Periodic Reporting for period 1 - NU 4 nu (NU 4 ν: nuclear ab initio methods for neutrino physics)
Período documentado: 2022-09-01 hasta 2024-08-31
The project “NU 4 ν: Nuclear Ab Initio Methods for Neutrino Physics” is driven by the requirements of next-generation neutrino oscillation experiments such as Hyper-Kamiokande (HyperK) and the Deep Underground Neutrino Experiment (DUNE). These experiments aim to achieve precise measurements of fundamental neutrino properties, with a particular focus on the charge-parity violating phase—a key to answering some of the most profound questions in particle physics. As these programs enter the Precision Frontier over the next decade, their success hinges on minimizing systematic errors, with the dominant uncertainty currently arising from the modeling of neutrino-nucleus interactions. This includes understanding the nuclear response of detectors across a broad spectrum of energies and nuclear targets, specifically medium-mass nuclei.
The “NU 4 ν” initiative is advancing nuclear structure calculations using state-of-the-art ab initio nuclear theory, extending these methods to new applications and higher energy regions relevant to electroweak interactions in neutrino oscillation experiments. The project has focused on delivering the first ab initio theoretical calculations of electroweak cross-sections for medium-mass nuclei, with rigorous estimation of theoretical uncertainties—a critical factor for Precision Frontier searches. Furthermore, these results have been integrated into the Monte Carlo event generators employed by experimental collaborations, creating a direct bridge to experimental data analysis.
The “NU 4 ν” initiative is advancing nuclear structure calculations using state-of-the-art ab initio nuclear theory, extending these methods to new applications and higher energy regions relevant to electroweak interactions in neutrino oscillation experiments. The project has focused on delivering the first ab initio theoretical calculations of electroweak cross-sections for medium-mass nuclei, with rigorous estimation of theoretical uncertainties—a critical factor for Precision Frontier searches. Furthermore, these results have been integrated into the Monte Carlo event generators employed by experimental collaborations, creating a direct bridge to experimental data analysis.
The project focused primarily on the quasi-elastic scattering mechanism (a single nucleon knockout from a nucleus), a crucial process for the HyperK and DUNE experiments, addressing both medium and high energy transfer regimes. In the initial phase, we developed the spectral function for the 16O nucleus, with particular emphasis on assessing theoretical uncertainties arising from the reconstruction of spectral functions. To achieve this, we employed a novel method based on expanding the nuclear response using orthogonal polynomials.
This spectral reconstruction method was later applied to derive the spin response in infinite neutron matter, leading to the first consistent ab initio calculation of the nuclear response, enabling rigorous uncertainty quantification, and paving the way for further applications in astrophysical contexts.
The 16O spectral function was also validated against inclusive electron scattering data, with the associated theoretical uncertainties propagated to the observables. Additionally, the spectral functions were implemented into Monte Carlo event generators, allowing for direct comparison of theoretical predictions with neutrino scattering experiments.
The spectral function approach is based on impulse approximation. Within this approach, the final state interactions of the knocked-out nucleon are initially neglected and can be accounted for only in an approximate way. To achieve a fully consistent calculation, we employed the Lorentz Integral Transform (LIT) method, applying it for the first time to obtain the transverse response of the 40Ca nucleus. This development enabled a fully consistent prediction of the electron scattering cross-section and marked a critical milestone toward analyzing the 40Ar nucleus. Our calculations revealed an unexpected excess of strength compared to experimental data. We explored the source of this discrepancy by developing a Bayesian neural network that retrieves longitudinal and transverse responses from inclusive electron scattering cross-sections, trained on multiple light and medium-mass nuclei. This novel machine-learning approach lays the foundation for further advancements in understanding neutrino-nucleus scattering cross-sections. It also pointed out to some tensions between various experimental campaigns which collected data of 40Ca and so validates the LIT results which we obtained.
Finally, both the LIT and SF approaches, developed within the same many-body framework and nuclear dynamics, were employed to perform a comparative analysis of longitudinal and transverse responses for the 4He nucleus, enabling us to investigate the effects of final-state interactions and two-body currents.
Overview of Main Results
- Calculation of the spectral function for the 16O nucleus using a novel spectral reconstruction method, allowing for rigorous uncertainty quantification.
- Implementation of the spectral function into Monte Carlo event generators, directly impacting neutrino oscillation experiments.
- First ab initio calculation of the transverse response in a medium-mass system (40Ca), enabling accurate predictions of electron scattering cross-sections.
- Development of a novel machine-learning approach, based on a Bayesian neural network, to extract nuclear responses from electron scattering data.
- Creation of a framework to calculate spin responses in nuclear matter, with applications in astrophysics.
Dissemination
The results of this project have been published in three articles in high-impact, peer-reviewed journals, with three additional articles currently under review. Over the past two years, these findings were also presented at five international workshops and three conferences, where the Principal Investigator (PI) was invited to give talks.
The PI organized two international workshops directly related to the “NU 4 ν” project. The first workshop, "Neutrino Nucleus Scattering at Low and Intermediate Energies," was held at MITP (Mainz, Germany) in June 2023 and brought together experts from the theoretical and experimental neutrino scattering communities, along with developers of Monte Carlo event generators. A second workshop, "Uncertainty Quantification in Nuclear Physics," at MITP in June 2024, targeted the nuclear theory community. Additionally, the PI was involved in several conferences, including serving as a local organizer for the 25th European Conference on Few-Body Problems in Physics in August 2023 and acting as a convener for NuInt24, the leading conference series dedicated to neutrino-nucleus interactions.
The PI was also invited to give seminars and colloquia at several institutions and deliver lectures at summer schools for Ph.D. students (e.g. the ECT* Doctoral Program and the International Neutrino Summer School 2024). Furthermore, during the project, the PI was appointed a board member of NuSTEC (Neutrino Scattering Theory Experiment Collaboration), contributing expertise in many-body ab initio theory.
These activities enabled the dissemination of the project’s findings to a broad audience of researchers from various experimental and theoretical communities, primarily focused on nuclear and particle physics.
This spectral reconstruction method was later applied to derive the spin response in infinite neutron matter, leading to the first consistent ab initio calculation of the nuclear response, enabling rigorous uncertainty quantification, and paving the way for further applications in astrophysical contexts.
The 16O spectral function was also validated against inclusive electron scattering data, with the associated theoretical uncertainties propagated to the observables. Additionally, the spectral functions were implemented into Monte Carlo event generators, allowing for direct comparison of theoretical predictions with neutrino scattering experiments.
The spectral function approach is based on impulse approximation. Within this approach, the final state interactions of the knocked-out nucleon are initially neglected and can be accounted for only in an approximate way. To achieve a fully consistent calculation, we employed the Lorentz Integral Transform (LIT) method, applying it for the first time to obtain the transverse response of the 40Ca nucleus. This development enabled a fully consistent prediction of the electron scattering cross-section and marked a critical milestone toward analyzing the 40Ar nucleus. Our calculations revealed an unexpected excess of strength compared to experimental data. We explored the source of this discrepancy by developing a Bayesian neural network that retrieves longitudinal and transverse responses from inclusive electron scattering cross-sections, trained on multiple light and medium-mass nuclei. This novel machine-learning approach lays the foundation for further advancements in understanding neutrino-nucleus scattering cross-sections. It also pointed out to some tensions between various experimental campaigns which collected data of 40Ca and so validates the LIT results which we obtained.
Finally, both the LIT and SF approaches, developed within the same many-body framework and nuclear dynamics, were employed to perform a comparative analysis of longitudinal and transverse responses for the 4He nucleus, enabling us to investigate the effects of final-state interactions and two-body currents.
Overview of Main Results
- Calculation of the spectral function for the 16O nucleus using a novel spectral reconstruction method, allowing for rigorous uncertainty quantification.
- Implementation of the spectral function into Monte Carlo event generators, directly impacting neutrino oscillation experiments.
- First ab initio calculation of the transverse response in a medium-mass system (40Ca), enabling accurate predictions of electron scattering cross-sections.
- Development of a novel machine-learning approach, based on a Bayesian neural network, to extract nuclear responses from electron scattering data.
- Creation of a framework to calculate spin responses in nuclear matter, with applications in astrophysics.
Dissemination
The results of this project have been published in three articles in high-impact, peer-reviewed journals, with three additional articles currently under review. Over the past two years, these findings were also presented at five international workshops and three conferences, where the Principal Investigator (PI) was invited to give talks.
The PI organized two international workshops directly related to the “NU 4 ν” project. The first workshop, "Neutrino Nucleus Scattering at Low and Intermediate Energies," was held at MITP (Mainz, Germany) in June 2023 and brought together experts from the theoretical and experimental neutrino scattering communities, along with developers of Monte Carlo event generators. A second workshop, "Uncertainty Quantification in Nuclear Physics," at MITP in June 2024, targeted the nuclear theory community. Additionally, the PI was involved in several conferences, including serving as a local organizer for the 25th European Conference on Few-Body Problems in Physics in August 2023 and acting as a convener for NuInt24, the leading conference series dedicated to neutrino-nucleus interactions.
The PI was also invited to give seminars and colloquia at several institutions and deliver lectures at summer schools for Ph.D. students (e.g. the ECT* Doctoral Program and the International Neutrino Summer School 2024). Furthermore, during the project, the PI was appointed a board member of NuSTEC (Neutrino Scattering Theory Experiment Collaboration), contributing expertise in many-body ab initio theory.
These activities enabled the dissemination of the project’s findings to a broad audience of researchers from various experimental and theoretical communities, primarily focused on nuclear and particle physics.
This project advanced state-of-the-art calculations in several key areas, achieving impactful progress by:
- Extending ab initio electroweak response calculations to medium-mass nuclei, specifically addressing the 40Ca nucleus.
- Establishing a direct connection between ab initio calculations and experimental efforts by implementing these results in Monte Carlo event generators—the first time such ab initio results have been incorporated.
- Developing a novel method for spectral reconstruction and applying it to both nuclear matter (with applications in astrophysics) and finite nuclei.
- Creating an innovative machine-learning approach, based on a Bayesian neural network, to extract nuclear responses from electron scattering data.
- Extending ab initio electroweak response calculations to medium-mass nuclei, specifically addressing the 40Ca nucleus.
- Establishing a direct connection between ab initio calculations and experimental efforts by implementing these results in Monte Carlo event generators—the first time such ab initio results have been incorporated.
- Developing a novel method for spectral reconstruction and applying it to both nuclear matter (with applications in astrophysics) and finite nuclei.
- Creating an innovative machine-learning approach, based on a Bayesian neural network, to extract nuclear responses from electron scattering data.