(643b) Optimizing Fabrication of Polysulfone Ultrafiltration Membranes Using a Green Solvent, Cyrene | AIChE

(643b) Optimizing Fabrication of Polysulfone Ultrafiltration Membranes Using a Green Solvent, Cyrene

Authors 

Hackett, C. - Presenter, University of Arkansas
Thompson, A., Prairie View A&M University
Wickramasinghe, R., University of Arkansas
Qian, X., University of Arkansas-Fayetteville
Hale, D., University of Notre Dame
Even though membranes can lead to more environmentally sustainable separation processes, membrane casting involves the use of toxic organic solvents. The demand for filtration membranes made with nontoxic green solvents has increased over the past several years. Cyrene (dihydrolevoglucosenone) is a solvent that can be used for fabricating membranes by phase inversion, and it is less toxic than traditional solvents such as n-methylpyrrolidone (NMP). Nearly all previous studies that used Cyrene for fabricating membranes used water as the nonsolvent in the coagulation bath. However, other nonsolvents, such as ethanol, have been investigated in conjunction with traditional solvents and sometimes led to enhancements in membrane performance. In this study, polysulfone ultrafiltration membranes were fabricated by nonsolvent-induced phase separation (NIPS), using Cyrene as the solvent and an ethanol/water mixture as the nonsolvent coagulation bath. Membranes were optimized by adjusting the ethanol content of the coagulation bath. Performance was tested by dead-end filtration. Membranes cast from Cyrene solution into pure water had low pure water flux (40.7 L/m2/h) and bovine serum albumin (BSA) rejection (18.1%). When 70 v% ethanol was used in the coagulation bath, pure water flux increased to 106.4 L/m2/h and BSA rejection increased to 84.7%. Membranes were characterized to better understand the performance increase. Fourier-transform infrared spectroscopy (FTIR) analysis of membrane samples showed that membranes cast into pure water retained significant residual Cyrene even after 24 hours of storage in water. The inability of pure water to adequately remove Cyrene led to diminished membrane performance. However, precipitation baths with 25 v% or more ethanol fully removed Cyrene from the membranes. The closer hydrogen bonding solubility parameters of ethanol and Cyrene (vs. water and Cyrene) contributed to the improved Cyrene removal, as ethanol and Cyrene have better miscibility. Scanning electron microscopy (SEM) showed that membranes cast from Cyrene solution into 70 v% ethanol solution had a flatter and more porous surface compared to membranes cast into pure water, which had a crumpled surface morphology. All membranes cast from Cyrene solution had sponge-like pores, indicative of delayed demixing. Dissipative particle dynamics simulations are being used to further elucidate how differences in solvent and coagulation bath composition affect pore formation. Understanding the relationship between coagulation bath composition and performance parameters can lead to the determination of an optimal membrane fabrication protocol. The result of this study is a pathway for fabricating polysulfone ultrafiltration membranes with adequate performance, reduced toxicity, and guidance for further optimization.