Quasars are considered to be the most luminous subclass among Active Galactic Nuclei. Although they appear to be distributed everywhere in the Universe, only a small fraction of them happen to be part of a fortuitous alignment with a galaxy or a galaxy cluster that lies along the line-of-sight, which results in multiple images of the background quasar by means of gravitational lensing. Compact objects, such as stars in the bulges or halos of the galaxies, may induce further changes in the mean lensing magnification that are uncorrelated among images. This microlensing effect due to unseen star-sized objects can be used to extract information about both the source quasar and the lens galaxy.
The small size of the emitting quasar regions, together with involved distances in the Gpc range, make spatially resolved observations of the background quasar hopeless, even with the largest available optical telescopes. Direct observation of the microlenses at such distances is also impossible. Microlensing can help overcoming these difficulties, since microlensing magnification is very sensitive to the degree of alignment of the system (thus, relative movement of the microlenses with respect to the background quasar causes measurable changes in brightness over finite periods of time). Additionally, flux measures can be done across different constituents of the spectra that arise from regions with different sizes.
In fact, nearly all we know from quasars has been inferred by studying their extraordinary rich spectra. There is a time-variable continuum emission spanning several orders of magnitude in wavelength, whose generating mechanism can be explained by a dissipative accretion disk around a very massive object. Broad emission lines follow the continuum variability pattern after different time delays, that have been used to impose upper limits on the distance to the central engine. This technique, known as Reverberation Mapping can be applied to any of the known quasars as long as there is a high enough S/N ratio, but observation campaigns over long periods of time are required.
Microlensing is based upon a different approach that can help overcoming the difficulties of Reverberation Mapping, though it faces another challenging circumstances. Quasars undergoing strong lensing are only a small and rare subset. Moreover, the fraction of them that are brilliant enough as to allow detailed spectra is small. In addition to that, the observations are technically challenging, since they must consist on two independent, simultaneous long slit spectra with no contamination from each other, being the images only a few arc seconds apart. This has only been achieved for roughly a couple dozens of these objects. The authors did in each case their best to derive properties based on the detected microlensing signal, although the involved physical quantities are strongly unconstrained when dealing with a single object.
That is the starting point of this thesis, in which we bring together as many as possible of such spectra from the literature, with the goal of studying properties that could not be studied individually, of both the source quasar and the lens galaxy. The work requires collecting the data, designing the appropriate measuring process, generating sets of simulations to compare the measures with, applying the Bayesian inference methods, and offering an interpretation of the results.
The thesis is structured in seven chapters:
Chapter 1 presents a theoretical introduction gathering the essentials required to work on extragalactic microlensing.
Chapter 2 presents a short motivation statement, explaining the need for this thesis work and why it may be useful for the scientific community. The main motivation: processing a sample of lensed quasar spectra and build a single-epoch statistical study out of them, that overcomes the limitations of the individual object approach and the lack of observation campaigns over long periods of time.
In Chapter 3 I present a consistent overview of all the spectra, brought together over a common wavelength scale, after their continua have been subtracted, and scaled so that they roughly overlap. This is intended to be a quick visual reference to the information and possibilities contained in each spectral pair.
Chapters 4 to 7 present the four research papers that constitute the bulk of this thesis.
Chapter 8 summarizes the main conclusions of the thesis and offers some suggestions for future work in the field.