Glycine is a common amino acid with relatively complex chemistry in solid state. Although several polymorphs (α, β, δ, γ, ε) of crystalline glycine are known, for NMR spectroscopy the most important is a polymorph, which is used as a standard for calibration of spectrometer performance and therefore it is intensively studied by both experimental methods and theoretical computation. The great scientific interest in a glycine results in a large number of crystallographic information files (CIFs) deposited in Cambridge Structural Database (CSD). The aim of this study was to evaluate the influence of the chosen crystal structure of α glycine obtained in different crystallographic experimental conditions (temperature, pressure and source of radiation of α glycine) on the results of periodic DFT calculation. For this purpose the total of 136 GIPAW calculations of α glycine NMR parameters were performed, preceded by the four approaches (“SP”, “only H”, “full”, “full+cell”) of structure preparation. The analysis of the results of those computations performed on the representative group of 34 structures obtained at various experimental conditions revealed that though the structures were generally characterized by good accuracy (R < 0.05 for most of them) the results of the periodic DFT calculations performed using the unoptimized structures differed significantly. The values of the standard deviations of the studied NMR parameters were in most cases decreasing with the number of optimized parameters. The most accurate results (of the calculations) were in most cases obtained using the structures with solely hydrogen atoms positions optimized. © 2018 Wiley Periodicals, Inc.

中文翻译：

---

晶体结构的选择是否会影响周期性 DFT 计算的结果？甘氨酸α多晶型GIPAW核磁共振参数计算案例

甘氨酸是一种常见的氨基酸，在固态时具有相对复杂的化学成分。尽管已知结晶甘氨酸的几种多晶型物 (α, β, δ, γ, ε)，但对于 NMR 光谱而言，最重要的是多晶型物，它被用作校准光谱仪性能的标准，因此两种实验都对其进行了深入研究。方法和理论计算。对甘氨酸的巨大科学兴趣导致在剑桥结构数据库 (CSD) 中存储了大量晶体学信息文件 (CIF)。本研究的目的是评估在不同晶体学实验条件（温度、压力和 α 甘氨酸的辐射源）中获得的 α 甘氨酸的所选晶体结构对周期性 DFT 计算结果的影响。为此，对 α 甘氨酸 NMR 参数进行了总共 136 次 GIPAW 计算，然后是四种结构制备方法（“SP”、“仅 H”、“完整”、“完整+细胞”）。对在不同实验条件下获得的 34 个代表性结构组进行的计算结果的分析表明，虽然这些结构通常具有良好的精度（其中大多数 R < 0.05），但周期性 DFT 计算的结果使用未优化的结构差异很大。在大多数情况下，研究的 NMR 参数的标准偏差值随着优化参数的数量而减少。在大多数情况下，最准确的（计算的）结果是使用仅优化氢原子位置的结构获得的。