Periodic Reporting for period 2 - NuBridge (Neutrino experiments and data as a bridge to new physics)
Periodo di rendicontazione: 2025-05-01 al 2026-04-30
Sintesi del contesto e degli obiettivi generali del progetto
The Standard Model of particle physics is the theory that incorporates our current knowledge about the nature of fundamental particles and their interactions. Despite being an incredibly successful predictive framework, the Standard Model appears to be incomplete, as several observations cannot be addressed within it (the massive nature of neutrinos, the existence of dark matter, and the matter-antimatter asymmetry of the Universe). This motivates the search for New Physics beyond the Standard Model. In the past, this search has mostly focused on exploring the high-energy frontier, by looking for new particles that can be produced in the interaction of Standard Model particles at increasingly high values of energy (this is carried out at collider facilities such as the LHC). This activity provided decisive confirmation tests of the validity of the Standard Model, but has so far produced no hints for the nature of the physics that lies beyond it.
The aim of NuBridge is to explore the possibility that the New Physics particles instead lie at a lower energy scale, but have very feeble couplings with the Standard Model ones, effectively forming a "Dark Sector". An additional motivation to focus on such low-scale dark sectors comes from the recent detection, by pulsar timing array experiments, of a stochastic gravitational-wave background in the Universe: while this signal does not by itself require physics beyond the Standard Model, it is in tension with its standard astrophysical interpretation in terms of supermassive black hole binaries, and could naturally arise from a cosmological phase transition in a low-scale dark sector, thus offering a complementary observational window onto this scenario. The test of the dark sector hypothesis requires a different kind of experiments, where a very large number of interactions is generated in fixed-target experiments, thus exploring the high-intensity frontier. NuBridge employs neutrinos as a possible bridge between the Standard Model and New Physics, following three main research axes. First, it uses data from current and next-generation neutrino experiments to test the dark sector hypothesis; neutrino experiments provide powerful beams of primary particles and large sensitive detectors, making them ideal settings to search for feebly interacting particles. Second, it looks for possible indirect manifestations of New Physics in precision neutrino data, by performing global fits of experimental data and searching for deviations with respect to the standard predictions. Third, it produces public and open-source scientific tools that can be used by researchers worldwide to study dark sector and neutrino physics, enhancing the long-term impact of the project beyond its duration.
The aim of NuBridge is to explore the possibility that the New Physics particles instead lie at a lower energy scale, but have very feeble couplings with the Standard Model ones, effectively forming a "Dark Sector". An additional motivation to focus on such low-scale dark sectors comes from the recent detection, by pulsar timing array experiments, of a stochastic gravitational-wave background in the Universe: while this signal does not by itself require physics beyond the Standard Model, it is in tension with its standard astrophysical interpretation in terms of supermassive black hole binaries, and could naturally arise from a cosmological phase transition in a low-scale dark sector, thus offering a complementary observational window onto this scenario. The test of the dark sector hypothesis requires a different kind of experiments, where a very large number of interactions is generated in fixed-target experiments, thus exploring the high-intensity frontier. NuBridge employs neutrinos as a possible bridge between the Standard Model and New Physics, following three main research axes. First, it uses data from current and next-generation neutrino experiments to test the dark sector hypothesis; neutrino experiments provide powerful beams of primary particles and large sensitive detectors, making them ideal settings to search for feebly interacting particles. Second, it looks for possible indirect manifestations of New Physics in precision neutrino data, by performing global fits of experimental data and searching for deviations with respect to the standard predictions. Third, it produces public and open-source scientific tools that can be used by researchers worldwide to study dark sector and neutrino physics, enhancing the long-term impact of the project beyond its duration.
Lavoro eseguito dall’inizio del progetto fino alla fine del periodo coperto dalla relazione e principali risultati finora ottenuti
The main research aim of NuBridge, which is the study and exploration of low scale dark sectors, received a strong motivation at the very start of the project in June 2023, when the NANOGrav collaboration published the analysis of their 15 years data det, providing for the first time evidence for a stochastic gravitational-wave background in the Universe. Given that the observed signal is in tension with common astrophysical predictions, the project focused on exploring possible new physics explanations for its origin. The analysis identified a minimal dark sector model that can reproduce the observed data from a primordial first order phase transition in the early Universe, carefully taking into account many subtleties that are often overlooked in the literature, and highlighted strong phenomenological relations that can be used by experiments to test the model. In a related study it was then established, for the first time, the quantitative impact of such low-scale phase transitions scenarios on the possible mechanisms at the origin of the matter-antimatter asymmetry of the Universe, thus establishing a new phenomenological connection between these a priori disconnected topics.
In order to effectively study the proposed models, we developed the scientific software ELENA, aimed at computing the cosmological bubble nucleation rates by using the so called tunnelling potential formalism, that allows for a fast numerical computation of the tunnelling action, as an alternative to existing tools that are instead based on the more computationally demanding bounce equation solving. The software ELENA has been released for public use as free and open-source code.
On the connections between neutrinos and physics beyond the Standard Model, we identified a new solution to the dark matter problem within the minimal seesaw mechanism, that is the minimal extension of the Standard Model that can account for massive neutrinos. In a series of works, we first proposed that dark matter can be produced in the form of sterile neutrinos in the decay of heavier neutral leptons, then carefully computed the dark matter production rates taking into account the significant thermal effects, and finally verified that simultaneous solutions exist for reproducing both neutrino masses and dark matter properties, as well as identified the observables that experiments can use to test the proposed model.
Moreover, we developed a new theoretical framework in which a "dark sector" (a hidden set of particles interacting only very weakly with ordinary matter) simultaneously explains why neutrinos have mass and provides a viable dark matter candidate, while remaining consistent with all known cosmological observations. The framework predicts observable signatures in current and future neutrino experiments, and offers a possible explanation for a long-standing unresolved gamma-ray signal from the centre of our Galaxy. To pursue direct experimental tests of these ideas, we joined the REDTOP, DUNE, NA64 and ICARUS particle physics experiments, and initiated within ICARUS a dedicated working group focused on dark sector searches.
We are also working on a global fit of neutrino data to extract the values of oscillation parameters from the combination of all available experiments worldwide. Neutrino oscillations are quantum-mechanical phenomena through which neutrinos change identity as they travel; combining data from all experiments worldwide allows for a more precise determination of neutrino properties and for the search of deviations that would signal new physics. More than seventeen experimental datasets have been implemented and validated within the framework, drawing from experiments including Super-Kamiokande, Daya Bay, IceCube, SNO, and JUNO, among others. While other analysis of this kind exist, they all use private algorithms. Within NuBridge we are extending the public code GAMBIT to include neutrino oscillations analysis, providing the first study of such that employs a fully open-source software, making the analysis totally transparent and reproducible. As part of this ongoing work, we developed and publicly released the new software PEANUTS, for the fast computation of oscillation probabilities of solar and atmospheric neutrinos; PEANUTS is based on semi-analytical algorithms and modern machine learning techniques, that greatly improve performance with respect to simple numerical solutions. PEANUTS has been used to reproduce the results of the SNO experiment and is currently implemented as backend in the GAMBIT pipeline.
Finally, we released DarkAgents, a multi-agent artificial intelligence system that combines large language models with human-written scientific code, to assist scientists in performing theoretical astroparticle physics research in a transparent and controllable way.
In order to effectively study the proposed models, we developed the scientific software ELENA, aimed at computing the cosmological bubble nucleation rates by using the so called tunnelling potential formalism, that allows for a fast numerical computation of the tunnelling action, as an alternative to existing tools that are instead based on the more computationally demanding bounce equation solving. The software ELENA has been released for public use as free and open-source code.
On the connections between neutrinos and physics beyond the Standard Model, we identified a new solution to the dark matter problem within the minimal seesaw mechanism, that is the minimal extension of the Standard Model that can account for massive neutrinos. In a series of works, we first proposed that dark matter can be produced in the form of sterile neutrinos in the decay of heavier neutral leptons, then carefully computed the dark matter production rates taking into account the significant thermal effects, and finally verified that simultaneous solutions exist for reproducing both neutrino masses and dark matter properties, as well as identified the observables that experiments can use to test the proposed model.
Moreover, we developed a new theoretical framework in which a "dark sector" (a hidden set of particles interacting only very weakly with ordinary matter) simultaneously explains why neutrinos have mass and provides a viable dark matter candidate, while remaining consistent with all known cosmological observations. The framework predicts observable signatures in current and future neutrino experiments, and offers a possible explanation for a long-standing unresolved gamma-ray signal from the centre of our Galaxy. To pursue direct experimental tests of these ideas, we joined the REDTOP, DUNE, NA64 and ICARUS particle physics experiments, and initiated within ICARUS a dedicated working group focused on dark sector searches.
We are also working on a global fit of neutrino data to extract the values of oscillation parameters from the combination of all available experiments worldwide. Neutrino oscillations are quantum-mechanical phenomena through which neutrinos change identity as they travel; combining data from all experiments worldwide allows for a more precise determination of neutrino properties and for the search of deviations that would signal new physics. More than seventeen experimental datasets have been implemented and validated within the framework, drawing from experiments including Super-Kamiokande, Daya Bay, IceCube, SNO, and JUNO, among others. While other analysis of this kind exist, they all use private algorithms. Within NuBridge we are extending the public code GAMBIT to include neutrino oscillations analysis, providing the first study of such that employs a fully open-source software, making the analysis totally transparent and reproducible. As part of this ongoing work, we developed and publicly released the new software PEANUTS, for the fast computation of oscillation probabilities of solar and atmospheric neutrinos; PEANUTS is based on semi-analytical algorithms and modern machine learning techniques, that greatly improve performance with respect to simple numerical solutions. PEANUTS has been used to reproduce the results of the SNO experiment and is currently implemented as backend in the GAMBIT pipeline.
Finally, we released DarkAgents, a multi-agent artificial intelligence system that combines large language models with human-written scientific code, to assist scientists in performing theoretical astroparticle physics research in a transparent and controllable way.
Progressi oltre lo stato dell’arte e potenziale impatto previsto (incluso l’impatto socioeconomico e le implicazioni sociali più ampie del progetto fino ad ora)
We explicitly verified that a supercooled dark scalar phase transition in the early Universe can account for the NANOGrav gravitational-wave data, by addressing the challenges previously raised in the literature that had put under question the feasibility of this explanation. Going beyond previous studies, we carefully considered the effects of a vacuum domination phase and explicitly tracked the phase transition from its onset to its completion. We showed that a conformal-like potential is required to reproduce the observed data, and that this feature imposes a strong correlation on the mass spectrum of the theory that can be targeted by dedicated experiments. Extending this result, we established for the first time a direct quantitative connection between low-scale supercooled phase transitions and the leptogenesis mechanism (a process that can be responsible for the matter-antimatter asymmetry of the Universe) showing that the entropy released by such transitions can dramatically alter the viable parameter space of leptogenesis models. This constitutes a novel phenomenological bridge between two previously disconnected frontiers of particle cosmology.
We greatly improved the algorithm for the numerical computation of the tunnelling action in the first order phase transitions dynamics, which represents the most numerically demanding step in the process, by developing the code ELENA, which requires approximately 10 milliseconds for each computation on consumer-grade computers. The software has been released as free and open-source code.
We identified a new solution to the dark matter problem within the minimal seesaw framework, a minimal extension of the Standard Model originally conceived to generate neutrino masses. We computed for the first time the dark matter production rate in this framework by coherently taking into account thermal effects, employing thermal quantum field theory in the real-time formalism, and determined the parameter space of solutions and the observables that experiments can use to test this simultaneous solution to dark matter and neutrino masses.
In a parallel line of research, we constructed a broader sub-GeV dark sector framework that simultaneously accounts for neutrino masses, provides a dark matter candidate, and satisfies all known cosmological constraints, while predicting testable signatures in neutrino experiments. This framework also offers a candidate explanation for the 511 keV gamma-ray line from the Galactic centre, an astrophysical anomaly that has remained unexplained for over fifty years.
We developed and released PEANUTS, a free and open-source code for the fast computation of solar neutrino oscillation probabilities, based on a semi-analytical approach that greatly improves numerical efficiency over purely numerical methods, enabling more effective large-scale statistical analyses. PEANUTS has been interfaced with the GAMBIT global analysis framework and subsequently extended to cover atmospheric neutrino oscillations, incorporating modern machine learning techniques to handle the more numerically demanding underlying computations. Within GAMBIT, the NeutrinoBIT neutrino module has been substantially advanced: more than seventeen experimental datasets have been implemented and validated from scratch, and an efficient approach to the statistical treatment of nuisance parameters has been introduced that makes the global parameter space exploration significantly more tractable. This work will deliver the first fully open-source global fit of neutrino oscillation data, making the analysis transparent and reproducible by any research group worldwide.
Finally, we produced DarkAgents, a multi-agent artificial intelligence system designed for theoretical astroparticle physics research. The system combines large language model agents with deterministic, human-written scientific code to build orchestrated pipelines capable of performing research tasks (such as scanning parameter spaces and computing physical observables) in an automated yet human-controllable manner. Released as free and open-source software, it represents a novel approach to AI-assisted fundamental research.
We greatly improved the algorithm for the numerical computation of the tunnelling action in the first order phase transitions dynamics, which represents the most numerically demanding step in the process, by developing the code ELENA, which requires approximately 10 milliseconds for each computation on consumer-grade computers. The software has been released as free and open-source code.
We identified a new solution to the dark matter problem within the minimal seesaw framework, a minimal extension of the Standard Model originally conceived to generate neutrino masses. We computed for the first time the dark matter production rate in this framework by coherently taking into account thermal effects, employing thermal quantum field theory in the real-time formalism, and determined the parameter space of solutions and the observables that experiments can use to test this simultaneous solution to dark matter and neutrino masses.
In a parallel line of research, we constructed a broader sub-GeV dark sector framework that simultaneously accounts for neutrino masses, provides a dark matter candidate, and satisfies all known cosmological constraints, while predicting testable signatures in neutrino experiments. This framework also offers a candidate explanation for the 511 keV gamma-ray line from the Galactic centre, an astrophysical anomaly that has remained unexplained for over fifty years.
We developed and released PEANUTS, a free and open-source code for the fast computation of solar neutrino oscillation probabilities, based on a semi-analytical approach that greatly improves numerical efficiency over purely numerical methods, enabling more effective large-scale statistical analyses. PEANUTS has been interfaced with the GAMBIT global analysis framework and subsequently extended to cover atmospheric neutrino oscillations, incorporating modern machine learning techniques to handle the more numerically demanding underlying computations. Within GAMBIT, the NeutrinoBIT neutrino module has been substantially advanced: more than seventeen experimental datasets have been implemented and validated from scratch, and an efficient approach to the statistical treatment of nuisance parameters has been introduced that makes the global parameter space exploration significantly more tractable. This work will deliver the first fully open-source global fit of neutrino oscillation data, making the analysis transparent and reproducible by any research group worldwide.
Finally, we produced DarkAgents, a multi-agent artificial intelligence system designed for theoretical astroparticle physics research. The system combines large language model agents with deterministic, human-written scientific code to build orchestrated pipelines capable of performing research tasks (such as scanning parameter spaces and computing physical observables) in an automated yet human-controllable manner. Released as free and open-source software, it represents a novel approach to AI-assisted fundamental research.