Data-Driven Theoretical Modeling of Centrifugal Step Emulsification and Its Application in Comprehensive Multiscale Analysis

Tailored droplet generation is crucial for droplet microfluidics that involve samples of varying sizes. However, the absence of precise predictive models forces droplet platforms to rely on empiricism derived from extensive experiments, underscoring the need for comprehensive modeling analysis. To address this, a novel customized assembled centrifugal step emulsifier (CASE) is presented by incorporating a “jigsaw puzzles” design to efficiently acquire large-scale experimental data. Numerical simulations are utilized to analyze fluid configurations during step emulsification, identifying a key connection tube that determines droplet size. By training and verifying with the experimental and simulation datasets, a comprehensive theoretical model is established that allows for the preliminary design of the droplet size and generation frequency with an average error rate of 4.8%, successfully filling a critical gap in existing field. This predictive model empowers the CASE to achieve all-in-one functionality, including droplet pre-design, generation, manipulation, and on-site detection. As a proof of concept, multiscale sample analysis ranging from nanoscale nucleic acids to microscale bacteria and 3D cell spheroids is realized in the CASE. In summary, this platform offers valuable guidance for customized droplet generation by centrifugal step emulsifiers and promotes the adoption of droplet microfluidics in biochemical assays.