The development of NMR crystallography methods requires a reliable database of chemical shifts measured for systems with known crystal structure. We measured and assigned carbon and hydrogen chemical shifts of twenty solid natural amino acids of known polymorphic structure, meticulously determined using powder X-ray diffraction. We then correlated the experimental data with DFT-calculated isotropic shieldings. The small size of the unit cell of most amino acids allowed for advanced computations using various families of DFT functionals, including generalized gradient approximation (GGA), -GGA and hybrid DFT functionals. We tested several combinations of functionals for geometry optimizations and NMR calculations. For carbon shieldings, the widely used GGA functional PBE performed very well, although an improvement could be achieved by adding shielding corrections calculated for isolated molecules using a hybrid functional. For hydrogen nuclei, we observed the best performance for NMR calculations carried out with structures optimized at the hybrid DFT level. The high fidelity of the calculations made it possible to assign additional signals that could not be assigned based on experiments alone, for example signals of two non-equivalent molecules in the unit cell of some of the amino acids.

中文翻译：

---

氨基酸的核磁共振晶体学

核磁共振晶体学方法的发展需要一个可靠的数据库，其中包含针对已知晶体结构的系统测量的化学位移。我们测量并指定了二十种已知多晶型结构的固体天然氨基酸的碳和氢化学位移，并使用粉末 X 射线衍射仔细确定。然后，我们将实验数据与 DFT 计算的各向同性屏蔽相关联。大多数氨基酸的晶胞尺寸较小，允许使用各种 DFT 泛函进行高级计算，包括广义梯度近似 (GGA)、-GGA 和混合 DFT 泛函。我们测试了几种用于几何优化和 NMR 计算的泛函组合。对于碳屏蔽，广泛使用的 GGA 泛函 PBE 表现非常好，尽管可以通过添加使用混合泛函为孤立分子计算的屏蔽校正来实现改进。对于氢核，我们观察到在混合 DFT 水平上优化结构时进行 NMR 计算的最佳性能。计算的高保真度使得可以分配仅根据实验无法分配的额外信号，例如一些氨基酸的晶胞中两个非等价分子的信号。