A number of computational schemes based on different Density Functional Theory (DFT) functionals in combination with a number of basis sets were tested in the calculation of 1H and 13C NMR chemical shifts of strychnine, as a typical representative of the vitally important natural products, and used as a challenging benchmark and a rigorous test for such calculations. It was found that the most accurate computational scheme, as compared with experiment, was PBE0/pcSseg‐4//pcseg‐3 characterized by a mean absolute error of 0.07 ppm for the range of about 7 ppm for 1H NMR chemical shifts and that of only 1.13 ppm for 13C NMR chemical shifts spread over the range of about 150 ppm. For more practical purposes, including investigation of larger molecules from this series, a much more economical computational scheme, PBE0/pcSseg‐2//pcseg‐2, characterized by almost the same accuracy and much less computational demand, was recommended.

中文翻译：

---

天然产物的 1 H 和 13 C NMR 化学位移的 DFT 计算方案，以士的宁为例

在计算马钱子碱的 1H 和 13C NMR 化学位移时，测试了许多基于不同密度泛函理论 (DFT) 泛函结合多个基组的计算方案，作为极其重要的天然产物的典型代表，并且用作此类计算的具有挑战性的基准和严格的测试。结果发现，与实验相比，最准确的计算方案是 PBE0/pcSseg-4//pcseg-3，其特征在于 1H NMR 化学位移约 7 ppm 的平均绝对误差为 0.07 ppm，而13C NMR 化学位移只有 1.13 ppm，分布在约 150 ppm 的范围内。出于更实际的目的，包括研究本系列中更大的分子，一个更经济的计算方案，PBE0/pcSseg-2//pcseg-2，