多晶型溶析结晶过程两相CFD-Micromixing-PBE模型与模拟

More than 30% of crystals in the pharmaceutical industry exibit polymorphism. Antisolvent crystallization, which is widely used in the pharmaceutical industry, is greatly affected by the micro-mixing effect, making the scale-up and control of the size, crystal size distribution (CSD) and polymorphism very difficult. In the current project, on the basis of the already developed two-phase FM-PDF (Finite-mode probability density function) micromixing model, and by incorporating heat from mixing and crystallization on the micro-scale, the two-phase CFD (computational fluid dynamics) and the polulation balace equation (PBE) for each polymorph, a coulped model, i.e., the two-phase CFD-Micromixing-PBE, will be developed for the polymorphic antisolvent crystallization process. And a souce code solver in the framework of OpenFOAM (Open source field operation and manipulation) will be developed. Then the two-phase DQMOM-IEM (Direct quadrature method of moments coupled with the interaction by exchange with the mean) micromixing model will be developed based on the two-phase FM-PDF of this project. Another coulped model with the DQMOM-IEM as the micromixing model for the polymorphic system, together with its corresponding OpenFOAM solver, will be developed then. Experiments on the antisolvent crystallization are conducted for the systems of L-histidine-ethanol-water and indomethacin-acetone-heptane. The experimentally measured data are used to validate the reliability and applicability of the model. The influence of some important factors, such as operating parameters, the internals and the type of the impeller, on the size, CSD and the content of each polymorph will be numerically investigated. The results might be useful for the design, control and scale-up for the processes of antisolvent crystallization such as L-histidine-ethanol-water and indomethacin-acetone-heptane etc. One-dimensional PBE is being used in the simulations of actual crystallization processes. This cannot accurately describe crystals of very high aspect ratio such as needle- or sheet-like crystals. As two-dimensional growth rates for the actual crystallization systems are lacking, the existing one-dimensional grwoth rate will be reassigned to the two dimensions based on some principles, such as preserving added mass and specific area. The coulped model and a solver with the two-dimensional PBE will be developed, and the antisolvent crystallzation process will be simulated. It can hopefully predict simultaneously the size, CSD, content of polymorphs and crystal habit in real crystallization processes.

超过30%的药物晶体存在多晶型现象。医药工业中应用广泛的溶析结晶过程，混合影响明显、放大效应显著，晶体大小、分布(CSD)与目标晶型含量调控困难。基于现有的两相FM-PDF微观混合模型，考虑混合与结晶热，耦合两相CFD与每一晶型PBE，建立多晶型两相CFD-Micromixing-PBE模型，开发相应的OpenFOAM求解器。发展精度和物理意义更合理的两相DQMOM-IEM微观混合模型，建立相应的耦合模型与求解器。实验以L-组氨酸-乙醇-水和吲哚美辛-丙酮-庚烷为对象。模拟结合实验揭示操作条件等对CSD、晶型含量等的影响规律，指导相关过程的放大设计。实际结晶过程的模拟PBE均采用一维尺度，难以准确描述针形等纵横比变化显著的晶体。实际体系二维生长动力学缺乏，拟将已有一维生长动力学按照生长质量和比表面积增量守恒原则二维分配，模拟实际溶析结晶过程，有望同时预测药物结晶的CSD、形貌、晶型含量。

超过30%的药物晶体存在多晶型现象。医药工业中应用广泛的溶析结晶过程，混合影响明显、放大效应显著，晶体大小、分布(CSD)与目标晶型含量调控困难。项目以吲哚美辛多晶型溶析结晶过程为研究对象，对吲哚美辛在不同混合溶剂体系中的溶解度和溶解行为进行了实验测量与分析，并采用三种热力学模型对不同晶型的溶解度进行关联。研究表明吲哚美辛的溶解行为在不容溶剂体系中存在非常大的差别。提出的溶解度参数可简单高效地用于溶析结晶过程溶剂初步筛选。多晶型的有效控制离不开准确的结晶动力学数据。基于实验、离散化粒数衡算方程（PBE）以及非线性优化算法，项目首次获得了吲哚美辛两种晶型（α与γ）溶析结晶动力学。结果表明γ晶型其二次成核相对α晶型更明显。根据PBE模型计算，γ晶型的初级成核速率仅仅影响溶析结晶最初始阶段的晶型组成，而最终产品晶型有效控制的关键在于α晶型的生长速率。在获得吲哚美辛溶析结晶热力学和动力学数据的基础上，考虑混合效应，耦合两相CFD与每一晶型PBE，建立了吲哚美辛多晶型溶析结晶过程耦合模型，开发了相应的OpenFOAM求解器。模拟结果揭示了操作条件等对CSD、晶型含量等的影响规律，可指导多晶型溶析结晶过程的放大设计。实际体系二维生长动力学缺乏。项目建立了离散化二维PBE，同时进行了一维生长动力学分解获取二维生长动力学。采用具有分析解的例子以及β L-谷氨酸结晶过程进行验证，表明建立的二维PBE求解算法以及所提出的采用一维动力学分解来获取二维生长动力学的思路是可行的。