基于TALYS的Geant4中子产额修正方法研究

Study on a TALYS-Based Correction Method for Neutron Yield in Geant4

摘要: 深地低本底粒子探测实验中，宇宙线缪子产生的中子是重要本底来源，中子产额模拟精度直接影响本底评估可靠性。已有研究表明，Geant4 在有机液体闪烁体中的中子产额预测存在系统偏差，主要与其中低能区(E <200 MeV)强子-核非弹性反应截面参数化近似有关。本文提出一种基于TALYS 核反应计算的Geant4 中子产额修正方法，在保持Geant4 级联结构和未态运动学分布不变的基础上，利用TALYS 截面数据对中低能次级中子、质子和伽马射线与¹²C 的非弹性反应概率进行离线逐顶点重赋权。与大亚湾实验测量值比较，FTFP_BERT_HP 偏差由+20.0% 降至+6.04%，QGSP_BIC_HP 由+13.0% 降至-0.39%。结果表明，该方法可显著改善液闪中子产额模拟精度，为深地低本底实验中子本底研究提供有效修正途径。

Abstract: In deep underground low-background particle detection experiments, neutrons induced by cosmic-ray muons constitute one of the dominant background sources limiting the detector sensitivity, The neutron yield, which characterizes the neutron production capability of muons traversing matter, is a key physical observable whose simulation accuracy directly impacts the reliability of muon-induced background studies,Previous studies have indicated that the Geant4 Monte Carlo particle transport toolkit exhibits systematic deviations in predicting neutron yields in organic liquid scintillators. One of the primary sources of this discrepancy arises from the use of parameterized approximations for hadron-nucleus inelastic cross sections in the intermediate and low energy range (E< 200 MeV).To address this issue, we propose a correction method for neutron yield simulations in Geant4 based on nuclear reaction cross sections calculated with the TALYS code. While preserving the original cascade structure and final-state kinematic distributions in Geant4, this method employs high-precision cross section data from TALYS to perform an offline, vertex-by-vertex reweighting of inelastic interaction probabilities for secondary neutrons, protons, and gamma rays on ¹²C in the intermediate and low energy region.A comparison with measurements from the Daya Bay neutrino experiment demonstrates that, after incorporating TALYS-based cross sections, the relative deviations in neutron yield are signifcantly reduced for both physics lists. Specifcally, the discrepancy for FTFP_BERT_HP decreases from +20.0% to +6.04%, while that for QGSP_ BIC_HP improves from +13.0% to -0.39%. The corrected simulation results show substantially improved agreement with experimental data.This approach provides an effective pathway to enhance the predictive accuracy of Geant4 for neutron yields in liquid scintillators, and offers a refined methodology for modeling muon-induced neutron backgrounds in deep underground low-background experiments.