Skin layer in cyclic loading-cleaning of a nanofiber filter in filtering nano-aerosols

Highlights • Cyclic loading/cleaning nanofiber filter with nano-aerosols. • Thin skin layer upstream of filter controls loading behavior, permeability drops to 10–20% of initial value. • Capillary model for aerosol deposit, pore bridging, cake model match 6-cycle loading tests. • Transit from depth-to-surface-filtration – tests and modelling.
Nano-aerosols from viruses to virgin pollutant particulates from combustion, 100 nm or smaller, are harmful to our health as they penetrate readily into our body causing various diseases. Nanofiber filter can capture effectively these nano-aerosols. However, over time the pressure drop increases dramatically and cleaning of the filter by backpulse/backblow is essential for filter reuse. The cyclic loading-and-cleaning of a nanofiber filter has been investigated for the first time experimentally and theoretically. The “skin” layer, a thin region upstream of the nanofiber filter, plays a pivoting role in controlling the pressure drop excursion of the filter. We model the skin layer to be made up of numerous fine capillaries and examine how continuous aerosols deposited in the capillaries affect rapid rise in pressure drop followed by bridging of aerosols across the capillary openings leading to more bridging and ultimately formation of cake on top of the bridges and filter surface. We have been able to describe the deposition of aerosols in the capillary pores for depth filtration, the deposition of aerosols in the cake (surface filtration), and the intermediate bridging regime between these two. We can depict the complete pressure drop excursion including the S-shaped curve behavior from depth filtration transiting to surface filtration for a filter with severe skin effect. Our prediction matches extremely well with the 6 cycles of loading/cleaning on a 280-nm nanofiber filter subject to challenging nano-aerosols, 50–400 nm. During cyclic loading and cleaning, the porosity and permeability in the skin layer for our experiment drop to 68% and 11–21% of their original values, respectively, and the effective pore diameter also drops from 1.2 to 0.6 μm.