CDS&E: Molecular Modeling of Solute Precipitate Nucleation

结晶是一种在食品、特种化学品和药物分子的生产中发挥关键作用的过程，但人们对此知之甚少。例如，许多活性药物成分难以结晶，溶解性差，或具有多种晶型，具有不同的物理性质，但这些性质直接影响药物的疗效。因此，能够预测分子结晶行为的计算机模型在指导药物开发策略方面可能有价值。结晶的第一个阶段涉及一个称为成核的过程，在这个过程中，分子聚集在一起，形成一个大到足以生长成晶体的类似固体的团簇。成核通常只涉及几个分子，并且可能受到表面小杂质或缺陷的强烈影响，这使得实验研究具有挑战性。成核作用也与药物的结晶性和出现的结晶形式直接相关。在化学、生物工程、环境和运输系统司以及化学司化学理论、模型和计算方法计划的支持下，该项目将开发新的、普遍适用的计算机模拟方法和软件，以研究晶体分子从溶液中成核的过程。这些方法将被用来研究活性药物成分的成核作用，并了解溶剂和实验条件对其结晶的作用。在项目过程中开发的软件和基本理解将通过生产新药、先进材料和新的高科技工作岗位而直接造福社会。这项研究将被整合到旨在教育和接纳传统上在高等STEM教育中代表不足的少数族裔的外联工作中。该项目旨在开发新的、普遍适用的分子模拟方法，以研究分子晶体从溶液中成核。这些方法将直接模拟恒定过饱和度的降水，并将明确包括衡量溶剂结构的集体变量，而不仅仅是溶质。这些方法将用于研究溶剂如何定性地改变成核机制，并从分子水平上解释在许多化学相似分子对中观察到的不同沉淀行为。为了实现这一目标，该项目将：(1)开发一种方法来生成对本体溶剂结构敏感的有序参数，并使用它来分析以前关于溶剂对成核机制影响的研究数据；(2)开发一种通用方法来生成恒定过饱和条件下溶质沉淀的最小自由能路径，包括对溶剂结构敏感的有序参数；以及(3)使用新方法来获得磺胺二甲嘧啶和磺胺二甲嘧啶这两种结晶行为非常不同的几乎相同的分子的成核路径，并解释这些差异的原因。该方法是集合变量弦方法(SMCV)的一种新形式，可用于渗透系综、巨正则系综和Gibbs系综等开放系统系综中。PI(Santiso)已经成功地使用SMCV研究了过冷熔体中纯物质的成核，这一新的公式将使模拟现实条件下的溶质沉淀成为可能。这项研究产生的工具将使科学界和工程界能够使用现实的分子模型来模拟和研究结晶。这个工具箱将加速药物和其他固体产品计算设计的创新。此外，能够在开放系统中模拟激活过程的方法和软件将能够应用于其他领域，如催化、分离和溶液化学。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Crystallization is a process that plays a crucial role in the production of food products, specialty chemicals, and pharmaceutical molecules but remains poorly understood. For example, many active pharmaceutical ingredients are difficult to crystallize, have poor solubility, or have multiple crystalline forms with different physical properties, yet these properties have a direct impact of the efficacy of the drug. Accordingly, computer models capable of predicting molecular crystallization behavior can be valuable in guiding drug development strategies. The first stages of crystallization involve a process called nucleation, where molecules come together into a solid-like cluster large enough to grow into a crystal. Nucleation typically involves only a few molecules and can be drastically affected by small impurities or defects in surfaces, making it challenging to study experimentally. Nucleation is also directly related to the crystallizability of a drug and to the crystalline form that emerges. With support from the Division of Chemical, Bioengineering, Environmental, and Transport Systems and from the Chemical Theory, Models, and Computational Methods program in the Division of Chemistry, this project will develop new, generally applicable computer simulation methods and software to study the nucleation of crystal molecules from a solution. These methods will be used to study the nucleation of active pharmaceutical ingredients and to understand the role of the solvent and experimental conditions on their crystallization. The software and fundamental understanding developed over the course of the project will yield direct benefits to society through the production of new drugs, advanced materials, and new high-tech jobs. The research will be integrated in outreach efforts geared toward the education and inclusion of minorities traditionally underrepresented in higher STEM education.This project aims to develop new, generally applicable molecular simulation methods to study the nucleation of molecular crystals from solution. These methods will directly model precipitation at constant supersaturation and will explicitly include collective variables measuring the structure of the solvent, rather than just the solute. The methods will be used to study how solvents qualitatively change the nucleation mechanism and to provide a molecular-level explanation for the different precipitation behavior observed in many pairs of chemically similar molecules. Toward this goal, the project will: (1) develop a method to generate order parameters sensitive to the structure of bulk solvents, and use it to analyze data from previous studies on solvent effects on nucleation mechanisms; (2) develop a general method to generate minimum free energy paths for solute precipitation, under constant supersaturation conditions, including order parameters sensitive to solvent structure; and (3) use the new methods to obtain nucleation paths for sulfadiazine and sulfamerazine, two nearly identical molecules with very different crystallization behavior, and elucidate the causes for those differences. The proposed method is a new formulation of the String Method in Collective Variables (SMCV) that can be used in open-system ensembles such as the osmotic, grand canonical, and Gibbs ensembles. The PI (Santiso) has successfully used the SMCV to study the nucleation of pure substances from undercooled melts, and this new formulation will enable modeling solute precipitation under realistic conditions. The tools resulting from this research will enable the scientific and engineering community to simulate and study crystallization using realistic molecular models. This toolbox will accelerate innovation in the computational design of drugs and other solid products. Furthermore, methods and software enabling simulation of activated processes in open systems will enable applications in other areas such as catalysis, separations, and solution chemistry.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.