Crystal Structure Predictions for 4-Amino-2,3,6-trinitrophenol Using a Tailor-Made First-Principles-Based Force Field
Crystal Structure Predictions for 4-Amino-2,3,6-trinitrophenol Using a Tailor-Made First-Principles-Based Force Field
复制标题
使用定制的基于第一性原理的力场预测 4-氨基-2,3,6-三硝基苯酚的晶体结构
DOI：
10.1021/acs.cgd.1c01117
复制
发表时间：
2022
影响因子：
3.8
通讯作者：
Szalewicz, Krzysztof
中科院分区：
文献类型：
--
作者：
Metz, Michael P.;Shahbaz, Muhammad;Song, Hongxing;Vogt-Maranto, Leslie;Tuckerman, Mark E.;Szalewicz, Krzysztof
利用第一性原理电子结构计算和分子模拟对含能分子4-氨基-2，3，6-三硝基苯酚的晶体结构进行了预测。这种基于物理的方法由一系列步骤组成。首先，一个定制的两体势能面（PES）构建与最近开发的软件，autoPES，使用基于密度泛函理论描述的单体的自适应微扰理论[SAPT（DFT）]。的拟合过程，确保渐近正确的PES采用严格的渐近多极展开，无缝集成与SAPT（DFT）的相互作用能。接下来，通过用刚性分子生成可能的晶体结构，使用SAPT（DFT）力场最小化这些结构，并基于定制的SAPT（DFT）分子间力场和通用/SAPT（DFT）分子内力场运行柔性分子的等温等压分子动力学（MD）模拟来进行晶体结构预测（CSP）。这种工作流程导致实验观察到的结构被确定为具有最低晶格能量的形式之一，证明了CSP的第一原理，自下而上方法的成功。重要的是，我们认为，分子间的潜力，在这里的SAPT（DFT）为基础的潜力，是决定性的晶体结构，而通用/SAPT（DFT）力场可以用来代表分子内的潜力。这种力场方法简化了CSP工作流程，而不会显著影响预测的准确性。
Predictions of crystal structures from first-principles electronic structure calculations and molecular simulations have been performed for an energetic molecule, 4-amino-2,3,6-trinitrophenol. This physics-based approach consists of a series of steps. First, a tailor-made two-body potential energy surface (PES) was constructed with recently developed software, autoPES, using symmetry-adapted perturbation theory based on a density-functional theory description of monomers [SAPT(DFT)]. The fitting procedure ensures asymptotic correctness of the PES by employing a rigorous asymptotic multipole expansion, which seamlessly integrates with SAPT(DFT) interaction energies. Next, crystal structure prediction (CSP) was performed by generating possible crystal structures with rigid molecules, minimizing these structures using the SAPT(DFT) force field, and running isothermal–isobaric molecular dynamics (MD) simulations with flexible molecules based on the tailor-made SAPT(DFT) intermolecular force field and a generic/SAPT(DFT) intramolecular one. This workflow led to the experimentally observed structure being identified as one of the forms with the lowest lattice energy, demonstrating the success of a first-principles, bottom-up approach to CSP. Importantly, we argue that the accuracy of the intermolecular potential, here the SAPT(DFT)-based potential, is determinative of the crystal structure, while generic/SAPT(DFT) force fields can be used to represent the intramolecular potential. This force field approach simplifies the CSP workflow, without significantly compromising the accuracy of the prediction.