The positive correlation between planet detection rate and host star iron abundance lends strong support to the core accretion theory of planet formation. However, iron is not the most significant mass contributor to the cores of giant planets. Since giant planet cores are thought to grow from silicate grains with icy mantles, the likelihood of gas giant formation should depend heavily on the oxygen and silicon abundance of the planet formation environment. Here we compare the silicon and oxygen abundances of a set of 76 planet hosts and a control sample of 80 metal-rich stars without any known giant planets. Our new, independent analysis was conducted using high resolution, high signal-to-noise data obtained at McDonald Observatory. Because we do not wish to simply reproduce the known planet-metallicity correlation, we have devised a statistical method for matching the underlying [Fe/H] distributions of our two sets of stars. We find a 99\% probability that planet detection rate depends on the silicon abundance of the host star, over and above the observed planet-metallicity correlation. We do not detect any such correlation for oxygen. Our results would thus seem to suggest that grain nucleation, rather than subsequent icy mantle growth, is the important limiting factor in forming giant planets via core accretion. Based on our results and interpretation, we predict that planet detection should correlate with host star abundance for refractory elements responsible for grain nucleation and that no such trends should exist for the most abundant volatile elements responsible for icy mantle growth. / text

The PPAK Integral Field Spectroscopy (IFS) Nearby Galaxies Survey: PINGS, a 2-dimensional spectroscopic mosaicking of 17 nearby disk galaxies in the optical wavelength range. This project represents the first attempt to obtain continuous coverage spectra of the whole surface of a galaxy in the nearby universe. The final data set comprises more than 50000 individual spectra, covering in total an observed area of nearly 80 arcmin square. The powerful capabilities of wide-field 2D spectroscopic studies areproven. The chemical composition of the whole surface of a spiral galaxy ischaracterised for the first time as a function not only of radius, but of theintrinsic morphology of the galaxy, allowing a more realistic determination oftheir physical properties. The methodology, analysis and results of thisdissertation will hopefully contribute in a significant way to understand thenature of the physical and chemical properties of the gas phase in spiralgalaxies.

Within the currently well-established ΛCDM cosmological framework we still lack a satisfactory un- derstanding of the processes that trigger, regulate and eventually quench star formation on galactic scales. Gas flows (including inflows from the cosmic web and supernovae-driven outflows) are con- sidered to act as self-regulatory mechanisms, generating the scaling relations between stellar mass, star formation rate and metallicity observed in the local Universe by large spectroscopic surveys. These surveys, however, have so far been limited by the availability of only one spectrum per galaxy. The aim of this dissertation is to expand the study of star formation and chemical abundances to resolved scales within galaxies by using integral field spectroscopy (IFS) data, mostly from the ongoing SDSS- IV MaNGA survey. In the first part of this thesis I demonstrate the ubiquitous presence of extended low ionisation emission-line regions (LIERs) in both late- and early-type galaxies. By studying the Hα equivalent width and diagnostic line ratios radial profiles, together with tracers of the underlying stellar popula- tion, I show that LIERs are not due to a central point source but to hot evolved (post-asymptotic giant branch) stars. In light of this, I suggest a new classification scheme for galaxies based on their line emission. By analysing the colours, star formation rates, morphologies, gas and stellar kinematics and environmental properties of galaxies with substantial LIER emission, I identify two distinct popula- tions. Galaxies where the central regions are LIER-like, but show star formation at larger radii are late types in which star formation is slowly quenched inside-out. This transformation is associated with massive bulges. Galaxies dominated by LIER emission at all radii, on the other hand, are red-sequence galaxies harbouring a residual cold gas component, acquired mostly via external accretion. Quiescent galaxies devoid of line emission reside in denser environments, which suggests environmental effects as a likely cause for the existence of line-less galaxies on the red sequence. In the second part of this dissertation I focus on the study of resolved chemical abundances by characterising the gas phase oxygen and nitrogen abundance gradients in a large sample of star forming galaxies. I analyse the deviations from an exponential profile at small and large radii and the dependence of the gradients on stellar mass. These findings are interpreted in the context of the inside-out paradigm of disc growth. I then demonstrate the necessity of gas flows, which are responsible for the observed flattening of the metallicity and N/O ratio gradients at large radii. Finally, I present a case study based on one nearby galaxy (NGC 628), in which I combine IFS and cold gas data to derive a spatially resolved metal budget and estimate the mass of metals lost by the galaxy throughout its life- time. By using simple physically-motivated models of chemical evolution I infer the average outflow loading factor to be of order unity.

Contexte. Les étoiles naissent ensemble dans des nuages moléculaires géants. Si nous faisons l’hypothèse qu’ils étaient à l’origine chimiquement homogènes et bien mélangés, nous nous attendrions à ce que les étoiles issues d’un même nuage aient la même composition chimique. La plupart des groupes d’étoiles sont perturbés lors de leur évolution dans la galaxie et l’information dynamique est perdue. Ainsi la seule possibilité que nous ayons de reconstruire l’histoire de la formation stellaire est d’analyser les abondances chimiques que l’on observe aujourd’hui.But. La technique de marquage chimique a pour but de retrouver les amas d’étoiles dissociés en se basant uniquement sur leur composition chimique. Nous évaluons la viabilité de cette technique pour retrouver les étoiles qui sont nées dans un même amas mais qui ne sont plus gravitationnellement liées.Méthodes. Nous avons créé une librairie de spectres stellaires de haute qualité afin de faciliter l’évaluation des analyses spectrales. Nous avons développé notre propre outil d’analyse spectrale, nommée iSpec, capable d’homogénéiser les spectres stellaires venant de tous types d’instruments et de dériver les paramètres atmosphériques et les abondances chimiques. Finalement, nous avons compilé des spectres stellaires d’étoiles de 32 amas ouverts, nous avons dérivé de façon homogène les paramètres atmosphériques et les abondances de 17 espèces, et nous avons utilisé des algorithmes d’apprentissage automatique pour grouper les étoiles en se basant sur leur composition chimique.Résultats. Nous avons trouvé que les étoiles à des étapes d’évolution différentes ont des motifs chimiques distincts qui peuvent être dus à des effets NLTE,de diffusion atomique, de mélange et de corrélation à partir des déterminations de paramètres atmosphériques. Quand nous séparons les étoiles suivant leur stade d’évolution, nous observons qu’il y a un important degré de recouvrement dans la détermination des signatures chimiques des amas ouverts. Ceci rend difficile de retrouver les groupes d’étoiles nées ensemble en utilisant la technique de marquage chimique. / Context. Stars are born together from giant molecular clouds and, if weassume that they were chemically homogeneous and well-mixed, we expect them toshare the same chemical composition.Most of the stellar aggregates are disrupted while orbiting the Galaxy and thedynamic information is lost, thus the only possibility to reconstruct the stellarformation history is to analyze the chemical abundances that we observe today.Aims. The chemical tagging technique aims to recover disrupted stellarclusters based merely on their chemical composition. We evaluate the viability of thistechnique to recover conatal stars that are not gravitationally bound anymore.Methods. We built a high-quality stellar spectra library to facilitate theassessment of spectral analyses. We developed our own spectral analysisframework, named iSpec, capable of homogeneizing stellar spectra and derivingatmospheric parameters/chemical abundances. Finally, we compiled stellar spectrafrom 32 Open Clusters, homogeneously derived atmospheric parameters and 17abundance species, and applied machine learning algorithms to group the starsbased on their chemical composition. This approach allows us to evaluate theviability of the chemical tagging technique.Results. We found that stars in different evolutionary stages havedistinguished chemical patterns may be due to NLTE effects, atomic diffusion, mixingand correlations from atmospheric parameter determinations. When separating starsper evolutionary stage, we observed a high degree of overlapping among OpenCluster’s chemical signatures, making it difficult to recover conatal aggregates byapplying the chemical tagging technique.

The present thesis was born and evolved within the RAdial Velocity Experiment (RAVE) with the goal of measuring chemical abundances from the RAVE spectra and exploit them to investigate the chemical gradients along the plane of the Galaxy to provide constraints on possible Galactic formation scenarios. RAVE is a large spectroscopic survey which aims to observe spectroscopically ~10^6 stars by the end of 2012 and measures their radial velocities, atmospheric parameters and chemical abundances. The project makes use of the UK Schmidt telescope at Australian Astronomical Observatory (AAO) in Siding Spring, Australia, equipped with the multiobject spectrograph 6dF. To date, RAVE collected and measured more than 450,000 spectra. The precision of the chemical abundance estimations depends on the reliability of the atomic and atmosphere parameters adopted (in particular the oscillator strengths of the absorption lines and the effective temperature, gravity, and metallicity of the stars measured). Therefore we first identified 604 absorption lines in the RAVE wavelength range and refined their oscillator strengths with an inverse spectral analysis. Then, we improved the RAVE stellar parameters by modifying the RAVE pipeline and the spectral library the pipeline rely on. The modifications removed some systematic errors in stellar parameters discovered during this work. To obtain chemical abundances, we developed two different processing pipelines. Both of them perform chemical abundances measurements by assuming stellar atmospheres in Local Thermodynamic Equilibrium (LTE). The first one determines elements abundances from equivalent widths of absorption lines. Since this pipeline showed poor sensibility on abundances relative to iron, it has been superseded. The second one exploits the chi^2 minimization technique between observed and model spectra. Thanks to its precision, it has been adopted for the creation of the RAVE chemical catalogue. This pipeline provides abundances with uncertains of about ~0.2dex for spectra with signal-to-noise ratio S/N>40 and ~0.3dex for spectra with 20>S/N>40. For this work, the pipeline measured chemical abundances up to 7 elements for 217,358 RAVE stars. With these data we investigated the chemical gradients along the Galactic radius of the Milky Way. We found that stars with low vertical velocities |W| (which stay close to the Galactic plane) show an iron abundance gradient in agreement with previous works (~-0.07$ dex kpc^-1) whereas stars with larger |W| which are able to reach larger heights above the Galactic plane, show progressively flatter gradients. The gradients of the other elements follow the same trend. This suggests that an efficient radial mixing acts in the Galaxy or that the thick disk formed from homogeneous interstellar matter. In particular, we found hundreds of stars which can be kinetically classified as thick disk stars exhibiting a chemical composition typical of the thin disk. A few stars of this kind have already been detected by other authors, and their origin is still not clear. One possibility is that they are thin disk stars kinematically heated, and then underwent an efficient radial mixing process which blurred (and so flattened) the gradient. Alternatively they may be a transition population" which represents an evolutionary bridge between thin and thick disk. Our analysis shows that the two explanations are not mutually exclusive. Future follow-up high resolution spectroscopic observations will clarify their role in the Galactic disk evolution. / Die vorliegende Doktorarbeit wurde im Rahmen des RAdial Velocity Experiment (RAVE) angefertigt. Ihr Ziel ist es, chemische Elementhäufigkeiten an RAVE-Spektren zu messen und zur Untersuchung chemischer Gradienten in der Milchstrassenebene zu benutzen, um verschieden Szenarien der Galaxienentstehung einzugrenzen. RAVE ist eine große spektrokopische Durchmusterung, deren Ziel es ist, bis zum Ende des Jahres 2012 insgesamt 10^6 Sterne zu spektroskopieren, um deren Radialgeschwindigkeiten, sternatmosphärische Parameter und chemische Häufigkeiten zu messen. Das Projekt benutzt das UK Schmidt Teleskop am Australian Astronomical Observatory (AAO) in Siding Spring, Australien, welches mit dem Multiobjekt-Spektrographen 6dF bestückt ist. Bis heute hat RAVE die Spektren von mehr als 450,000 Sternen gesammelt und untersucht. Die Genauigkeit, mit der die Elementhäufigkeiten abgeschätzt werden können, hängt von der Zuverlässigkeit der verwendeten Parameter, (insbesondere der Oszillatorstärken der Absorptionslinien sowie von der effektiven Temperatur, Schwerebeschleunigung und der Metallizität des gemessenen Sterns) ab. Daher identifizierten wir zunächst 604 Absorptionslinien im Wellenlängenbereich von RAVE und verbesserten deren Oszillatorstärken durch eine inverse Spektralanalyse. Dann wurden die stellaren Parameter von RAVE verbessert, indem die RAVE Pipeline und die stellaren Parameter, auf denen sie beruht, modifiziert wurden. Die Änderungen eliminierten einen Teil der systematischen Fehler von stellaren Parametern, die im Laufe dieser Arbeit gefunden wurden. Um Elementhäufigkeiten zu bestimmen, haben wir zwei verschiedene Prozessierungs-Pipelines entwickelt. Beide berechnen die Elementhäufigkeiten unter der Annahme von Sternatmosphären im lokalen thermischen Gleichgewicht (local thermal equilibrium, LTE). Die erste Pipeline berechnete Elemenhäufigkeiten anhand der Äquivalentbreiten von Absorptionslinien. Da diese Methode eine geringe Empfindlichkeit für die Elementhäufigeiten relativ zu Eisen erreichte, wurde sie ersetzt. Die neue Pipeline benutzt chi^2-Fits von Modellspektren an die beobachteten Spektren. Dank Ihrer Präzision wurde diese für die Erstellung des RAVE-Katalogs von Elementhäufigkeiten verwendet. Diese Pipeline liefert Elementhäufigkeiten mit einer Genauigkeit von ~0.2dex, während für Spektren mit 20>S/N>40 immerhin noch ~0.3dex Genauigkeit erreicht werden. Für die vorliegende Arbeit wurden für 217.358 Sterne die Häufigkeiten von sieben chemischen Elementen bestimmt. Mit diesen Daten wurde der radiale chemische Gradient unserer Milchstraße untersucht. Wir finden, dass Sterne mit kleinen vertikalen Geschwindigkeiten |W|, die also nahe der galaktischen Ebene bleiben, einen radialen Gradienten der Eisenhäufigkeit zeigen, der mit früheren Studien übereinstimmt (~-0.07 dex Kpc^-1), während Sterne mit großen |W|, also solche, die größere galaktische Höhen erreichen, einen progressiv flachere Gradienten zeigen. Die Gradienten der anderen Element folgen dem gleichen Trend. Das lässt darauf schließen, dass entweder die Durchmischung der galaktischen dicken Scheibe effizient arbeitet oder aber dass die dicke Scheibe aus interstellarer Materie gebildet wurde, die chemisch recht homogen war. Speziell fanden wir hunderte von Sternen, die zwar kinematisch als zur dicken Scheibe zugehörig klassifiziert werden können, die aber die typische chemische Zusammensetzung der dünnen Scheibe aufweisen. Einige wenige dieser Sterne wurden bereits von anderen Autoren entdeckt, aber ihre Herkunft bleibt immer noch unklar. Eine Möglichkeit ist, dass die Sterne der dünnen Scheibe kinematische geheizt werden, sodass sie effizienter radial gemischt werden, was die chemischen Gradienten verwischt und auch flacher macht. Alternativ dazu könnten diese Sterne einer "Übergangspopulation" angehören, welche hinsichtlich der Scheibenevolution die Verbindung zwischen der dünnen und der dicken Scheibe darstellt. Unsere Untersuchung zeigt, dass sich diese beiden Erklärungen gegenseitig nicht ausschließen. Künftige Nachspektroskopierung mit hoher Auflösung wird die Rolle dieser Sterne in der Entwicklungsgeschichte der galaktischen Scheibe aufklären.

Damped Lyman α systems (DLAs) are useful probes of the chemical enrichment of the universe as they provide accurate abundance measurements of many chemical species. Using a sample of 30 DLAs (with large metal column densities) observed with the High Resolution Echelle Spectrometer on the Keck I telescope, the abundances of several elements (i.e. iron, zinc, chromium, silicon, sulphur, phosphorus, manganese, and boron) are derived and presented. A comparison is drawn between the abundances from these metal-rich DLAs with literature samples encompassing the largest compilation of high resolution observations of other DLAs, and stars from the Milky Way and its satellite galaxies to understand the astrophysical nature of DLAs. Furthermore, the first ever extragalactic study of boron is presented. Using the sample of 30 metal-rich DLAs, two 3σ detections and one near detection 2.97σ) were found. From the comparison of [B/O] and, for the first time, [B/S], with studies in the Milky Way, there appears to be an excess of boron relative to its parent nucleus (oxygen) in these three DLA systems, suggesting that there may be a higher cosmic ray flux in DLAs than in the Milky Way. / Graduate / 0606

Damped Lyman α systems (DLAs) are useful probes of the chemical enrichment of the universe as they provide accurate abundance measurements of many chemical species. Using a sample of 30 DLAs (with large metal column densities) observed with the High Resolution Echelle Spectrometer on the Keck I telescope, the abundances of several elements (i.e. iron, zinc, chromium, silicon, sulphur, phosphorus, manganese, and boron) are derived and presented. A comparison is drawn between the abundances from these metal-rich DLAs with literature samples encompassing the largest compilation of high resolution observations of other DLAs, and stars from the Milky Way and its satellite galaxies to understand the astrophysical nature of DLAs. Furthermore, the first ever extragalactic study of boron is presented. Using the sample of 30 metal-rich DLAs, two 3σ detections and one near detection 2.97σ) were found. From the comparison of [B/O] and, for the first time, [B/S], with studies in the Milky Way, there appears to be an excess of boron relative to its parent nucleus (oxygen) in these three DLA systems, suggesting that there may be a higher cosmic ray flux in DLAs than in the Milky Way. / Graduate / 0606

O presente trabalho propõe uma abordagem abrangente para descrever a evolução da região central da Via Láctea, compreendendo-se aí o bojo, a barra e as interfaces dos mesmos com o limite interno do disco e com a região central do halo. Pretende-se investigar as propriedades químicas e cinemáticas destas estruturas, que são interconectadas, com o objetivo de separá-las e aplicar os resultados daí obtidos a um modelo de formação e evolução do bojo e da região interna do disco que descreva simultaneamente distintos aspectos da evolução da região central da Galáxia. Na primeira parte do trabalho, uma amostra de nebulosas planetárias (NPs) localizadas no disco interno e no bojo da Galáxia é utilizada para encontrar a distância galactocêntrica que melhor separa estas duas populações, do ponto de vista das abundâncias. Foram utilizadas escalas de distâncias estatísticas para o estudo da distribuição das abundâncias na interface bojo-disco. A aplicação do teste Kolmogorov-Smirnov mostrou que, em média, a população interna não segue o gradiente radial de abundâncias do disco na direção do centro galáctico. Baseado neste estudo, propõe-se uma distância galactocêntrica de 1.5 kpc para definir a interface bojo-disco. Na segunda parte do trabalho, foram realizadas observações espectrofotométricas de 21 NPs localizadas na direção do centro da Galáxia com o telescópio SOAR. Estes objetos estão localizados bem próximos ao plano galáctico na direção central da Via Láctea, onde não existem dados de NPs na literatura. Os resultados mostram que as NPs localizadas nesta região apresentam baixas abundâncias de oxigênio comparadas com as NPs do disco interno e de outras regiões do bojo. Os resultados indicam que o bojo apresenta uma complexa composição de populações estelares. Por um lado, a presença de nebulosas com baixas abundâncias mostra que o bojo pode ter se formado a partir de um disco galáctico antigo através de uma evolução secular. Por outro lado, existem alguns objetos do bojo para os quais as abundâncias coincidem com o limite do gradiente radial do disco nesta região. Esta é uma evidência para um bojo composto por duas ou mais populações: uma originada do disco fino, e outra originada do disco espesso. Na última parte do trabalho propõe-se a inclusão de fluxos radiais de gás em um modelo de evolução química para simular os efeitos de uma barra localizada no centro da Galáxia nas distribuições de abundâncias, densidade de gás e taxa de formação estelar (SFR). Os resultados das simulações indicam que os modelos com fluxos de gás apresentam uma SFR mais alta no bojo e que os perfis da SFR e da densidade de gás na região central são melhor reproduzidos após a inclusão dos fluxos radiais no modelo. As simulações indicam ainda que o gradiente de abundâncias do disco é mais plano para o caso da inclusão da barra. Estes resultados indicam que a barra e os fluxos de gás exercem um importante papel na formação de estrelas no centro das galáxias espirais barradas. / This project proposes a comprehensive approach to describe the evolution of the central region of the Galaxy, comprising the bulge, the bar and their interfaces with the inner disk and the central region of the halo. We intend to investigate the chemical and kinematic properties of these structures, which are interconnected, aiming to separate them and apply these results to a model for the formation and evolution of the bulge and inner disk, capable to describe simultaneously distinct aspects of the evolution of the central region of the Galaxy. First, a sample of planetary nebulae (PNe) located in the inner-disk and bulge of the Galaxy is used in order to find the galactocentric distance that better separates these two populations, from the point of view of abundances. Statistical distance scales were used to study the distribution of abundances across the disk-bulge interface. A Kolmogorov-Smirnov test was used to find the distance in which the chemical properties of these regions better separates. The results of the statistical analysis indicate that, on the average, the inner population has lower abundances than the outer. Additionally, for the $\\alpha$-elements abundances, the inner population does not follow the disk radial gradient towards the galactic centre. Based on our results, we suggest a bulge-disk interface at 1.5 kpc, marking the transition between the bulge and inner-disk of the Galaxy, as defined by the intermediate mass population. Second, we present spectrophotometric observations for a sample of 21 PNe located towards the galactic centre of the Galaxy. The abundances are derived based on observations in the optical domain made at the SOAR telescope. Their location is interesting since there are no observations of PNe in this region. The data show lower oxygen abundances compared to those from PNe located in the inner disk and other bulge regions. The results show that the bulge has a complex composition of stellar populations. The presence of PNe with low abundances indicates that the bulge might be formed from an old galactic disk through secular evolution. On the other hand, other objects from our sample have abundances compared to those from inner disk PNe. This is evidence that two or more populations might compose the bulge: one originated from the thin disk, and the other from the thick disk. Last, we propose a chemical evolution model that includes radial gas flows. This is done in order to mimic the effects of the galactic bar on the chemical abundances distributions and the gas density profiles and the star formation rate (SFR). The results of the models with radial flows point to a high SFR in the bulge and, additionally, the SFR and gas density profiles in the inner Galaxy are better reproduced after the inclusion of radial gas flows in the model. After including a specific velocity pattern for the bar, the results show a flattening of the radial abundance gradient. Our results indicate that radial gas flows may play an important role in the star formation near the centre of barred spiral galaxies.

<p>Chemical evolution addresses the problem of the formation of the chemical elements and their evolution throughout the history of the universe. This thesis discusses in particular the chemical evolution in the young universe and what we may learn from the observations of the oldest stars. The present day production of carbon in the Galaxy is also discussed. Interstellar media of young, metal-poor, star-forming systems are expected to show large chemical abundance inhomogeneities due to local supernova explosions. These inhomogeneities are reflected in the surface abundances of the population of longlived, low-mass stars. A stochastic model of the chemical evolution in such systems is presented and used to study the metallicity distribution and the scatter in chemical abundance ratios. The model takes into account mixing of the enriched material by turbulent motions and cloud collisions in the interstellar medium as well as infall of pristine matter. The predicted metallicity distribution shows, in accordance with observations of extreme Pop II strars in the Galactic halo, a distinct cut-off at [Fe/H]~-4. However, the fraction of stars below [Fe/H]=-4 agrees with observatrion only if a population of metal-free stars (Pop III) was never able to form. The predicted scatter in abundance ratios is demonstrated to be crucially dependent on the as yet uncertain supernova yields and the relatively small star-to-star scatter is tentatively explained by the averaging of a large number of contributing supernovae and by the selection effects favouring contributions from supernovae in a certain mass range for the most metal-poor stars. Furthermore, stars enriched by one single supernova are predicted to be found in very narrow sequences in the abundance ratio diagrams (so called A/A diagrams). Verification of the existence of such features, called single supernova sequences, is observationally challenging. Abundance analysis of carbon was performed in a large sample of solar-type stars in the Galactic disk using the forbidden [C I] line at 8727 Å. A comparison between the relation of [C/O] with metallicity for the Galactic stars and that of dwarf irregular galaxies suggests that large amounts of carbon are produced today by massive, so called Wolf-Rayet stars. Low-mass stars are less important. This was also demonstrated by modelling the chemical evolution of carbon.</p>

Chemical evolution addresses the problem of the formation of the chemical elements and their evolution throughout the history of the universe. This thesis discusses in particular the chemical evolution in the young universe and what we may learn from the observations of the oldest stars. The present day production of carbon in the Galaxy is also discussed. Interstellar media of young, metal-poor, star-forming systems are expected to show large chemical abundance inhomogeneities due to local supernova explosions. These inhomogeneities are reflected in the surface abundances of the population of longlived, low-mass stars. A stochastic model of the chemical evolution in such systems is presented and used to study the metallicity distribution and the scatter in chemical abundance ratios. The model takes into account mixing of the enriched material by turbulent motions and cloud collisions in the interstellar medium as well as infall of pristine matter. The predicted metallicity distribution shows, in accordance with observations of extreme Pop II strars in the Galactic halo, a distinct cut-off at [Fe/H]~-4. However, the fraction of stars below [Fe/H]=-4 agrees with observatrion only if a population of metal-free stars (Pop III) was never able to form. The predicted scatter in abundance ratios is demonstrated to be crucially dependent on the as yet uncertain supernova yields and the relatively small star-to-star scatter is tentatively explained by the averaging of a large number of contributing supernovae and by the selection effects favouring contributions from supernovae in a certain mass range for the most metal-poor stars. Furthermore, stars enriched by one single supernova are predicted to be found in very narrow sequences in the abundance ratio diagrams (so called A/A diagrams). Verification of the existence of such features, called single supernova sequences, is observationally challenging. Abundance analysis of carbon was performed in a large sample of solar-type stars in the Galactic disk using the forbidden [C I] line at 8727 Å. A comparison between the relation of [C/O] with metallicity for the Galactic stars and that of dwarf irregular galaxies suggests that large amounts of carbon are produced today by massive, so called Wolf-Rayet stars. Low-mass stars are less important. This was also demonstrated by modelling the chemical evolution of carbon.