Systematic Parameter Optimization of Quantum Molecular Dynamics Models for Hadron Therapy Using Multi-Ion Fragmentation Data
Quantum molecular dynamics (QMD) models are widely used to simulate nuclear fragmentation in hadron therapy, but their predictive accuracy depends strongly on parameters that are often selected empirically. We developed an optimized QMD framework by systematically calibrating three parameters for a relativistic mean-field model with the NS2 parameter set and Skyrme models with the SLy4 and SkM* parameter sets: the wave-packet width L, maximum evolution time Tm, and impact-parameter envelope factor benv. The wave-packet width was determined from experimental charge radii, whereas Tm and benv were parameterized as functions of incident kinetic energy and reaction-system mass and optimized using proton- and heavy-ion-induced fragmentation data over 30-400 MeV/u. Performance was compared with the original LiQMD, Binary Cascade, and Liege Intranuclear Cascade models. The optimized Tm depended strongly on incident energy but only weakly on system mass, indicating that the transition from the dynamical QMD stage to statistical de-excitation is governed mainly by collision energy. In contrast, benv showed model-dependent behavior: NS2 favored larger peripheral-collision contributions for lighter systems at low energies, whereas the Skyrme models showed relatively weak energy and mass dependence. The optimized parameterizations improved agreement with experimental fragment production cross sections, angular distributions, and energy distributions. The optimized Skyrme models achieved the best overall performance and outperformed the cascade models for most datasets. This framework provides a physically consistent description of nuclear fragmentation across multiple observables and may improve calculations of secondary-particle transport, dose deposition, and linear energy transfer in hadron therapy.