a-SiO2 structures
收藏资源简介:
该数据集包含1000个非晶态二氧化硅(a-SiO2)结构,通过熔融-淬火-退火分子动力学模拟生成,涵盖了从2.0到2.7 g/cm³的密度范围。数据集的创建过程包括使用LAMMPS进行分子动力学模拟,并采用CHIK势能模型进行淬火和退火处理,以确保结构的多样性和高质量。该数据集主要用于训练和评估机器学习模型,特别是在预测固态核磁共振参数方面,旨在解决复杂材料系统中的静态和动态行为预测问题。
This dataset contains 1000 amorphous silica (a-SiO₂) structures generated through melt-quench-annealing molecular dynamics simulations, spanning a density range from 2.0 to 2.7 g/cm³. The construction of this dataset utilizes LAMMPS to perform molecular dynamics simulations, and adopts the CHIK interatomic potential model for quenching and annealing processes to ensure structural diversity and high-quality results. This dataset is primarily employed for training and evaluating machine learning models, especially for predicting solid-state nuclear magnetic resonance (NMR) parameters, aiming to address the prediction of static and dynamic behaviors in complex material systems.

- 1Graph-neural-network predictions of solid-state NMR parameters from spherical tensor decomposition牛津大学 · 2024年



