Book Description

Removal and recovery of volatile organic compounds (VOCs) from industrial wastewater and groundwater has become increasingly important due to stringent environmental regulations. Membrane air-stripping (MAS), using microporous polypropylene hollow fiber membrane modules, is one of the most promising processes for this purpose. The mass transfer of water and VOCs in MAS was studied using such a module, with air-flow on the lumen side and liquid cross-flow on the shell side. Chloroform, toluene and their mixture were used as model VOCs. Water transport experiments showed that mass transport was significantly decreased when the membrane had been in contact with water for prolonged periods. It was hypothesized the increased mass transfer resistance was due to water condensation in a fraction of the membrane pores. MAS of chloroform from aqueous solutions confirmed the additional mass transfer resistance with prior exposure to water. It was concluded that membrane pores were completely air-filled at the start and became partially wetted with water after prolonged period during the MAS process. The currently existing models are able to predict the performance only for either completely air-filled or liquid-filled pores. A modification of an existing model was proposed to take into account diffusion through the partially water-filled pores, as well as the partially air-filled pores. It was found that the model predictions agreed well with the experimental data. This hypothesis also provided a plausible explanation for the conflicting literature values of the membrane mass transfer resistance. It was also found that the membrane mass transfer resistance of the partially water-filled pores was two orders of magnitude higher than that of air-filled pores. Leveque's (1928) correlation overestimates the local mass transfer coefficient in a cylindrical tube at low velocities. A modification of this correlation has been proposed to predict the local air film mass transfer coefficient at low air velocities. The proposed correlation predictions matches well with the experimental data. The overall mass transfer coefficients of chloroform obtained in this work for liquid cross-flow on the shell side were up to twice as high as those reported in the literature, even though our experiments were carried out at much lower water and air velocities. However, the air pressure drop on the lumen side was significantly higher than that for system with air flow on the shell side. The overall mass transfer coefficients did not change when the initial chloroform concentration in the feed ranged from 81 to 908 ppm. MAS process was found effective in concentrating chloroform to more than 90% from a feed aqueous solution of ppm levels. The adsorption of toluene had strong detrimental impact on the performance of the polypropylene hollow fiber module. It is hypothesized that the toluene sorption resulted in swelling of the polypropylene fibers causing a reduction of the effective pore diameter and as a result of this, the toluene transport was substantially lower than expected. Due to this effect, the presence of toluene in the binary aqueous solution with chloroform significantly reduced the mass transport of chloroform compared to that with only chloroform. Henry's law constants were determined for individual chloroform and toluene as well as for their mixtures at 23°C and are reported. The effect of initial chloroform concentrations on Henry's law constant was experimentally examined.