Developing Powder Crystallography for Molecules using NMR Parameters and Computations

核磁共振（NMR）实验在科学中被广泛用于阐明化学结构和动力学的许多方面。近年来，核磁共振实验已被应用于固态有机分子，以进行晶体结构的细化、选择、验证和测定。在上述任务中，将NMR数据定期与基于输入结构生成的计算输出进行比较。广义上讲，这一过程被称为“NMR结晶学”（NMR-X）。该研究计划旨在改进当前协议的几个方面，并提出额外的创新，旨在提高NMR-X的准确性，效率和鲁棒性。我们的研究计划在未来五年的三个重点领域总结如下。1.我们将提高计算的结晶分子有机物的NMR参数的准确性，同时使效率的牺牲最小。目前最先进的NMR-X计算主要依赖于赝势来描述核心电子，平面波来描述价电子，并使用周期性边界条件来复制晶格的平移对称性。这种方法在计算NMR参数时相当准确，但有几个基本的限制。例如，执行包括精确电子交换的计算是非常低效的。这在目前的NMR-X方法中造成了依赖系统的“盲点”，必须加以纠正。该研究计划将开发一种局部嵌入方法，该方法将为更大的有机分子组产生比目前情况更准确的NMR参数。2.该研究计划将产生更强大的NMR-X指标。从本质上讲，指标是“统治者”，其中那些在NMR-X社区建立一个拟议的候选结构（或其他结构表征任务）的质量，这是我们的论点，目前的指标是不健全的，因为它们可能会导致相互冲突的结构质量的估计。对于几种核素，放射性核素目前以某种形式存在（例如，1H，13 C，15 N，17 O）和几个NMR参数（例如，化学位移）。我们将通过首先建立一个可靠的NMR参数和NMR信号数据的数据库，其结构是众所周知的化学系统，产生更强大的指标。然后，我们将扩展目前的指标，包括温度的影响，并随后考虑替代的数学结构的指标本身。3.通过将获得更准确的NMR参数计算和强大的NMR-X度量所实现的进步结合起来，我们将开发NMR-X软件，该软件可用于改进使用衍射方法确定的化学结构，重点关注分辨率差的晶体结构。这项研究计划将取得有意义的成果，对制药行业产生影响。其他应用领域包括抗菌剂和共晶材料。

Nuclear magnetic resonance (NMR) experiments are widely used in science to elucidate many aspects of chemical structure and dynamics. Recently, NMR experiments have been applied to organic molecules in the solid state to perform crystal structure refinements, selections, verifications, and determinations. In the above tasks, the NMR data are regularly compared against computational outputs generated based on an input structure. Broadly speaking, this process is known as `NMR crystallography' (NMR-X). This research program seeks to improve several aspects of current protocols and proposes additional innovations with the aim of enhancing the accuracy, efficiency, and robustness of NMR-X. The three focus areas of our research program over the next five years are summarized below. 1. We will improve the accuracy of calculated NMR parameters for crystalline molecular organics while making minimal sacrifices in efficiency. Current state-of-the-art NMR-X computations rely largely on pseudopotentials to describe core electrons, planewaves to describe valence electrons, and use periodic boundary conditions to replicate the translational symmetry of the crystal lattice. This approach is fairly accurate when calculating NMR parameters yet has several fundamental limitations. For example, performing calculations that include exact electron exchange are highly inefficient. This creates system-dependent `blind spots' in present NMR-X approaches, which must be rectified. This research program will develop a local embedding approach that will produce more accurate NMR parameters for a larger set of organic molecules than is currently the case. 2. This research program will generate more robust NMR-X metrics. Essentially, metrics are the `rulers' by which those in the NMR-X community establish the quality of a proposed candidate structure (or other structure characterization task), and it is our contention that current metrics are not robust as they can lead to conflicting estimates of structure quality. Metrics currently exist in some form for several nuclides (e.g., 1H, 13C, 15N, 17O) and several NMR parameters (e.g., chemical shifts). We will generate more robust metrics by first establishing a database of reliable NMR parameters and NMR signal data for chemical systems whose structures are well known. We will then extend current metrics to include temperature effects and subsequently consider alternative mathematical structures for the metrics themselves. 3. By pairing the advances realized from access to more accurate NMR parameter computations and robust NMR-X metrics, we will develop NMR-X software that can be used to refine chemical structures determined using diffraction methods, with a focus on poorly resolved crystal structures. This research program will achieve meaningful results with impact in the pharmaceutical industry. Further application areas include antimicrobials and co-crystalline materials.