Synthesis of Stellar Populations in Galaxies
Jean Michel Gomes
Spectral synthesis of stellar populations is a very powerful tool to investigate the formation and evolution of galaxies. This course is intended to discuss the general aspects of this method from its infancy to present days. It's aimed at graduate students and researchers willing to know more about the subject.
The classes will be divided according to the following schedule:
- Feb 23: Synthesis of Stellar Populations: historical approach
- March 2: Synthesis of Stellar Populations: Simple Stellar Populations as building block of galaxies
- March 9: Synthesis of Stellar Populations: the STARLIGHT code
- March 16: Synthesis of Stellar Populations: Some Applications
Session 1: Synthesis of Stellar Populations: historical approach
23 February 2011
One of the key steps in extragalactic astronomy is to understand the formation and evolution of galaxies. By means of spectral synthesis, which is the comparison of models to data, astronomers have achieved an unprecedented development in the last 10 years to derive the star-formation and chemical histories of galaxies with successful applications to huge sets of galaxy spectra, such as the Two-Degree Field Galaxy Redshift Survey, Six-Degree Field Galaxy Survey and the Sloan Digital Sky Survey.
Therefore, it's worthwhile to tell the synthesis' history from its infancy to the present days, in order to have a broader image of what were, what is and the distinct methods of synthesis, as well as old problems and plagues that still disturb scientists of the field today. Nevertheless, a complete revision with all possible references would be impossible, so I will present my point of view of the principal achievements in this field. This is the objective of the first class.
We will start with a qualitative introduction of the synthesis' problem. This technique comes in two scopes, evolutionary and semi-empirical population synthesis. The latter will be the main focus of our study, since the first works of Tinsley, Faber and O'Connell in the 60's up to now. We will also discuss the different observables used in the models and the degeneracies involved.
Session 2: Synthesis of Stellar Populations: Simple Stellar Populations as building block of galaxies
2 March 2011
We saw in the first class that semi-empirical population synthesis (sometimes called as an inverse approach or only population synthesis) explores the question of galaxy evolution from a more practical viewpoint. Its main goal is to decompose the spectrum of a galaxy into the spectra of its main constituent stellar populations. We can select a library of stellar spectra or, since more recently, Single Stellar Populations (SSPs), i.e. stellar populations forming instantaneously and being fully characterised by their age, chemical composition and Initial Mass Function (IMF). The derived distribution of these "SSP building blocks", referred to as "population vector", corresponds to a discretized approximation to the Star Formation History of a galaxy.
We will study the basic equations and ingredients to create Composite Stellar Populations. A simple approximation is then used for an instantaneous burst to create Simple Stellar Populations (SSP). The best SSP models available in the market will be discussed, together with problems and the expectations for the future.
Session 3: Synthesis of Stellar Populations: the STARLIGHT code
9 March 2011
The power of spectral synthesis as a means to estimate physical properties of galaxies has been well established. We have entered in a new era with the availability of large and high-quality data bases of observed galactic spectra. The combination of a wide set of synthetic and observed stellar libraries with meanwhile much refined population synthesis codes permit us to significantly improve our understanding on the formation and evolution of galaxies. The STARLIGHT (Cid Fernandes et al. 2005), provides one of the most powerful tools currently available.
In the current version of STARLIGHT, the Simple Stellar Populations (SSPs) library may contain up to 300 elements, sufficient to densely cover both the age and metallicity parameter space. The SSP population vector returned by the model holds essentially all information needed to fully characterize the SFH of a galaxy, at least in theory.
We describe the mathematical model in which STARLIGHT is based. We explain the probabilities of the parameters using Bayes' theorem and statistical mechanics. The numerical procedures will be outlined. They are based on Markov Chain Monte Carlo chains with simulated annealing and Gelman and Rubin convergence method.
Session 4: Synthesis of Stellar Populations: Some Applications
16 March 2011
We show some applications of the STARLIGHT synthesis method to SDSS and 6dF galaxies. Emission lines are also studied, with and without subtracting the stellar contributions to show how the classical BPT diagram to classify galaxies according to star-forming and AGN changes. The principal by-product of the synthesis, the population vector, will be discussed for different types of galaxies. Several other physical properties, such as the stellar mass, mean stellar age and metallicity will also be discussed. At last, we will show an special application for possible miss-classified AGN, called retired galaxies, using the methods developed in this course.