Experimental study of decontamination factor in pool scrubbing with two-phase flow evolution

Pool scrubbing is a key mitigation process in nuclear severe accidents (SAs), in which aerosols and gaseous fission products (FPs) are injected into water and removed through gas–liquid interactions. Although the decontamination factor (DF) depends strongly on particle size and hydrodynamic conditions, systematic DF datasets obtained under high-flow, non-condensing conditions—together with corresponding detailed two-phase flow-field measurements—remain scarce. This study addresses this gap by providing a high-accuracy, size-resolved DF dataset consistent with independently measured flow-field data. A 0.5-m-diameter, 8-m-high test facility was used to reproduce the characteristic severe-accident flow evolution from large globules to bubble breakup and the formation of a fine-bubble swarm. Using Welas, SMPS, and ELPI+, size-resolved DF values for soluble CsI aerosols were obtained over a wide aerodynamic-diameter range (sub-0.05 μm to 10 μm). The mixture water level was varied continuously from 0 to approximately 5 m, enabling DF measurements across evolving flow structures. The DF exhibited clear dependencies on both particle size and water level, and the transition from the injection/breakup region to the swarm region was directly reflected in the DF behavior. These trends were consistent with detailed axial bubble-size evolution measured previously in the same facility. An empirical DF correlation was developed as a function of mixture water level and aerodynamic diameter. Comparison of this correlation with the DF measurements, flow-field data, and the mechanistic MELCOR/SPARC90 model showed that, while the major hydrodynamic transitions were consistent with the model, the experimentally observed particle-size dependence and gas-flow-rate independence revealed characteristics associated with strongly turbulent, multidimensional bubble motion in the swarm region. Furthermore, SMPS measurements demonstrated a pronounced DF increase for ultrafine particles (<0.05 μm) due to Brownian diffusion—providing new experimental evidence in a particle-size range for which reliable DF data have been largely unavailable. A comparison with insoluble BaSO₄ aerosols generated under identical conditions showed consistently lower DF values for BaSO₄ despite nearly identical particle density, indicating that aerosol solubility has a significant influence on removal efficiency. The systematic DF dataset obtained in this study elucidates the governing mechanisms of particle-size-dependent aerosol removal under realistic pool-scrubbing flow regimes. Combined with detailed flow-field measurements from the same facility, these results provide a robust benchmark for improving mechanistic DF models, particularly under high-flow conditions relevant to severe-accident source-term assessments. • The decontamination factor (DF) of CsI aerosol was measured under varying water levels. • Three instruments (Welas, ELPI+, SMPS) were used to obtain DF across a wide particle size range. • DF increased exponentially with mixture water level, showing strong particle-size dependency. • A new empirical model was proposed for DF in the sub-0.1 μm range considering Brownian diffusion. • The combined model enables DF prediction from 0.03 to 10 μm across practical scrubbing conditions.