Revealing the role of polymer in the robust preparation of the 2,4-dichlorophenoxyacetic acid metastable crystal form by AI-based image analysis

Abstract

The selective preparation of pure polymorphs is of great significance in the pharmaceutical industry. This work aims to prepare the pure metastable Form II of the 2,4-dichlorophenoxyacetic acid (2,4-D) and investigate the effects of polymer on the solution-mediated phase transformation (SMPT) with the assistance of AI-based image analysis. Firstly, a convolutional neural network model was built to segment and classify crystal Form II (needle) and crystal Form I (bulk). Then, the polymorphic transition kinetics based on the changes in the area of the two 2,4-D crystal forms over time was precisely and efficiently analyzed. The results revealed that the polymer additives mainly inhibited the dissolution of crystal Form II and the nucleation of crystal Form I, thereby obtaining pure metastable Form II of 2,4-D. The AI-based image analysis of the process evolution strategy offers insightful guidance on the manufacture of pure metastable crystal forms in the pharmaceutical and chemical industry. Metastable needle-like Form II of 2,4-D could be robustly prepared with polymer additive assistance. A convolutional neural network model, Crys-form CNN, was built to visualize and quantify the additives' effects on the crystal transformation process and determine the rate-controlling steps. • Pure metastable of the 2,4-D was robustly prepared with the assistance of polymer. • An AI-based image analysis model, Crys-form CNN, was built. • The SMPT of 2,4-D in the cuvette was in-situ monitored and visualized. • The effects of polymers on the SMPT were quantified by Crys-form CNN.