Many properties of bulk polymers are closely related to the molecular chain conformation. It is ideal if guidance exists to link molecular features to polymer chain conformation to significantly narrow down the design space. Here, we analyze a series of molecular descriptors derived from empirical force-fields and relate them to single chain conformation characterized by radius of gyration calculated from molecular dynamics simulations. Using data analyses and machine learning techniques, we identify that the weakest dihedral angle, characterized by the energy constants, along the chain backbone is the most influential descriptor determining the single chain radius of gyration.

本体聚合物的许多性质与分子链构象密切相关。如果存在将分子特征链接到聚合物链构象以显着缩小设计空间的指导，则是理想的。在这里，我们分析了一系列来自经验力场的分子描述子，并将它们与以分子动力学模拟计算的回转半径为特征的单链构象联系起来。使用数据分析和机器学习技术，我们确定沿链主干的以能量常数为特征的最弱二面角是确定单链回转半径最有影响力的描述子。