Selection of Solvent Composition in RP-HPLC by Artificial Intelligence (AI) for Sustainable Pharmaceutical Analysis

Abstract

The main objective of the work is to assess if AI-assisted solvent selection improves chromatographic performance by lowering experimental workload, solvent consumption, and environmental impact. Green analytical chemistry promotes environmentally friendly solvents, but it takes time to optimize them through trial and error. Both conventional solvents (methanol, acetonitrile) and environmental alternatives (ethanol, glycerol, and natural deep eutectic solvents) were employed for the chromatographic runs. To achieve optimal resolution, solvent ratios were sequentially modified in the conventional mode. An innovative method for accurately predicting ideal solvent systems is provided by artificial intelligence (AI). Chromatographic conditions and solvent physicochemical properties were utilized to train machine learning models for AI-based optimization. Optimal green solvent mixtures were identified by the AI system and determined experimentally employing model analytes. Whereas the conventional method optimization was more expensive in terms of solvent consumption and required many trial runs. Prediction by AI pinpointed the optimal ethanol–glycerol mixtures and saved experimental runs by nearly 60%. The AI-based method significantly minimized cost and environmental footprint while providing retention time, peak resolution, and symmetry that were the same with or better than the original. Finally, it concludes that the optimization of green solvent systems with artificial intelligence (AI) offers an effective and sustainable substitute for traditional RP-HPLC method development and in pharmaceutical analysis, leading to more efficient and eco-friendly results.