Research

Several puzzles remain unsolved in particle physics, cosmology, and gravity: what is the nature of dark matter and dark energy? How did the matter-antimatter asymmetry form in the early Universe? What is the particle nature of inflation? And how can we reconcile general relativity with quantum mechanics?

My research tackles these questions by combining theoretical modelling with insights from astrophysical observations and terrestrial experiments. In recent years, I have been particularly intersted in understanding how we can use gravitational wave observations to answer these fundamental questions. As part of this effort, back in 2020, I joined the NANOGrav Collaboration. There, I leverage my background as a theorist to design and lead searches for signatures of new physics in pulsar-timing arrays (PTAs). In 2023, I helped establis the NANOGrav New Physics Working Group, which I have co-chaired since its inception. In this role, I coordinate the research efforts aimed at identifying new physics signals in PTA data.

Gravitational wave backgrounds

For centuries, we have studied the Universe through the electromagnetic radiation emitted by celestial objects. This approach has been extraordinarily successful, but it also has fundamental limitations: many interesting phenomena are electromagnetically faint, obscured, or simply do not emit light at all. For example, in the early Universe, photons were tightly coupled to the hot plasma of charged particles, making the cosmos opaque until recombination. As a result, any electromagnetic radiation produced at earlier times was repeatedly scattered and cannot reach us directly today.

However, the first direct detection of Gravitational Waves by the LIGO and Virgo collaborations provided us with a new probe to study environments and phenomena inaccessible to electromagnetic radiation. For example, thanks to their extremely weak interactions with matter, gravitational waves can travel unperturbed in the primordial plasma. Therefore, any gravitational wave signal produced in the early Universe would travel through cosmic history largely unperturbed, carrying information about these early stages of the Universe’s history.

In the summer of 2023, several PTA collaborations announced compelling evidence for a Gravitational Wave Background (GWB), a constant hum of gravitational waves permeating our Universe (you can think about it as the gravitational wave analogues of the Cosmic Microwave Background). This detection, which was enabled by a decade-long monitoring of a collection of millisecond pulsars, opens a range of fundamental questions with far-reaching implications for early Universe physics and astrophysics. First, what is the source of this signal? Is it generated by a cosmic population of supermassive black hole binaries (SMBHB) residing at the center of most massive galaxies, or could it be a relic from the early Universe? And beyond this, once a source is identified, how can we reconstruct its properties?

A large fraction of my current research work revolves around answering these two questions. I have led the first analysis trying to discriminate between an astrophysical and primordial origin of the GWB. I have developed a publicly available code, PTArcade, that allows for easy implementation of Bayesian inference analysis of PTA data, and developed a broad array of data analysis techniques aimed at discriminating and characterizing sources of GWBs (see the sections below).

GWB anisotropies

One of the most promising ways to identify the source of the GWB observed by PTA collaborations is to reconstruct the sky distribution of its power. Indeed, if the GWB is produced by a population of inspiraling SMBHBs, clustering of host galaxies and source discreteness are expected to induce a detectable level of anisotropies in this sky distribution. On the other hand, the level of anisotropies produced by most primordial sources is well below the sensitivity of present and future PTAs.

Therefore, detecting anisotropies would provide strong evidence in favor of an astrophysical origin of the signal. At the same time, a null anisotropy detection could be used to constrain the properties of the SMBHB population and eventually rule out this interpretation in favor of a cosmological interpretation. Indeed, in a series of works, I have shown how GWB anisotropy detection (or lack thereof) conveys specific information about the SMBHB population that can be used to break degeneracies and allow for a more accurate reconstruction of its properties than possible with spectral information alone. 

At the same time, fully exploiting this opportunity requires analysis methods that are both sensitive and robust to real-world PTA data. Many of the techniques currently used to reconstruct anisotropies are not yet optimal, and developing improved, end-to-end anisotropy pipelines is a key part of my research activities. As part of this effort, I have recently developed anisotropy detection strategies that –for the first time– take into account the impact of cosmic variance, and are frequency resolved – a crucial ingredient in any realistic anisotropy detection pipeline since the anisotropies produced by SMBHB are expected to show strong frequency dependence. And I am currently developing a new detection strategy based on simulation-based inference.

Testing dark matter at GW observatories

The remarkable precision of gravitational-wave experiments can also be leveraged to search for signals beyond gravitational waves themselves. By measuring tiny, correlated perturbations—either as phase shifts in laser interferometers or as pulse-arrival-time variations in pulsar timing arrays—these instruments can act as sensitive probes of new, weakly coupled physics. In particular, I am interested in how gravitational-wave datasets can be used to test dark-matter scenarios that produce subtle oscillatory, transient, or stochastic signatures.

For example, I have studied how small clumps of dark matter drifting through the Milky Way could give the Earth and/or a pulsar a tiny “nudge,” i.e. a slight Doppler acceleration. In pulsar timing data, that would show up as a subtle, time-dependent shift in the pulse arrival times—something we can explicitly search for (or use to set limits if we don’t see it). This allowed us to set constraints on the local abundance of these objects (which are predicted by several dark matter models).

Another kind of dark matter models that can be tested at gravitational-wave experiments are models of ultralight dark matter (a.k.a fuzzy dark matter). In these models, the dark matter is a bosonic particle with a mass m << eV, which coherently oscillates as a classical wave with angular frequency set by the dark matter mass (the field also presents stochastic oscillations at lower frequencies, which can be similarly looked for). These oscillations in the dark matter field source metric oscillations that we can look for using gravitational-wave observatories. In the most recent analysis of the NANOGrav data, we found no evidence of such oscillations, which allowed us to set a constraint on the local abundance of these ultralight fields.

If these ultralight dark matter fields couple directly to Standard Model particles, their phenomenology becomes more interesting. Indeed, in this cases, the oscillations in the dark matter fields produce oscillations in the value of fundamental constants, which can be tested in a variety of ways in pulsar timing or interferometers. Looking for these kinds of signals in pulsar timing data, we were able to set constraints on a variety of DM-Standard Model couplings that rival the ones placed by other terrestrial probes.

Dark matter phenomenology

I am also interested in studying the phenomenology of dark matter particles in the early universe and in direct detection experiments. During my PhD I investigated the role that bound states of dark matter particles can play in the formation of the dark matter relic abundance. The results of my research showed that these bound states –if unstable– can significantly alter the dark matter relic density, or — if stable– provide a viable dark matter candidate (even if the constituents are charged under QCD!).

I have also studied how several light dark matter candidates can be searched for in direct detection experiments using electronic excitations in semiconductors, or collective crystal excitations.