, R., Buonanno, R., Hardy, E., & Marconi, G. 2005, in IAU Colloq. 198: Near-fields cosmology with dwarf elliptical galaxies, ed. H. Jerjen & B. Binggeli, 25–29 Gilmore, G., Wyse, R. F. G., & Norris, J. E. 2002, ApJ, 574, L39 Goerdt, T., Moore, B., Read, J. I., Stadel, J., & Zemp, M. 2006, MNRAS, 368, 1073 Gratton, R., Sneden, C., & Carretta, E. 2004, ARA&A, 42, 385 Gratton, R. G., Carretta, E., Claudi, R., Lucatello, S., & Barbieri, M. 2003, A&A, 404, 187 Gray, D. F. 1992, The Observation and Analysis of Stellar Photospheres (The Ob- servation and Analysis of Stellar Photospheres, by David F. Gray, pp. 470. ISBN 0521408687. Cambridge, UK: Cambridge University Press, June 1992.) Grebel, E. K. 1998, Highlights of Astronomy, 11, 125 Gustafsson, B., Bell, R. A., Eriksson, K., & Nordlund, A. 1975, A&A, 42, 407 Gustafsson, B., Edvardsson, B., Eriksson, K., Mizuno-Wiedner, M., Jørgensen, U. G., & Plez, B. 2003, in ASP Conf. Ser. 288: Stellar Atmosphere Modeling, ed. I. Hubeny, D. Mihalas, & K. Werner, 331–+ Harris, W. E. 1991, ARA&A, 29, 543 Harris, W. E. & Harris, G. L. H. 2002, AJ, 123, 3108 Heger, A. & Woosley, S. E. 2002, ApJ, 567, 532 Heiter, U. & Eriksson, K. 2006, A&A, 452, 1039 Helmi, A., Irwin, M. J., 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 Hill, V. 1997, A&A, 324, 435 Hill, V., Andrievsky, S., & Spite, M. 1995, A&A, 293, 347 Hill, V., François, P., Spite, M., Primas, F., & Spite, F. 2000, A&A, 364, L19 Hodge, P. W. 1961, PASP, 73, 328 Ibata, R. A., Gilmore, G., & Irwin, M. J. 1994, Nature, 370, 194 Iwamoto, K., Brachwitz, F., Nomoto, K., Kishimoto, N., Umeda, H., Hix, W. R., & Thielemann, F.-K. 1999, ApJS, 125, 439 James, G., François, P., Bonifacio, P., Carretta, E., Gratton, R. G., & Spite, F. 2004, A&A, 427, 825 Johnson, J. A. & Bolte, M. 2001, ApJ, 554, 888 Jorgensen, U. G. & Jimenez, R. 1997, A&A, 317, 54 Kappeler, F., Beer, H., & Wisshak, K. 1989, Reports of Progress in Physics, 52, 945 Koch, A., Grebel, E. K., Kleyna, J. T., Wilkinson, M. I., Harbeck, D. R., Gilmore, G. F., Wyse, R. F. G., & Evans, N. W. 2007, AJ, 133, 270 Koch, A., Wilkinson, M. I., Kleyna, J. T., Gilmore, G. F., Grebel, E. K., Mackey, A. D., Evans, N. W., & Wyse, R. F. G. 2006, ArXiv Astrophysics e-prints Kustner, F. 1921, Veroeffentlichungen des Astronomisches Institute der Universitaet Bonn, 15, 1 Langer, G. E., Hoffman, R., & Sneden, C. 1993, PASP, 105, 301 Langer, G. E. & Hoffman, R. D. 1995, PASP, 107, 1177 Lecureur, A., Hill, V., Zoccali, M., Barbuy, B., Gomez, A., Minniti, D., Ortolani, S., & 148 BIBLIOGRAPHY

Renzini, A. 2006, ArXiv Astrophysics e-prints Letarte, B., Hill, V., Jablonka, P., Tolstoy, E., François, P., & Meylan, G. 2006, A&A, 453, 547 Lotz, J. M., Miller, B. W., & Ferguson, H. C. 2004, ApJ, 613, 262 Mac Low, M. & Ferrara, A. 1999, ApJ, 513, 142 Mateo, M., Olszewski, E., Welch, D. L., Fischer, P., & Kunkel, W. 1991, AJ, 102, 914 Mateo, M. L. 1998, ARA&A, 36, 435 Mathews, G. J., Bazan, G., & Cowan, J. J. 1992, ApJ, 391, 719 Matteucci, F. 2003, Ap&SS, 284, 539 Mayer, L., Governato, F., Colpi, M., Moore, B., Quinn, T., Wadsley, J., Stadel, J., & Lake, G. 2001, ApJ, 547, L123 McWilliam, A. 1997, ARA&A, 35, 503 McWilliam, A., Preston, G. W., Sneden, C., & Searle, L. 1995, AJ, 109, 2757 Monaco, L., Bellazzini, M., Bonifacio, P., Ferraro, F. R., Marconi, G., Pancino, E., Sbordone, L., & Zaggia, S. 2005, A&A, 441, 141 Moore, B., Ghigna, S., Governato, F., Lake, G., Quinn, T., Stadel, J., & Tozzi, P. 1999, ApJ, 524, L19 Navarro, J. F., Frenk, C. S., & White, S. D. M. 1995, MNRAS, 275, 56 Nissen, P. E. & Schuster, W. J. 1997, A&A, 326, 751 Olszewski, E. W., Mateo, M., Harris, J., Walker, M. G., Coleman, M. G., & Da Costa, G. S. 2006, AJ, 131, 912 Pagel, B. E. J. 1997, Nucleosynthesis and Chemical Evolution of Galaxies (Nucleosyn- thesis and Chemical Evolution of Galaxies, by Bernard E. J. Pagel, pp. 392. ISBN 0521550610. Cambridge, UK: Cambridge University Press, October 1997.) Pasquini, L., Avila, G., Blecha, A., Cacciari, C., Cayatte, V., Colless, M., Damiani, F., de Propris, R., Dekker, H., di Marcantonio, P., Farrell, T., Gillingham, P., Guinouard, I., Hammer, F., Kaufer, A., Hill, V., Marteaud, M., Modigliani, A., Mulas, G., North, P., Popovic, D., Rossetti, E., Royer, F., Santin, P., Schmutzer, R., Simond, G., Vola, P., Waller, L., & Zoccali, M. 2002, The Messenger, 110, 1 Plez, B., Brett, J. M., & Nordlund, A. 1992, A&A, 256, 551 Pompeia, L., Hill, V., Spite, M., Cole, A., Primas, F., Romaniello, M., Pasquini, L., Cioni, M. ., & Smecker-Hane, T. 2006, ArXiv Astrophysics e-prints Pont, F., Zinn, R., Gallart, C., Hardy, E., & Winnick, R. 2004, AJ, 127, 840 Pritzl, B. J., Venn, K. A., & Irwin, M. 2005, AJ, 130, 2140 Putman, M. E., Staveley-Smith, L., Freeman, K. C., Gibson, B. K., & Barnes, D. G. 2003, ApJ, 586, 170 Raiteri, C. M., Gallino, R., & Busso, M. 1992, ApJ, 387, 263 Ramírez, I. & Meléndez, J. 2005, ApJ, 626, 465 Reddy, B. E., Lambert, D. L., & Allende Prieto, C. 2006, MNRAS, 367, 1329 Reddy, B. E., Tomkin, J., Lambert, D. L., & Allende Prieto, C. 2003, MNRAS, 340, 304 Russell, S. C. & Bessell, M. S. 1989, ApJS, 70, 865 Sandage, A. 1970, ApJ, 162, 841 Saviane, I., Held, E. V., & Bertelli, G. 2000, A&A, 355, 56 Searle, L. & Zinn, R. 1978, ApJ, 225, 357 BIBLIOGRAPHY 149

Seth, A., Olsen, K., Miller, B., Lotz, J., & Telford, R. 2004, AJ, 127, 798 Shapley, H. 1938, Nature, 142, 715 Shetrone, M., Venn, K. A., Tolstoy, E., Primas, F., Hill, V., & Kaufer, A. 2003, AJ, 125, 684 Shetrone, M. D. 1996a, AJ, 112, 1517 —. 1996b, AJ, 112, 2639 Shetrone, M. D., Bolte, M., & Stetson, P. B. 1998, AJ, 115, 1888 Shetrone, M. D., Côté, P., & Sargent, W. L. W. 2001, ApJ, 548, 592 Smith, V. V., Hinkle, K. H., Cunha, K., Plez, B., Lambert, D. L., Pilachowski, C. A., Barbuy, B., Meléndez, J., Balachandran, S., Bessell, M. S., Geisler, D. P., Hesser, J. E., & Winge, C. 2002, AJ, 124, 3241 Sneden, C., Gratton, R. G., & Crocker, D. A. 1991, A&A, 246, 354 Sneden, C., Ivans, I. I., & Fulbright, J. P. 2004, in Origin and Evolution of the Elements, 172 Sneden, C., Kraft, R. P., Shetrone, M. D., Smith, G. H., Langer, G. E., & Prosser, C. F. 1997, AJ, 114, 1964 Sneden, C., Pilachowski, C. A., & Kraft, R. P. 2000, AJ, 120, 1351 Spite, M. 1967, Annales d’Astrophysique, 30, 211 Stephens, A. & Boesgaard, A. M. 2002, AJ, 123, 1647 Stetson, P. B., Hesser, J. E., & Smecker-Hane, T. A. 1998, PASP, 110, 533 Stoehr, F., White, S. D. M., Tormen, G., & Springel, V. 2002, MNRAS, 335, L84 Strader, J., Brodie, J. P., Forbes, D. A., Beasley, M. A., & Huchra, J. P. 2003, AJ, 125, 1291 Timmes, F. X., Brown, E. F., & Truran, J. W. 2003, ApJ, 590, L83 Timmes, F. X., Woosley, S. E., & Weaver, T. A. 1995, ApJS, 98, 617 Tinsley, B. M. 1979, ApJ, 229, 1046 Tolstoy, E., Hill, V., Irwin, M., Helmi, A., Battaglia, G., Letarte, B., Venn, K., Jablonka, P., Shetrone, M., Arimoto, N., Abel, T., Primas, F., Kaufer, A., Szeifert, T., Francois, P., & Sadakane, K. 2006, The Messenger, 123, 33 Tolstoy, E., Irwin, M. J., Cole, A. A., Pasquini, L., Gilmozzi, R., & Gallagher, J. S. 2001, MNRAS, 327, 918 Tolstoy, E., Irwin, M. J., Helmi, A., Battaglia, G., Jablonka, P., Hill, V., Venn, K. A., Shetrone, M. D., Letarte, B., Cole, A. A., Primas, F., Francois, P., Arimoto, N., Sadakane, K., Kaufer, A., Szeifert, T., & Abel, T. 2004, ApJ, 617, L119 Tolstoy, E., Venn, K. A., Shetrone, M., Primas, F., Hill, V., Kaufer, A., & Szeifert, T. 2003, AJ, 125, 707 Travaglio, C., Galli, D., Gallino, R., Busso, M., Ferrini, F., & Straniero, O. 1999, ApJ, 521, 691 Travaglio, C., Gallino, R., Arnone, E., Cowan, J., Jordan, F., & Sneden, C. 2004, ApJ, 601, 864 Travaglio, C., Hillebrandt, W., & Reinecke, M. 2005, A&A, 443, 1007 Truran, J. W. 1981, A&A, 97, 391 Tsujimoto, T., Nomoto, K., Yoshii, Y., Hashimoto, M., Yanagida, S., & Thielemann, F.-K. 1995, MNRAS, 277, 945 150 BIBLIOGRAPHY van den Bergh, S. 1981, A&AS, 46, 79 —. 2000, The Galaxies of the Local Group (The galaxies of the Local Group, by Sid- ney Van den Bergh. Published by Cambridge, UK: Cambridge University Press, 2000 Cambridge Astrophysics Series Series, vol no: 35, ISBN: 0521651816.) —. 2006, AJ, 131, 304 Venn, K. A. 1999, ApJ, 518, 405 Venn, K. A., Irwin, M., Shetrone, M. D., Tout, C. A., Hill, V., & Tolstoy, E. 2004, AJ, 128, 1177 —. 2006, AJ, 132, 1726 Walker, M. G., Mateo, M., Olszewski, E. W., Bernstein, R., Wang, X., & Woodroofe, M. 2006, AJ, 131, 2114 White, S. D. M. & Rees, M. J. 1978, MNRAS, 183, 341 Wilkinson, M. I., Kleyna, J. T., Wyn Evans, N., Gilmore, G. F., Read, J. I., Koch, A., Grebel, E. K., & Irwin, M. J. 2006, in EAS Publications Series, ed. G. A. Mamon, F. Combes, C. Deﬀayet, & B. Fort, 105–112 Willman, B., Blanton, M. R., West, A. A., Dalcanton, J. J., Hogg, D. W., Schneider, D. P., Wherry, N., Yanny, B., & Brinkmann, J. 2005a, AJ, 129, 2692 Willman, B., Dalcanton, J. J., Martinez-Delgado, D., West, A. A., Blanton, M. R., Hogg, D. W., Barentine, J. C., Brewington, H. J., Harvanek, M., Kleinman, S. J., Krzesinski, J., Long, D., Neilsen, Jr., E. H., Nitta, A., & Snedden, S. A. 2005b, ApJ, 626, L85 Woosley, S. E. & Weaver, T. A. 1995, ApJS, 101, 181 Wyse, R. F. G. & Gilmore, G. 1995, AJ, 110, 2771 Wyse, R. F. G., Gilmore, G., Norris, J. E., Wilkinson, M. I., Kleyna, J. T., Koch, A., Evans, N. W., & Grebel, E. K. 2006, ApJ, 639, L13 Young, L. M. 1999, AJ, 117, 1758 Zinn, R. & West, M. J. 1984, ApJS, 55, 45 Zoccali, M., Barbuy, B., Hill, V., Ortolani, S., Renzini, A., Bica, E., Momany, Y., Pasquini, L., Minniti, D., & Rich, R. M. 2004, A&A, 423, 507 Zoccali, M., Lecureur, A., Barbuy, B., Hill, V., Renzini, A., Minniti, D., Momany, Y., Gómez, A., & Ortolani, S. 2006, A&A, 457, L1 Zucker, D. B., Belokurov, V., Evans, N. W., Kleyna, J. T., Irwin, M. J., Wilkinson, M. I., Fellhauer, M., Bramich, D. M., Gilmore, G., Newberg, H. J., Yanny, B., Smith, J. A., Hewett, P. C., Bell, E. F., Rix, H.-W., Gnedin, O. Y., Vidrih, S., Wyse, R. F. G., Willman, B., Grebel, E. K., Schneider, D. P., Beers, T. C., Kniazev, A. Y., Barentine, J. C., Brewington, H., Brinkmann, J., Harvanek, M., Kleinman, S. J., Krzesinski, J., Long, D., Nitta, A., & Snedden, S. A. 2006a, ApJ, 650, L41 Zucker, D. B., Belokurov, V., Evans, N. W., Wilkinson, M. I., Irwin, M. J., Sivarani, T., Hodgkin, S., Bramich, D. M., Irwin, J. M., Gilmore, G., Willman, B., Vidrih, S., Fellhauer, M., Hewett, P. C., Beers, T. C., Bell, E. F., Grebel, E. K., Schneider, D. P., Newberg, H. J., Wyse, R. F. G., Rockosi, C. M., Yanny, B., Lupton, R., Smith, J. A., Barentine, J. C., Brewington, H., Brinkmann, J., Harvanek, M., Kleinman, S. J., Krzesinski, J., Long, D., Nitta, A., & Snedden, S. A. 2006b, ApJ, 643, L103 Nederlandse samenvatting

it proefschrift draagt de titel “Chemische analyse van het Fornax dwerg sterrenstelsel”, en D het belangrĳkste doel van dit proefschrift is om te bepalen wat de chemische elementen in de sterren van dit sterrenstelsel zĳn. Dit willen we achterhalen om zo te proberen de evolutie van dit sterrenstelsel te begrĳpen. Sterrenstelsels zĳn geen “statische” objecten; ze bewegen, ze vormen sterren en ze kunnen interacties met andere sterrenstelsels hebben. Het bestuderen van de sterren die een sterrenstelsel vormen kan ons informatie over het verleden van het stelsel verschaﬀen. Sommige sterren kunnen zo oud zĳn als het sterrenstelsel zelf, terwĳl andere sterren veel jonger kunnen zĳn. Deze informatie kunnen we gebruiken om voor de gehele geschiedenis van de sterformatie in dit sterrenstelsel te bestuderen hoe de spectra van de sterren zĳn veranderd in de tĳd. De interpretatie van de resultaten is gebaseerd op de goed begrepen fysica van steratmosferen. Hier zal ik de achtergrond en de resultaten van dit proefschrift samenvatten.

Wat zĳn dwerg sterrenstelsels?

Dwerg sterrenstelsels behoren in principe tot het meest simpele type stelsels dat we kennen. Door deze stelsels te bestuderen kunnen vele theorieën over de formatie en evolutie van sterren en sterrenstelsels over een breed scala van omgevingen getest worden. Bolvormige dwerg sterrenstelsels zĳn kleine, bĳna ronde sterrenstelsels die men voornamelĳk in de nabĳheid van grotere stelsels, zoals de Melkweg, vindt. Normaal gesproken is er in deze stelsels geen stervorming en wordt er geen gas met deze stelsels geassociëerd. De abondantie ratio van de verschillende elementen in individuele sterren van verschillende leeftĳden verschaﬀen ons een gedetailleerde kĳk op de verschillende verrĳkkings-processen (Oftewel, supernovas, ster winden) die op hun buurt ons begrip vergroten van de globale processen van de formatie en evolutie van het gehele sterrenstelsel.

De formatie van de elementen

Op dit moment wordt aangenomen dat het Heelal is begonnen als een explosie, the Big Bang. In deze explosie werden waterstof, deuterium, helium en lithium gecreërd. Deze elementen worden dus gezien als de primaire elementen en alle andere elementen zĳn naderhand gevormd door middel van nucleosynthesis in de sterren. Nucleosynthese is dus verantwoordelĳk voor 152 Nederlandse samenvatting bĳna alles wat we vandaag de dag op aarde zien. Dit proces werd in de jaren ’50 voor het eerst uitgelegd in werk van Fowler en Hoyle, wat uiteindelĳk leidde tot het B2FH (Burbidge, Burbidge, Fowler, & Hoyle 1957) artikel.

Het eerste, meest fundamentele proces om waterstof in zwaardere elementen te veranderen, is waterstof verbranding. Dit is het proces waarbĳ waterstof in helium wordt veranderd. In lage massa sterren met een lage temperatuur in de kern gebeurt dit via de proton-proton ketting en in de zwaardere sterren met hogere temperaturen via het invangen van protonen door koolstof, stikstof en zuurstof atomen (in de CNO cyclus). Het volgende stadium is helium verbranding. Daar waar helium zich verzameld in de kern van de ster, zal de kern gaan inkrimpen totdat de temperatuur en de dichtheid hoog genoeg zĳn voor een reactie waarbĳ helium de brandstof is. Hierna volgt shell burning: koolstof verbranding, zuurstof verbranding, silicium verbranding. Deze processen kunnen uiteindelĳk elementen vormen zo zwaar als 56Fe. Dit is het zwaarste element dat door kernfusie in de kern van een ster kan worden gevormd.

De belangrĳkste zwaardere elementen worden alpha elementen genoemd. Deze elementen hebben een kern die meervouden zĳn van helium. Dit zĳn de elementen O, Mg, Ca, Si en Ti.Deze elementen worden voornamelĳk gevormd door alpha vangsten tĳdens de verschillende verbrandings fases van de massieve sterren. Deze elementen worden door supernova II explosies gedeponeerd in het interstellaire medium. En andere belangrĳke groep elementen is de ĳzerpiek. Deze groep wordt voornamelĳk geproduceerd door SN Ia explosies en ook door deze explosie terug gebracht in het ISM. Men verondersteld dat SN Ia explosie worden veroorzaakt door de explosie van een witte dwerg die een dubbelster vormt met een minder zware progenitor ster. Deze explosies komen voornamelĳk voor ∼1 Gyr na de eerste stervormings episode, in tegenstelling tot SN II explosies die kort levende massieve sterren als voorlopers hebben (Dit kan minder dan ∼ 5 − 10 Myrs zĳn).

Zwaardere elementen die voorbĳ de ĳzerpiek liggen worden gecreërd door het invangen van neutronen. De twee belangrĳkste processen (vanuit een sterrenkundig oogpunt) hierbĳ zĳn het s- en het r- proces. Als de neutronen flux niet hoog is dan krĳgen we het s-proces (of slow-process). Hierbĳ zĳn de intervallen tussen het invangen van een neutron lang vergeleken met de beta verval tĳdsschaal van een instabiele kern. De voorwaarden waaronder dit proces zich voordoet komen voornamelĳk voor in de omhulsels van thermisch pulserende AGB sterren en zĳn het meest effectief in sterren met massa’s 3-5 M . Het r-proces (of rapid-process) manifesteert zich als de neutronen flux hoog genoeg is voor het snel invangen van neutronen. Deze omstandigheden verwacht men voornamelĳk in omgevingen zoals die gevormd woorden door SNe II. Als de neutronen zo snel achter elkaar worden ingevangen kunnen ze accumuleren op instabiele kernen voordat deze kernen tĳd hebben voor een alpha of beta verval. De sterren die verantwoordelĳk zĳn voor deze explosies zĳn massief. Deze sterren hebben daarom een korte levensduur en worden verondersteld de eerste sterren te zĳn die zware elementen in het ISM brengen. Nederlandse samenvatting 153 Abondanties in sterrenstelsels

Tĳdens het leven van een ster worden de abondanties van elementen gepreserveerd∗ op het oppervlakte van de ster. Deze abondanties kunnen relatief makkelĳk worden achterhaald door het meten van absorptie lĳnen in spectra met een hoge resolutie. Hierdoor zĳn de abondanties van de elementen een belangrĳk instrument voor het begrĳpen van de oorsprong van een ster populatie geworden. De abondanties van verschillende elementen kunnen worden gemeten in sterren van verschillende leeftĳden. Dankzĳ de verschillende nucleosynthetische oorsprong van de verschillende elementen kunnen we achterhalen wanneer welk verrĳkings proces dominant was in de geschiedenis van het sterrenstelsel. Niet geheel onverwacht, hebben de eerste studies zich geconcentreerd op de Melkweg en pas onlangs zĳn vergelĳkbaar gedetailleerde studies gedaan in andere sterrenstelsels zoals de Magellaanse wolken en meest recentelĳk de dichtbĳ gelegen bolvormige dwerg sterrenstelsels.

Deze studies suggereren dat de ster populaties van de satelliet stelsels die we vandaag de dag observeren geen signiﬁcante bĳdrage kunnen leveren aan de ster populatie van de Melk- weg. Een mogelĳke uitzondering hierop is de buiten-halo van de de Melkweg. In vergelĳking met de Melkweg zĳn bolvormige dwerg stelsels simpelere stelsels. De meeste van deze stelsel hebben een veel lagere hoeveelheid stervorming dan de Melkweg en al deze stelsels hebben een verschillend en uniek stervormings verleden. In dit proefschrift worden abondantie ratio bestudeerd in het Fornax dSph. Voor de eerste keer is het sample groot genoeg voor een gedetailleerde studie van de interne evolutie van deze stelsels en om het verschil met onze Melkweg te kwantiﬁceren.

Het kosmologisch belang van dwerg sterrenstelsels

Het simpelste model voor de formatie van sterrenstelsels is dat alle sterrenstelsels worden gevormd door snel in elkaar te storten in het vroege Heelal. (het zogenoemde monolitisch in elkaar storten, of Eggen, Lynden-Bell, & Sandage 1962). Vervolgens evolueren deze stelsel over de tĳd uitsluitend door hun gas in sterren te veranderen. Dit model verondersteld dat de het overgrote deel van de massa in sterrenstelsels al bĳ het vormen van de stelsels op zĳn plaats was. Echter, dit model is herzien (e.g., Searle & Zinn 1978) tot een model dat verondersteld dat sterrenstelsels niet in een keer worden gevormd maar dat ze worden opgebouwd uit kleinere segmenten. Tegelĳkertĳd met deze theorie kwam er het zeer succesvolle “cold dark matter” (CDM) idee voor structuur formatie in het Heelal. CDM verondersteld dat de donkere materie in een sterrenstelsel word opgebouwd door het continu invangen van kleine klonten om zo de stelsels en cluster die we vandaag de dag zien te vormen (e.g., White & Rees 1978; Navarro, Frenk, & White 1995).

Het aantal kleine satellieten rond een groter sterrenstelsel zoals het onze, lĳkt door CDM te worden overschat. Dit probleem staat bekent als het “missende dwerg probleem” (e.g., Moore et al. 1999). De laatste jaren zĳn er vele zwakke satellieten rond om de Melkweg ontdekt en deze ontdekkingen hebben onze ideeën over de lokale groep verandert. Deze ontdekkingen suggereren dat de bolvormige dwerg stelsels die we tot vandaag de dag hebben bestudeerd nog maar het puntje van de ĳsberg zĳn. Deze stelsels zĳn de meest massieve

∗ Behalve enkele lichte elementen die door intern mixen kunnen worden beïnvloed: Li, C, N. 154 Nederlandse samenvatting satellieten van een grotere populatie van zwakkere satellieten met een lagere massa.

Wat hebben we geleerd?

Een belangrĳk aspect van dit proefschrift is de pipeline die ontwikkeld is voor het analyseren van een grote hoeveelheid stellaire spectra (∼ 100) op een consistente en statisch gezien robuuste manier. Deze pipeline gebruikt gereedschappen die normaal gesproken worden gebruikt voor spectra met twee keer hogere resolutie en veel groter bereik in golﬂengte. Dit vereiste het samenbrengen van een stel complexe taken zoals, accurate ster modellen, atomai- re data voor de absorptie lĳnen, codes voor lĳn formatie, EW metingen en signaal extractie methodes. Al deze taken moesten op de juiste wĳze in de pipeline worden geïncorpereerd en op de juiste wĳze worden gebruikt om nauwkeurige resultaten te krĳgen. De pipeline produ- ceert stellaire parameters en abondanties op een gecontroleerde manier. Dit vereiste ook de ontwikkeling van een fouten analyse en diagnostiek zodat de robuustheid van de resultaten kon worden getest.

Het Fornax bolvormige dwerg sterrenstelsel heeft vĳf globular clusters (GCs) met een breedte aan eigenschappen. Door gebruik te maken van de VLT/UVES heb ik de eerste gedetailleerde chemische abondanties van negen individuele sterren in drie van deze GCs ver- kregen. Uit onze resultaten is het duidelĳk dat deze GCs direct en vroeg in de geschiedenis van het Fornax dSph gevormd zĳn, netzoals de GCs in onze Melkweg. Ondanks het feit dat deze Fornax GCs in massa, morfologie en globale ster formatie geschiedenis zeer verschillen, zĳn de abondantie patronen van individuele sterren in deze GCs bĳna gelĳk aan de patronen die worden gevonden in GCs van de Melkweg. Ook abondantie patronen gevonden in sterren zeer speciﬁek voor GCs (diep-mixen) en zeldzame anomalieën (europium-rĳk) zĳn bĳna identiek. Dit suggereert dat sterren in GCs identiek zĳn ongeacht de grootte of het type stelsel waarmee zĳ geassocieerd worden.

Dankzĳ de multi-fibre mogelĳkheid van VLT/FLAMES heb ik gedetailleerde abondantie metingen kunnen doen van 81 RGB sterren in het centrale deel van Fornax. Dit is een signifi- cante, zelfs dramatische, verbetering van het vorige UVES sample dat bestond uit slecht drie sterren in het veld. Dit sample van Fornax veld sterren toont ongewoon lage [α/Fe] ratio, en de afhankelĳkheid met metallicity is verschillend van de Melkweg. Dit impliceerd een verschillende efficiëntie in gas verrĳking. Fornax veld sterren zĳn overduidelĳk voornamelĳk verrĳkt door s-proces elementen met een hoge metallicity. Dit toont de belangrĳke rol van (metaal arme) AGB sterren aan. Ons sample word gedomineerd dooreen relatief jonge, metaal rĳke ster populatie. Dit betekend dat we de chemische verrĳking van Fornax gedurende de laatste ∼4 Gyrs het meest gedetailleerd kunnen bekĳken. Ons sample bevat één veld ster en deze lĳkt oud en metaal arm te zĳn. De abondantie eigenschappen van deze ster zĳn niet te onderscheiden van de sterren in de globular clusters van Fornax.

Deze resultaten bevestigen en verdiepen de moeilĳkheden naar voren gebracht door eerdere meer gelimiteerde surveys in het begrĳpen van de rol die dit en vergelĳkbare stelsels hebben gespeeld bĳ het opbouwen van de Melkweg. Ook brengen deze resultaten een uitdaging ons begrip van de nucleosynthese. De ratio voor [Ni/Fe] zĳn bĳvoorbeeld ver onder het niveau dat normaal gesproken mogelĳk werd geacht. Résumé français

ette thèse s’intitule « Analyse chimique de la galaxie naine Fornax » et son but principal C est de tenter de comprendre l’évolution de cette dernière en déterminant les éléments chimiques présents dans ses étoiles. Les galaxies ne sont pas des objets « statiques » ; elles bougent, elles forment des étoiles et elles peuvent interagir avec d’autres galaxies. L’étude des étoiles composant une galaxie peut, en principe, nous informer sur son passé. Ainsi, certaines étoiles peuvent être aussi vieille que la galaxie elle-même ; d’autres peuvent être beaucoup plus jeunes. L’information sur le passé d’une galaxie peut servir à étudier la façon dont les spectres stellaires ont varié pendant la période où les étoiles se sont formées dans dans ladite galaxie. L’interprétation des résultats repose sur la physique des atmosphères stellaires.

Que sont les galaxies naines ?

En principe, les galaxies naines forment le type de galaxie le plus simple. On peut les étudier pour tester diverses théories sur la formation ainsi que l’évolution des étoiles et des galaxies dans un éventail d’environnements. Les galaxies naines sphéroïdes sont petites, grosso modo sphériques et sont habituellement trouvées dans les environs de galaxies plus larges, comme la Voie lactée. Elles n’ont généralement pas d’étoiles en formation ni semblent-elles avoir de gaz qui leur est associé. Les ratios d’abondance de diﬀérents éléments dans des étoiles d’âges divers fournissent un aperçu détaillé des processus variés d’enrichissement chimique (supernovas, vents stellaires) qui à leur tour, améliorent notre compréhension des processus globaux de la formation et de l’évolution des galaxies.

Formation des éléments chimiques

On croit que l’Univers a débuté par une explosion, le Big Bang, où l’hydrogène, le deutérium, l’hélium et le lithium ont été créés. Ces derniers sont alors considérés comme des éléments primordiaux, et tous les autres éléments sont formés subséquemment dans les étoiles par nucléosynthèse. La nucléosynthèse stellaire est donc responsable de presque tout ce qui nous entoure aujourd’hui sur Terre. Ceci a d’abord été expliqué dans les années 1950 par le travail de Fowler et Hoyle, et ﬁnalement dans l’article des B2FH (Burbidge, Burbidge, Fowler, & Hoyle 1957). 156 Résumé français

Le tout premier procédé fondamental de conversion d’hydrogène en éléments plus lourds est la combustion d’hydrogène. Il s’agit de convertir un noyau d’hydrogène en hélium, par la chaine proton-proton dans les étoiles de petites masses et dont le coeur a une basse tem- pérature et également par des captures de protons par des atomes de carbone, d’azote et d’oxygène (dans les cycles CNO) dans des étoiles plus massives et de températures plus élevées. L’étape suivante est la combustion de l’hélium. L’hélium s’accumule au coeur de l’étoile et le coeur se contracte jusqu’à ce que la température ainsi que la densité augmente assez pour produire une réaction dans laquelle l’hélium en est le carburant. Ensuite vient la combustion des couches, où respectivement le carbone, l’oxygène et le silicium brulent. Ce procédé peut produire des éléments aussi lourd que le 56Fe qui est l’élément le plus massif pouvant être formé par fusion au coeur d’une étoile.

Les éléments lourds les plus signiﬁcatifs sont appelés les éléments alpha, avec des noyaux qui sont des multiples de l’He (O, Mg, Ca, Si, Ti). Les éléments alpha sont majoritairement synthétisés par la capture d’alpha pendant les multiples phases de combustion dans les étoiles massives et ils sont expulsés dans le milieu interstellaire (interstellar medium – ISM) par des explosions de supernovas de type II (SN II). Un autre groupe important est formé par les éléments du pic du fer (iron peak), incluant le Fe lui-même. Ils sont produits essentiellement et expulsé dans le milieu interstellaire par des explosions de supernovas de type Ia (SN Ia), qu’on pense être dues à l’explosion d’une naine blanche dans un système binaire développé avec une étoile génitrice moins massive. Les SN Ia surviennent habituellement environ 1 mil- liard d’années après le premier épisode de formation d’étoile. Les SN II, quant à elles, ont des étoiles massives génitrices à vie courte, seulement 5-10 millions d’années.

Les éléments plus lourds, au-delà du pic du fer, sont créés par capture de neutrons où les deux procédés les plus importants (dans un contexte astrophysique) sont les processus S et R. Le processus S (lent) survient lorsque le flux de neutrons est peu élevé, pour que les intervalles entre les captures de neutrons sont longs par comparaison avec l’échelle de temps de désintégration bêta d’un noyau instable. Ces conditions sont trouvées dans les enveloppes d’étoiles AGB (Asymptotic Giant Branch – branche asymptotiques des géantes) en pulsation thermique et elles sont plus efficaces dans des étoiles de 3-5 masses solaires. Le processus R (rapide) survient lorsqu’il y a un flux de neutrons suffisant pour permettre une capture rapide de neutrons. On croit que cela survient généralement dans des environnements similaires à ceux produits par les SN II. Avec de telles captures rapides successives, les neutrons peuvent s’accumuler sur un noyau instable avant même de subir une désintégration alpha ou bêta. Les étoiles responsables de ces explosions sont massives, donc elles ont une courte vie, et on croit qu’elles sont les premiers objets fournissant des éléments lourds dans l’ISM.

Abondances dans les galaxies

Parce que les abondances d’éléments sont préservées∗ à la surface d’une étoile pendant toute sa durée de vie, et parce qu’elles peuvent être (relativement) facilement mesurées à partir de raies d’absorbtion dans des spectres stellaires à haute résolution, les abonances sont devenues un outil très important pour comprendre la genèse des populations stellaires. Les abondances de divers éléments peuvent être mesurés dans des étoiles de diﬀérents âges et (en raison de

∗ Sauf pour quelques éléments légers qui peuvent être affectés par mélange interne : Li, C, N. Résumé français 157 leur différente origine nucléosynthétique) elles nous permettent de déduire le procédé d’enrichissement qui a dominé à différentes époques de l’histoire de la galaxie. Les premières études se sont évidemment concentrées sur la Voie lactée (VL), et ce n’est que relativement récemment que des études détaillées similaires ont été faites dans d’autres galaxies, telles les Nuages de Magellan et, plus récemment, les galaxies naines sphéroïdales avoisinantes.

Ces études suggèrent que les populations stellaires des galaxies satellites que nous voyons aujourd’hui ne peuvent avoir contribué de façon signiﬁcative à la population stellaire de notre Galaxie, sauf peut-être pour le halo externe. Les galaxies naines sphéroïdales sont de simples systèmes en comparaison avec la VL. La plupart d’entre elles ont généralement des taux de formation d’étoiles plus faibles et chacunes d’entre elles ont une histoire unique de formation stellaire. Dans cette thèse, un échantillon statistiquement signiﬁcatif de ratios d’abondance est étudié pour la première fois dans la galaxie naine sphéroïdale Fornax. Cette thèse fait une étude détaillée de l’évolution interne de ce système et compare les résultats trouvés avec notre Galaxie.

Importance cosmologique des galaxies naines

Le modèle le plus simple de formation de galaxie stipule que toutes les galaxies se forment dans l’Univers primitif dans un scénario d’effondrement rapide (effondrement monolithique, Eggen, Lynden-Bell, & Sandage 1962). Au fil du temps, ces galaxies se développent ensuite uniquement en changeant leur masse gazeuse en masse stellaire. Ce modèle indique que la majorité de la matière de toutes les galaxies étaient en place à leur formation. Cependant, ce modèle simple a été surclassé (ex. Searle & Zinn 1978) par un modèle qui tient pour acquis que les galaxies ne se forment pas en un seul effondrement, mais plutôt qu’elles se constuisent avec le temps à partir de petits fragments. Cette théorie est venue en parallèle de la très po- pulaire vision de la formation de structure dans l’Univers, « matière sombre froide » (cold dark matter – CDM), qui établit que le contenu en matière sombre d’une galaxie se construit par l’accumulation continu de petits agglomérats, pour finalement former les galaxies et les amas de galaxies que nous voyons aujourd’hui (ex. White & Rees 1978; Navarro, Frenk, & White 1995).

La CDM semblent surestimer le nombre de petites galaxies satellites autour de galaxies plus larges (comme la nôtre) ; il s’agit d’une incohérence connue comme le « problème des naines manquantes » (missing dwarf problem, ex. Moore et al. 1999). Cependant, de récentes découvertes au sujet de plusieurs satellites de faible luminosité présents autour de la VL de- puis quelques années changent notre point de vue du Groupe local, suggérant ainsi que les galaxies naines sphéroïdales étudiées jusqu’à maintenant ne sont que le bout de l’iceberg. Ce sont les satellites les plus massifs d’une population plus large de satellites peu lumineux et de plus petites masses. 158 Résumé français Qu’avons-nous appris ?

Un aspet important de cette thèse est le système semi-automatique développé pour analy- ser un large nombre de spectres stellaires (∼ 100) de manière constante et statistiquement robuste, en utilisant des outils qui sont généralement utilisés sur des spectres ayant une résolution deux fois plus grande et une plus large couverture spectrale. Il a fallu joindre plusieurs tâches complexes, incluant des modèles d’atmosphères stellaires précis, des données atomiques pour les raies d’absorbtion, des codes de formation de raies spectrales, des mé- thodes de mesure de largeur équivalente et d’extraction du signal. Toutes ces tâches doivent être incluses adéquatement et être traitées pour obtenir des résultats précis. Nous obtenons donc des paramètres stellaires et des abondances de façon contrôlée, en tenant compte du calcul d’erreurs pour vériﬁer soigneusement la validité des résultats.

Fornax contient cinq amas globulaires (AG), ayant un éventail de propriétés. En utilisant le VLT/UVES, j’ai obtenu les premières abondances chimiques détaillées de neuf étoiles individuelles dans trois de ses AG. À partir de nos résultats, il appert qu’ils ont été formés rapidement et tôt dans l’histoire de Fornax tout comme l’ont été ceux de la VL. Ainsi, malgré leur masse, leur morphologie et leur histoire globale de formation d’étoiles complètement dif- férentes, les modèles d’abondance des étoiles individuelles dans les AG de Fornax sont presque identiques que ceux trouvés dans les AG de la Voie lactée, incluant des modèles d’abondance spéciﬁques aux amas d’étoiles (brassage des couches profondes, deep-mixing) et de rares anomalies (surabondance en europium) également observés dans d’autres AG. Ceci suggère que les étoiles dans les AG sont les mêmes sans égard pour le type ni pour la taille de la galaxie d’où l’amas globulaire provient.

Grâce à la capacité multifibre de VLT/FLAMES, j’ai pu prendre des mesures détaillées d’abondance de 81 étoiles de la branche des géantes rouges (dans la partie centrale de For- nax). Ceci est une amélioration significative par rapport au précédent échantillon de trois étoiles individuelles du champ central provenant de UVES. Les étoiles de Fornax sont sous- abondantes en éléments alpha, et leur dépendance en métallicité est différente de la VL. Ceci suggère une efficacité différente d’enrichissement du gaz. À haute métallicité, les étoiles sont essentiellement enrichies en éléments du processus S, démontrant le rôle important des étoiles AGB pauvre en métaux. Notre échantillon est dominé par une population relativement jeune et relativement riche en métaux. Nous avons donc obtenu un portrait détaillé de l’enrichissement chimique de Fornax pour les dernières ∼4 milliards d’années. Il n’y a qu’une d’étoiles dans notre échantillon qui semble vieille et pauvre en métaux, et ses propriétés d’abondance sont presque indifférenciable de celles des AG de Fornax.

Ces résultats confirment les difficultés trouvées dans les précédentes études (plus limi- tées) quant à la compréhension du rôle (s’il y en a un) de la galaxie Fornax (ou d’une galaxie similaire) dans la formation de la VL. Ces résultats mettent également au défi notre compré- hension des processus nucléosynthétiques de base, avec, par exemple, des ratios de [Ni/Fe] qui sont nettement plus bas que ce qui est perçu généralement comme possibe. Acknowledgements

There are many people and organisations I would like to thank for their help in the making this thesis. First, I would like to thank the Dutch society for giving me the opportunity to study in the Netherlands, and Groningen for being such a nice place to live. Second, the Netherlands Organisation for Scientiﬁc Research (NWO) for funding my research project, the Kapteyn Astronomical Institute and the Leids Kerkhoven-Bosscha Fonds (LKBF) for funding my travel. And for the printing of this thesis, I’m grateful to the LKBF, the Central Student Administration Oﬃce and the Kapteyn Astronomical Institute.

Thank you Eline and Vanessa for supervising me during the last four years. I know that the last part was tough and that time was ﬂying, but you kept pushing me in the right direction. Your help in reading, correcting and commenting my thesis was invaluable and I am grateful for that. You have been patient with me, given me the chance to retry when I made a mistake, shown me the way when I was lost and always made sure that at least one of you could help me when I was needing it. Thank you Mike, Kim, Matthew and Amina for your contributions to my work all over the years. And a special thank to Pascale: you were one of the most helpful person during my PhD for all the discussions we had together, always ready to give an advice, with your colourful style, direct, funny, and most of the time right. You were never afraid to say what had to be said and I’m glad I have been collaborating with you.

I’m also grateful to the reading committee of my thesis, Monique Spite, Piet van der Kruit and Jan Willem Pel, for their quick reading and insightful comments, helping to improve this thesis in many aspects. Your comments and suggestions were greatly appreciated and will probably help everyone who will my thesis. Many thanks to Bertrand Plez for making the stellar models we needed for Fornax; without your contribution, this work would not have been the same. Thanks to Marco Gullieuszik and Enrico Held for providing JHK photometry for some of our targets, and to Serge Demers for his carbon star list.

Thank you Giuseppina for being my house mate for three years, we really had some fun in this small cosy house on the Nieuwe Kĳk in’t Jatstraat! Thanks for helping me improve my now famous pasta alla Bruno. All the dinners we organised, the movies night, your home made liquors and the memorable parties and wine tasting nights we had (that would inevitably end up in Storm for one last beer). I have some really funny memories about this house, like that time when it was really late in the night and there was some noise outside and you... 160 Acknowledgements

Thank you Yang-Shyang for sharing your ﬂat with me in the last six months of my stay in the Netherlands. You were a great moral support in that diﬃcult time, cooking for me much more that I cooked for you, supporting me everyday in many aspects and trying to make sure that my life was not too boring even when I was working 16 hours per day. You were very understanding, always there for me and never asking anything in return for your generosity.

Thanks to César, Erik, Miguel and Wilfred for being my office mates. Thanks to Katarina, Miguel, Pablo and Fabrice for your moral support while going through the end of your thesis at the same time as me. Thank you Rien for all the political discussions we had (mostly) about Canada. Even though you were wrong most of the time, I enjoyed arguing with you on every occasions, and I will be missing this. Thanks to Bernard, Mark, Nigel, Saleem and Scott for all the political, social and scientific discussions we had. Thanks to Paolo and Simona for painting my beloved moped in pink, that was too generous of you, I will try to return the favour one day. Thanks to Jackie for finding a house for me to live in on the first day I arrived in the Netherlands. Rense, Jelte, Bjorn, Teffie, Edo ×2, Peter, Erwin, Michiel, Chris, Christiaan, Dieter and Wilfred, you helped me on countless occasions when I needed help to translate something in dutch, thank you. Teffie, thanks for making it all the way to Paris to visit me, and for all the dinners and discussions we had. Peter, Jan, Giussepina, Mirjam, Yang-Shyang and Alvaro, thanks for your company in Storm. Dieter, thank you for your guitar lessons and advices, and give my thanks again to your father for his nice historical tour of Gent. Thank you Patricia for your help with my postdoc applications. Thanks to Renzo for your advice on where to eat haring. Fabrice, Philippe and Ulrike, thanks for inviting me for dinner on several occasions, it was nice to have someone to speak french in Groningen. Thanks to Scott and Danny for helping me with Python and Linux. Thank you Eite for your quick response in getting a new desktop computer for me when my laptop failed. Thank you Monty Python for everything you’ve done. Thanks to Jeffrey Lebowski for being such a great character. Thank you Plume and Renaud for all the songs you wrote. Thanks to Pierre Falardeau for “Le temps des bouffons”. Thanks to Denis Drolet for being dressed in brown. Thanks to J.R.R. Tolkien and Douglas Adams for writing excellent stories. Thanks to de Pintelier for the good selection of beers, and to Storm for the good music and ambiance. Thanks to Images for the yearly French film festival. Thanks to Ole, Rense, Jelte, Teffie, Edo, Peter, Dieter, Wilfred, Paolo, Mirjam, Alvaro, Giussepina, Yang-Shyang, Martin, Matias, Facundo, Simona, Mercedes, Katarina, Elif, Seungyoup, Lodovico, Jeronimo, Emilio, César, Erik, Monica, Isabel, Léon, Patricia, Filippo, Philippe, Fabrice, Alicia, Eline and Andrew for all the dinners, parties, game nights and movie nights.

I also want to thank the Observatoire de Paris for welcoming me during my visits, and the CNRS for funding part of these visits. Special thanks to Vanessa, Jean-Noël, Selma, Pascale and Aurélie for taking good care of me while I was in Paris. Thank you Martin, Yuyan, Martine, Jesse and Keylan for your hospitality in December and January, when I became homeless after leaving the Netherlands. Thanks to Peter Kamphuis and Théane Lavigne for making the dutch and french summary for my thesis on such a short notice, 1 day! Thank you Jesse for the nice cover page you made for my thesis. Thank you Sylviane for all your help and ﬁnancial advices. Thank you Céline, Raynald, Martine and Simon for always encouraging me, always being there for me and providing stability in my life for the past 30 years. – Merci!

Bruno Letarte – February 2007 – Pasadena