使用基于多层感知机的神经网络确定 6 Li(p,α)3 He 反应的电子屏蔽势
Determination of electron screening potential of 6 Li(p,α)3 He reaction using MultiLayer Perceptron based neural network
AI总结：
针对轻核反应中电子屏蔽能难以精确测定的问题，提出利用基于多层感知机的人工神经网络分析 6 Li(p,α)3 He 反应，成功提取出 220 eV 的电子屏蔽势，证明该方法可作为估计轻核电子屏蔽势的有效替代工具。
摘要：
背景：理解亚库仑能区轻带电核之间的核反应在几个天体物理过程中具有重要意义。由于电子屏蔽效应，确定天体物理重要的 Gamow 区内的精确反应截面十分困难。各种多项式拟合、R-Matrix 和间接 Trojan horse 方法估计的电子屏蔽能远高于绝热极限。目的：使用基于多层感知机的人工神经网络分析获取 6 Li(p,α)3 He 的裸体天体物理 S 因子，并提取电子屏蔽能。方法：利用基于多层感知机的人工神经网络算法重新分析文献中可用的 6 Li(p,α)3 He 实验S因子，以获得能量依赖的天体物理 S 因子。同样使用该前馈人工神经网络，从 60 keV 以上的数据范围计算裸体天体物理 S 因子，在该范围电子屏蔽效应预计可忽略不计。然后通过取总屏蔽 S 因子与裸体 S 因子的比值获得电子屏蔽势。结果与结论：本研究通过基于人工神经网络的算法获得的电子屏蔽势为 220 eV。通过本分析提取的电子屏蔽势表明，人工神经网络可能是估计涉及轻核的电子屏蔽势的替代工具。
英文摘要：
Background: Understanding the nuclear reactions between light charged nuclei at sub-coulomb energy region holds significant importance in several astrophysical processes. Determination of the precise reaction cross-section within the astrophysically important Gamow range is difficult because of electron screening. Various polynomial fits, R-Matrix and Indirect Trojan horse method estimate much higher electron screening energies as compared to the adiabatic limit. Purpose: Obtain the bare astrophysical S-factor of 6 Li(p,α)3 He using Multi-Layer Perceptron based Artificial Neural Network based analysis and extract the electron screening energies. Methods: Experimental S-factor of 6 Li(p,α)3 He, available in literature, are reanalyzed using the Multi-LayerPerceptron based Artificial Neural Network based algorithm to obtain the energy dependent astrophysical S-factor. Bare astrophysical S-factor is also calculated using the same Feed-forward Artificial Neural Network from the data range above 60 keV where the electron screening effect is expected to be negligible. Electron screening potential is then obtained by taking the ratio of total shielded S-factor with the bare S-factor. Results and Conclusions: The electron screening potential obtained from the Present work through the Artificial Neural network based algorithm is found to be 220 eV. The extracted electron screening potential through the present analysis indicates that the Artificial Neural Network might be an alternative tools for estimation the electron screening potential involving light nuclei.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
