Renaissance of Historical Nuclear Fusion Data: How AI and Modern Data Analysis Decipher Fundamental Resonances of the Polarized 2H(d,p)3H Reaction
This paper demonstrates a new paradigm in nuclear physics: the computational reanalysis of historical experimental data ("dark data") using modern algorithms and Artificial Intelligence (AI). Focusing on tensor-polarized deuteron-deuteron fusion measurements conducted in 1966 at the Physics Institute of the University of Erlangen-Nuremberg (Nuclear Physics A122/2), the original evaluation was restricted to second-degree Legendre polynomials (d₂) due to the limitations of punch-card mainframes. Using an AI-generated and iteratively optimized Python framework, a high-resolution Legendre fit up to degree d₄ was achieved for the first time. The reanalysis uncovers a sharply defined resonance structure at 123 keV (laboratory acceleration voltage), marked by a synchronous sign reversal of all fit coefficients (c₀ to c₄) and an abrupt transition of quantum-mechanical partial waves from P-wave (S=1) dominance below 120 keV to S-wave (S=0) and D-wave (S=2) dominance above 130 keV. Calibrated against the Bosch-Hale parametrization, the results also reveal a significant branching divergence favoring the neutron channel above 120 keV.
Authors
- H. W. Franz
Institutions
- Friedrich-Alexander-Universität Erlangen-Nürnberg (DE)
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-16
- DOI
- https://doi.org/10.5281/zenodo.24291085
- Primary Topic
- Nuclear physics research studies
- Type
- article
- Field-Weighted Citation Impact
- 0.00