Heat-mass-momentum transfer in hollow fiber spinning
Abstract
The goal of this research is to develop the requisite fundamental knowledge necessary to tailor the production of hollow fiber membranes. Specifically, the focus of this work is a model describing the simultaneous heat, mass, and momentum transfer for hollow fiber membrane spinning. The model predicts radial concentration gradients and spinline dimensions as functions of axial position, as well as axial profiles of temperature, velocity, and core pressure that evolve during hollow fiber spinning. The modeling procedure requires little computational time and is applicable to hollow fiber membranes spun under various conditions. Sensitivity of model-predicted spinline variables to surface tension effects was explored in the thin filament limit. While viscous effects dominate surface tension effects for typical pure-polymer melt spinning, membrane spinning results show that surface tension effects alter the evolution of spinline variables during the process, which can affect spinning stability. Experimental diameter and axial velocity profiles obtained during spinning of hollow fiber membranes using a twin-screw extruder indicate that system viscosity can vary significantly due to diluent evaporation at the clad–air quench interface, which creates the concentration gradients modeled in this work. Experimental results show more rapid attenuation of the spinline than predicted by the model without accounting for the concentration dependence of viscosity. Incorporating the viscosity dependence on both concentration and temperature helps to resolve the discrepancy between modelpredicted and measured spinline diameter and velocity profiles. The sensitivity of hollow fiber membrane extent of anisotropy, the fraction of the fiber cross-section possessing a pore size gradient, to processing conditions and spinning system physical properties was examined. Results indicate that extent of anisotropy is sensitive to spinning temperature, core gas flow rate, air gap length, and diffusion coefficient, showing an increase in extent of anisotropy for an increase in these parameters. These results have important implications for membrane research, where development and optimization are largely trial-and-error approaches. This work is an important precursor to development of a complete model to predict membrane macrostructure (inner and outer diameters) and microstructure (pore size, pore size distribution, and anisotropy) as functions of spinning conditions and material properties.