Speaker
Description
Lipid nanoparticles (LNPs) are central to RNA- and gene-delivery therapeutics, and are increasingly produced using microfluidic mixers. However, LNP sterile filtration during downstream processing remains inherently challenging because LNP dimensions are comparable to the pore size of sterilizing-grade membranes (100 - 200 nm). In this regime, LNP passage and retention are governed by particle softness, deformability, and population heterogeneity leading to non-classical filtration that is not adequately captured by rigid-particle models. Conventional bulk characterization techniques, such as dynamic light scattering (DLS), provide limited insight into how population heterogeneity governs early membrane interactions and fouling behavior.
Here, we present a microfluidic framework that combines (i) controlled LNP synthesis, (ii) microfluidically enforced constant-flux filtration using a commercial sterilizing-grade syringe filter, and (iii) population-resolved LNP characterization by asymmetrical flow field-flow fractionation with UV and multiangle light scattering (AF4-UV-MALS). LNPs were produced by mixing a lipid-in-ethanol stream with an acidic aqueous stream in a microfluidic staggered herringbone mixer. AF4-UV-MALS analysis of LNPs formed under identical flow conditions revealed distinct population structures depending on cargo loading: unloaded LNPs form a predominantly single size population, whereas PolyA-loaded formulations exhibit an additional subpopulation of larger particles.
Microfluidic, pressure-driven flow control enabled precise low-flux operation (<50 LMH) and real-time transmembrane pressure (TMP) monitoring. This low flux operation creates a transport-controlled regime in which early-stage particle-membrane interactions can be probed before flux-induced cake compaction. In this way, a conventional sterilizing-grade filter is effectively turned into a microfluidic testbed for soft-particle filtration.
Under identical constant-flux conditions, PolyA-loaded LNPs exhibited a systematically stronger TMP increase than unloaded LNPs, indicating higher hydraulic resistance that cannot be explained by bulk-averaged size metrics alone. AF4 -UV-MALS analysis reveals that the minor, loading-induced large-particle subpopulation disproportionately contributes to an increased filtration resistance.
By directly linking microfluidically controlled filtration dynamics to population-resolved particle structure, this work shows that particle heterogeneity, rather than mean particle size, governs LNP filtration behavior. More broadly, it demonstrates how microfluidic flow control can be coupled to macroscale filter geometries to isolate and quantify early-stage, population-driven fouling phenomena. The presented approach offers a general microfluidic framework for mechanistic analysis of nanoparticle–membrane interactions, with implications for the design of microfluidic filtration modules and for the integration of sterile filtration in microfluidic LNP production and downstream processing.