arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2402.09128cond-mat.mtrl-scicond-mat.soft
收藏

机器学习辅助预测有机盐结构性质

Machine learning assisted prediction of organic salt structure properties

Ethan P. Shapera, Dejan-Krešimir Bučar, Rohit P. Prasankumar, Christoph Heil

AI总结:

提出基于晶体图奇异值的机器学习方法,利用未弛豫结构预测DFT弛豫后有机盐晶体的体积、每原子焓及相态,可低成本外推至新化学系统,用于快速筛选实验可行的晶体结构。

摘要:

我们展示了一种基于机器学习的方法,该方法能够在未弛豫结构的基础上预测弛豫后的晶体结构性质。与完整的晶体图表示相比,使用晶体图奇异值将描述晶体所需的特征数量减少了一个数量级以上。我们使用晶体图奇异值表示构建机器学习模型,以便基于随机生成的未弛豫晶体结构,预测DFT弛豫后的有机盐晶体的体积、每原子焓以及金属相与半导体相。初始基础模型经过训练,将89,949个随机生成的盐结构(由不同比例的1,3,5-三嗪和HCl形成)与DFT弛豫后结构对应的体积、每原子焓和相态关联起来。我们进一步证明,通过包含来自新系统的2,000到10,000个晶体结构,基础模型能够外推到新的化学系统。在用大量数据点训练单个模型后,可以以显著更低的成本进行扩展。所构建的机器学习模型可用于快速筛选大量随机生成的有机盐晶体结构,并高效地精选出最可能通过实验实现的结构。这些模型既可以作为独立的晶体结构预测器使用,也可以作为过滤步骤整合到更复杂的工作流程中。

英文摘要:

We demonstrate a machine learning-based approach which predicts the properties of crystal structures following relaxation based on the unrelaxed structure. Use of crystal graph singular values reduces the number of features required to describe a crystal by more than an order of magnitude compared to the full crystal graph representation. We construct machine learning models using the crystal graph singular value representations in order to predict the volume, enthalpy per atom, and metal versus semiconducting phase of DFT-relaxed organic salt crystals based on randomly generated unrelaxed crystal structures. Initial base models are trained to relate 89,949 randomly generated structures of salts formed by varying ratios of 1,3,5-triazine and HCl with the corresponding volumes, enthalpies per atom, and phase of the DFT-relaxed structures. We further demonstrate that the base model is able to extrapolate to new chemical systems with the inclusion of 2,000 to 10,000 crystal structures from the new system. After training a single model with a large number of data points, extension can be done at significantly lower cost. The constructed machine learning models can be used to rapidly screen large sets of randomly generated organic salt crystal structures and efficiently downselect the structures most likely to be experimentally realizable. The models can be used either as a stand-alone crystal structure predictor or incorporated into more sophisticated workflows as a filtering step.

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。

↑