Research

Gravitational Waves

Our group uses interferometer-based gravitational-wave experiments to constrain alternative models of the dark sector of the universe.

Dark Matter

Evidence for the existence of dark matter (DM) rests on its gravitational effects. Efforts to understand its nature have motivated an interpretation within particle physics that requires new dark particles beyond those predicted by the Standard Model of particle physics. Despite decades of experimental progress, the nature of DM remains a mystery. Even its most fundamental properties, such as its mass and the strength of its interactions with Standard Model particles and with other dark particles, remain unknown.

Our group studies the predictions of different dark-sector models for processes relevant to the history of the universe—such as the generation of relics during inflation and the production of dark matter—as well as for the evolution of large-scale cosmological perturbations and various astrophysical observables, including effects on pulsar timing measurements and gravitational-wave experiments. These predictions are compared with observational and experimental results.

Early Universe

In studying the origin and early evolution of cosmological perturbations, we investigate the robustness of the theoretical predictions of inflationary models, as well as different mechanisms for producing dark matter and other relics. This line of research includes the development of systematic methods for calculating correlation functions of quantum fields and other quantities relevant to the early universe.

Artificial Intelligence and Cosmology

Recent years have seen significant progress in applying machine learning methods to astrophysics and cosmology. Among these methods are physics-informed neural networks (PINNs), which solve systems of differential equations without relying on traditional numerical methods. A major advantage of this approach is a substantial reduction in computation times for parameter inference when solving the differential system is computationally demanding, as is particularly the case for cosmological models beyond the standard model. However, this method is still under development and does not provide estimates of its own error for every type of differential equation, making this an area in which we are actively working. We are also extending the application of PINNs to the study of structure growth in the universe.

We use PINNs to obtain the dynamics of the universe in different cosmological models as an alternative to traditional numerical methods. We also use neural-network regression methods to extract theoretical predictions of quantities relevant to cosmology from numerical codes.

Galaxy Formation and Evolution

Galaxies form as gas cools and collapses at the centre of dark matter halos, where star formation begins. Within the concordance cosmological model, dark matter halos and their galaxies grow and evolve hierarchically, strongly shaped by collisions and the continuous accretion of mass. That evolution is highly non-linear, lacks symmetries and spans many scales; in this context numerical simulations have become the tool of choice for studies of galaxy formation, following the evolution of each component through cosmic time.

We study in particular the formation and evolution of galactic discs in galaxies of mass and structure similar to our own Milky Way, their stability, and their connection to stellar components such as bulges and bars. We also use constrained cosmological simulations of the Local Group to investigate how environment and the history of structure formation influence the evolution of the Milky Way, Andromeda and their satellite galaxies. Complementing this, we study the chemical evolution of galaxies and its relation to star formation and galactic assembly.

We also work on code development, incorporating sub-grid physical models into the simulations. Among our recent developments is a model describing molecular hydrogen and its relation to star formation, which allows the connection between molecular gas, the birth of stars and the structure of galaxies to be studied in greater detail.

Pulsars

Pulsars—neutron stars with extraordinarily stable rotation periods—are among the most precise laboratories for studying gravitational physics in the strong-field regime. Our group seeks to use pulsar observations to constrain alternative models of dark matter and gravity. To this end, we use and develop statistical approaches and techniques to extract constraints on these models from the data.

Modified Gravity

The discovery of the accelerated expansion of the universe using Type Ia supernovae in 1998 raised a problem for theoretical cosmology: explaining the physical mechanism responsible for it. The standard cosmological model (ΛCDM) proposes including the cosmological constant in Einstein's equations, but this presents some theoretical problems. This issue, together with tensions in cosmological parameter values, motivates the study of alternative cosmological models. Among the most widely studied alternatives, we can distinguish three categories: (i) parametrised dark energy, which adds a fluid with a particular equation of state to the energy–momentum tensor; (ii) adding one or more scalar fields with a specified potential to the energy–momentum tensor; and (iii) describing gravity using a theory other than general relativity. This last option is known as modified gravity and is also used to describe phenomena attributed to dark matter.

We study modified-gravity models to explain the current accelerated expansion of the universe. We also test local constraints on these theories to assess their consistency. Models studied include the f(R) models proposed by Hu–Sawicki and Starobinsky, unimodular gravity, and teleparallel gravity with non-minimal coupling.

We study modified-gravity models as alternatives to describing the universe in terms of a dark sector, including models of spin-2 dark matter, with an emphasis on their theoretical consistency and phenomenological implications.

Data Analysis and Inference

We use data from Type Ia supernovae, cosmic chronometers, baryon acoustic oscillations, the cosmic microwave background, quasar ultraviolet and X-ray fluxes, and the ages of the oldest globular clusters to test predictions of cosmological models beyond the standard model and constrain their free parameters.

The significant increase in the quantity and quality of data in recent years has revealed tensions in cosmological parameter values inferred from different datasets. The best-known example is the Hubble tension: a disagreement reported at more than five standard deviations between the value inferred from cosmic microwave background data under the standard cosmological model and that obtained from Type Ia supernova and Cepheid measurements. Motivated by these discrepancies, we focus on tension metrics that assess the statistical significance of agreement or disagreement between cosmological parameter values inferred from different datasets.