Estimating Recovery in the Liquid–Liquid Extraction Chemical Space

Chemical characterization studies for medical devices rely on extract preparation techniques such as liquid–liquid extraction to render the more amenable instrumental analysis. Considering these are generally non-targeted, effect of full range potential analytes must be considered. However, there is currently no workflow evaluating impact non-targeted extractables. Herein, we present a framework approaching this problem by defining applicable chemical space, selecting model that appropriately predicts recovery from method, example chemicals with expected recoveries, and experimentally measuring recoveries verify performance under laboratory conditions. The demonstrated extraction, demonstrably dependent extractable pKa distribution coefficient. Method parameters including pH, volume ratios, iterations were also considered critical factors in determining recovery. analytical method was subsequently verified incorporating deliberate deviations part robustness determination using central composite design. Overall, selected work resulted coverage approximately 85% space. Further, various permutations evaluated, found predict root-mean-square error 19%. A result approach increased clarity use surrogate standards evaluation methods, indicating they should chosen based variables included predictive model. Additionally, can applied space so estimations made about which relevant likely poorly recovered during preparation. This understand analysis how considerations integrated into development validation.