Tailoring nonsolvent-thermally induced phase separation (N-TIPS) effect using triple spinneret to fabricate high performance PVDF hollow fiber membranes

In the field of membrane technology, the combined nonsolvent-thermally induced phase separation (N-TIPS) method is a versatile technique to prepare macroporous membranes with narrow pore size distribution. However, its full potential in hollow fiber preparation has been hindered due to the formation of a dense skin layer via rapid mass exchange at the solvent-nonsolvent interface. In this study, a simple and effective solution is proposed to eliminate the dense skin layer and prepare a highly porous, macroporous membrane. We employed a triple spinneret to apply a transient coating layer to prevent the polymer dope solution from directly contacting the nonsolvent, which is regarded as the main cause of dense skin layer formation. Interestingly, depending on the choice of transient coating fluid, the fluid acts as a selective layer to facilitate unidirectional diffusion of the solvent into the nonsolvent while blocking the nonsolvent inflow. By controlling the solvent-nonsolvent exchange and thermal gradient in tandem, we show that a convenient control over NIPS and TIPS kinetics can be achieved. Several solvents, ranging from conventional solvents such as dibutyl phthalate (DBP) and N-methyl-2- pyrrolidone (NMP) to environmentally-friendly solvents such as acetyl tributyl citrate (ATBC) and methyl-5- (dimethylamino)-2-methyl-5-oxopentanoate (PolarClean® ), were employed as the transient coating layer to investigate the effects of different solvents on the formation of macroporous hollow fiber membranes. It was observed that as the hydrophobicity of the coating layer increased, the prepared membrane exhibited a more porous morphology with a larger pore size. Moreover, the tensile strength of the prepared fibers improved from 3.4 ± 0.1 MPa to 7.7 ± 0.3 MPa, the pure water permeability increased from 1 to 988 L m-2 h-1 bar-1 , and the membrane exhibited a narrow pore size distribution.