Silicon nanoporous membranes (SNMs) are suitable for hemofiltration and dialysis applications. These 15 nm thick membranes, with a distribution of nanometer sized pores (average 8 nm), have been used for molecular diffusion which depends on their size and surface charge. In the present study, SNMs are fabricated using MEMS based techniques with reduced pore diameter (5.7 nm) for enzyme purification application. This membrane based enzyme purification is more scalable in volume and clearance rates as compared to conventional column chromatography methods, where enzymes traverse their path through a standing column filled with porous agarose beads.

A stack of 3 layers (SiOx: a-Si: SiOx) is deposited (by chemical vapour deposition) and subjected to rapid thermal annealing at a temperature of 950 ℃ to form pores. Pores were characterised using the Transmission Electron Microscope image of the membrane, which shows a distribution of pores ranging from 4 nm to 8.7 nm diameter, peaking at 5 nm. Present study looks at the filtration of an enzyme cocktail mixture through these SNMs. The in-house made teflon filtration setup contains two chambers - cis and trans (each 50 mL), and the 3x3 SNM array was held in the middle with the help of O-rings. Laminar flow path is incorporated inside each chamber in the counter directions to enhance the diffusion. Fermentation enzymes mixture (having different molecular weights) was filled in cis side and was found to retain molecules of size 8-10 nm and higher, after 3 hours of filtration. The overall retention of enzymes were 75% (580 µg mL-1 out of 800 µg mL-1) of the total enzyme mixture. In order to understand the mass transportation through these nanoporous domains, diffusivity values were extracted using simulations by matching the previously reported molecular separation results, to study the influence of the factors like analyte’s molecular weights and surface charge in diffusion through ultrathin membranes. Extracted diffusivity is then used to describe the analyte separation through SNMs. Diffusion coefficients were found to follow a power law relationship with molecular weights of analytes. Diffusivity of analytes in nanoporous medium seems to be a much stronger function of molecular weights (order of 4) as compared to that in aqueous medium (order of 0.33). Also, analytes with lower surface charge are shown to possess larger diffusivity values. The SNMs were used for real-time application for hemodialysis for the clearance of uremic toxins. The application was extended by integration of sensors critical for hemodialysis. These sensors were responsible for monitoring parameters like temperature, presence of bubbles in the tubings and the transmembrane pressure - which are essential during the dialysis process. Parameters were processed on a microcontroller and displayed on LCD.
Overall this work presents potential use of SNM for both Industrial and biomedical application.