Dark Substructures and Model Complexity in Strong Gravitational Lensing

Abstract

This work investigates our ability to use strong gravitational lensing as a tool to constrain the nature of dark matter. The assumption is that if we can detect dark subhalos below some cut-off mass in the subhalo mass function, then we can use these detections to place limits on the mass of a dark matter particle. In this Thesis, I first investigate two strong lenses imaged with the relatively new James Webb Space Telescope (JWST), attempting to model the mass distribution of the main deflector including different `multipole' perturbations to induce angular complexity and also including subhalos in the model. The results for multipole-only additions show strong bayesian evidence for angular mass complexity in on lens - SPT2147, with multipole strengths of 0.3-1.7 for and 2.4-9.5 for , whilst the other lens - SPT0418 shows no such preference. For lens models with only substructures, I find a strong preference for a substructure in SPT2147-50 with a Bayes factor of , however the addition of multipoles as well as substructure reduces the substructure bayesian evidence. With no reasonable prior information that can be placed upon what we expect of multipole strengths, I treat all model structures (multipole-only, substructure-only and multipole+substructure) as equally likely. I find that we return an increase in log bayesian evidence of (still corresponding to a detection) for a subhalo detection with a mass of (for a Navarro-Frenk-White model) in a model that also has orders and multipoles. I conclude that while SPT2147-50 could represent a new detection of a dark matter substructure in a strong lens, further analysis is needed to confirm that the signal is not due to systematics associated with the lens mass model.

In the second section, I analyse a selection of lenses that have Hubble Space Telescope imaging (from the SLACS and BELLS-GALLERY samples) and attempt to determine mass-to-light agreement, relationships of model complexity to the external shear, and the effects of multipole perturbations. Comparing a simple model that fits elliptical isophotes of the lens light, where the ellipses have radially constant parameters, I find that there is good agreement between the radially-constant light model, and the commonly used elliptical power-law (EPL) mass distribution (which has parameters that are also radially constant). Comparing this to a model where the angles and ellipticity of the isophotes can vary, I find that the model parameters of these isophotes around the einstein radius match the parameters of the constant model fairly well, demonstrating that for these lenses, it is unlikely that we need to change the EPL for a model with more angular freedom. For the external shear, I find strong evidence that shears that would be considered anomalously high strength () are correlated to the angular variance with the variable-ellipse model, with and for the p-value and linear relationship coefficient.

I also compare the mass and light multipole perturbations of these same lens galaxies, and find that these lens samples are not suitable for an in-depth study of multipoles as the GALLERY lenses are too faint and the SLACS lenses are too smooth. I find no correlation between a change in external shear magnitude and the multipole strength when going from the EPL model to EPLMultipoles, however there is a correlation between the multipole strength and the magnitude of external shear itself (, ). I find that there are generally correlations between and multipole strengths, and that mass and light are in general agreement to 3-sigma regarding strength, but exhibit less agreement on strength. I also discuss possible origins of mass-light mismatches from high central dark matter fractions and dark halos not being able to match the morphology of strong stellar disks.

Overall, this Thesis contributes to the growing body of work regarding galaxy-scale strong gravitational lensing in three ways. Firstly, it provides a detailed investigation into lenses from one of the newest sources (JWST) and provides an opportunity to discuss the possible degeneracies when trying to detect dark substructures. Second, the Thesis contributes to the understanding of what drives anomalously high external shear magnitudes. Third, it addresses the origins and meaning of different aspects of lens model complexity and considers whether the models we are using are accurate enough, which is essential for the new era of high-quality lenses from JWST and high-volume samples from Euclid.