School of Life Sciences, University of Bradford, Bradford, BD7 1DP, UK.
Phys Chem Chem Phys. 2011 Dec 7;13(45):20361-70. doi: 10.1039/c1cp22169h. Epub 2011 Oct 13.
Predicting the crystal structure of an organic molecule from first principles has been a major challenge in physical chemistry. Recently, the application of Density Functional Theory including a dispersive energy correction (the DFT(d) method) has been shown to be a reliable method for predicting experimental structures based purely on their ranking according to lattice energy. Further validation results of the application of the DFT(d) method to four organic molecules are presented here. The compounds were targets (labelled molecule II, VI, VII and XI) in previous blind tests of crystal structure prediction, and their structures proved difficult to predict. However, this study shows that the DFT(d) approach is capable of predicting the solid state structures of these small molecules. For molecule VII, the most stable (rank 1) predicted crystal structure corresponds to the experimentally observed structure. For molecule VI, the rank 1, 2 and 3 predicted structures correspond to the three experimental polymorphs, forms I, III and II, respectively. For molecules II and XI, their rank 1 predicted structures are energetically more stable than those corresponding to the experimental crystal structures, and were not found amongst the structures submitted by the participants in the blind tests. The rank 1 structure of molecule II is predicted to exist under high pressure, whilst the rank 1 structure predicted for molecule XI has the same space group and hydrogen bonding pattern as observed in the crystal of 1-amino-1-methyl-cyclopropane, which is structurally related to molecule XI. The experimental crystal structure of molecule II corresponds to the rank 4 prediction, 0.8 kJ mol(-1) above the global minimum structure, and the experimental structure of molecule XI corresponds to the rank 2 prediction, 0.4 kJ mol(-1) above the global minimum.
从第一性原理预测有机分子的晶体结构一直是物理化学中的一个主要挑战。最近，应用包括色散能校正的密度泛函理论（DFT(d)方法）已被证明是一种可靠的方法，可以根据晶格能对实验结构进行纯排名预测。本文进一步验证了 DFT(d)方法在四个有机分子中的应用。这些化合物是先前晶体结构预测盲测实验的目标（标记为分子 II、VI、VII 和 XI），它们的结构很难预测。然而，这项研究表明，DFT(d)方法能够预测这些小分子的固态结构。对于分子 VII，最稳定的（排名 1）预测晶体结构与实验观察到的结构相对应。对于分子 VI，排名 1、2 和 3 的预测结构分别对应于三种实验多晶型物，即 I、III 和 II。对于分子 II 和 XI，其排名 1 的预测结构在能量上比实验晶体结构更稳定，并且在盲测实验中参与者提交的结构中没有发现。分子 II 的排名 1 预测结构存在于高压下，而分子 XI 的排名 1 预测结构与观察到的 1-氨基-1-甲基环丙烷晶体中的结构具有相同的空间群和氢键模式，该晶体与分子 XI 结构相关。分子 II 的实验晶体结构对应于排名第 4 的预测，比全局最低结构高出 0.8 kJ mol(-1)，而分子 XI 的实验结构对应于排名第 2 的预测，比全局最低结构高出 0.4 kJ mol(-1)。
