Answers are grounded in this paper and cite their sources.
In this study, we present a deep learning framework based on Physics-Informed Neural Networks (PINNs) for solving the Schrödinger equation in neutron-nucleus scattering problems. For interaction analysis, an effective optical potential with a Woods–Saxon form factor was employed, whose parameters depend on the mass and atomic numbers of the target nucleus. The proposed framework is capable of addressing both forward problems—to predict physical quantities—and inverse problems to estimate potential parameters. A comprehensive partial wave analysis was conducted for neutron energies up to 100 MeV. The model successfully calculated phase shifts and differential cross sections for partial waves from l=0 to l=6, accurately reproducing the characteristic angular distributions of each wave, consistent with Legendre polynomials. Furthermore, a systematic study on various nuclei (from 12 C to 208 Pb) and the tin isotopic chain confirmed the R ∼ A 1/3 radius dependence and the isospin dependence of the potential depth. The total cross-section dependence on energy was also examined, revealing broad resonant structures in agreement with optical model predictions. The findings demonstrate that PINNs offer a powerful, mesh-free, and flexible alternative to traditional numerical methods for precise analysis of nuclear reactions, with significant potential for parameter estimation and model validation tasks in nuclear physics.
No takes yet. Share an insight, caveat, or question.
Vahid Mirzaei Mahmoud Abadi (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: