**Abstract:** Hierarchical zeolites, exhibiting interconnected micropores, mesopores, and macropores, are crucial materials for advanced separation and catalytic processes, particularly in multi-phase flow applications. Accurate prediction of multi-phase flow behavior within these complex structures remains a significant challenge. This paper introduces a novel, data-driven approach leveraging advanced pore-network modeling (PNM) combined with machine learning algorithms. By integrating high-resolution experimental data with dynamic PNM simulations, we develop a predictive framework capable of accurately forecasting multi-phase flow performance in hierarchical zeolites, offering significant advancements over deterministic PNM models.

Based on the synergistic effects of low-temperature microemulsions and Hofmeister ions, hierarchically mesoporous MOFs (HMMOFs) with adjustable mesopore sizes are successfully achieved, which are competent to detect femtomolar level alkaline phosphatase (ALP). Such high sensitivity for ALP detection highlights the importance of adjustable mesopore structures of HMMOFs in advanced sensing applications, and prefigures their potential for biomedical diagnostics. Abstract In addressing the demand for hierarchically mesoporous metal-organic frameworks (HMMOFs) with adjustable large mesopores, a method based on the synergistic effects of low-temperature microemulsions and Hofmeister ions is developed. Low temperature dramatically enhanced the solubility of hydrophobic solvent in the microemulsion core, enlarging the mesopores in HMMOFs replica. Meanwhile, Hofmeister salt-in ions continuously controlled mesopore expansion by modulating the permeability of swelling agent into the microemulsion core. The large mesopores up to 33 nm provided sufficient space for the alkaline phosphatase (ALP) enrichment, and retained the remaining channel to facilitate the free mass diffusion. Leveraging these advantages, a colorimetric sensor is successfully developed using large-mesopore HMMOFs for femtomolar ALP detection based on the enrichment and cycling amplification principles. The sensor exhibited a linear detection range of 100 to 7500 f m and a limit of detection of 42 f m , presenting over 4000 times higher sensitivity than classic para-nitrophenyl phosphate colorimetric methods. Such high sensitivity highlights the importance of adjustable mesoporous structures of HMMOFs in advanced sensing applications, and prefigures their potential for detecting large biomolecules in diagnostics and biomedical research.