Stellar population synthesis models

Development of stellar population models

Jesús Falcón-Barroso, Benjamin Roeck, Alexandre Vazdekis

A significant fraction of the observing time in large telescopes is dedicated to understanding the stellar content of galaxies at high and low redshift. In fact in the last decade such studies have provided severe restrictions to our current view of how galaxies form and evolve. The transformation from the observed light to meaningful physical parameters requires the comparison with predictions of stellar population synthesis models, which are fed by stellar spectral libraries.

Our group is one of the pioneers producing stellar population synthesis models in optical wavelengths for the study of galaxies. The stellar library MILES, containing 985 stars, is regarded nowadays as the standard library in the field as its spectral resolution, spectral-type coverage, flux-calibration accuracy and number of stars represent a substantial improvement over previous libraries. Based on the MILES spectra, we published a new generation of stellar population models that extend our optical SEDs from intermediate to very old age regimes and the metallicity coverage from super-solar down to [M/H] = -2.3. The library and models are available to the public through a user-friendly webpage (

We are currently working on the extension of our model predictions from the UV to NIR wavelengths, as well as producing stellar population models for different [Mg/Fe] abundance ratios (critical to properly study the populations of the more massive galaxies).

Analysis methods for the study of populations in galaxies

Ignacio Martín-Navarro, Alexandre Vazdekis

The study of the integrated light suffers several fundamental degeneracies that prevent us to obtain unique solutions. The most well known is the age-metallicity degeneracy. Due to this degeneracy a galaxy could look redder either because is older or more metal rich. Moreover, it is difficult to know the effects of the initial mass function (IMF) which is responsible for fixing the ratio between dwarf and giant stars. Another problem is our inability to disentangle the burst-age from the burst-strength. In the last decade new analysis tools based on some absorption lines have been able to partially mitigate the age-metallicity degeneracy. However, even the commonly used Balmer Hβ line index shows a non negligible sensitivity to metallicity, as well as to nebular emission.

We are searching for key spectral indices that allow us to unambiguously obtain information about age, metallicity, relative abundances ratios, IMF and formation histories. With such system of indices it is possible to propose diagnostic diagrams to break the fundamental degeneracies affecting the integrated light of the stellar populations. This work is specifically aimed to exploit the capabilities of the OSIRIS-GTC tunable filters.

Universally invariant IMF vs IGIMF theory

Alexandre Vazdekis, Carsten Weidner

Over the past years observations of young and populous star clusters have shown that the stellar initial mass function (IMF) can be conveniently described by a two-part power-law with an exponent α2 = 2.3 for stars more massive than about 0.5 M and an exponent of α1 = 1.3 for less massive stars. A consensus has also emerged that most, if not all, stars form in stellar groups and star clusters, and that the mass function of these can be described as a power-law (the embedded cluster mass function, ECMF) with an exponent β ≈ 2. These two results imply that the integrated galactic IMF (IGIMF) for early-type stars cannot be a Salpeter power-law, but that they must have a steeper exponent. An application to star-burst galaxies shows that the IGIMF can become top-heavy. This has important consequences for the distribution of stellar remnants and for the chemo-dynamical and photometric evolution of galaxies.

We are currently investigating which observational quantities can be best used to distinguish between the IGIMF-theory and an universal invariant IMF. To achieve this goal, we are computing a large set of models assuming standard IMFs and the IGIMF to be compared with the vast amount of existing observational data.