Direct-Contact Membrane Distillation: Simplified Flux Prediction, Mass Transfer Mechanisms, and Membrane Cleaning
AdvisorChildress, Amy E
Civil and Environmental Engineering
AltmetricsView Usage Statistics
Membrane distillation (MD) is a thermally-driven membrane separation process that uses the temperature difference at the membrane surfaces as the driving force to separate contaminants from potable water. The capability of MD to be combined with low-grade thermal energies to generate temperature gradients makes MD an attractive water treatment technology compared to the pressure-driven membrane processes, which utilize electricity as the power source. MD has versatile applications because it can achieve near 100% removal of salt and organic matter and the driving force of MD does not fluctuate significantly with variations in feed-water salinity.Low water flux of MD is desirable when treating feed waters with high fouling potentials, while high water flux of MD is desirable when treating feed waters with low fouling potentials. Therefore, the ability to select a MD membrane with proper (high or low) water flux is useful. Although several mass transfer models are available for flux prediction of MD membranes, the models are generally cumbersome and require information related to membrane properties. To overcome these issues, a simplified flux prediction model for direct-contact MD (DCMD) was developed using experimental flux and 28 structural parameters derived from physical properties of ten single-layer MD membranes. The membrane water fluxes were determined at the same experimental conditions, and the membrane properties included average pore size, porosity, tortuosity, thickness, liquid entry pressure, and contact angle. Additionally, an innovative membrane structural parameter, the membrane constant (Cm), that contains non-coupled membrane properties while still carrying the physical meaning of a relationship between thickness and porosity was developed. The correlation between water flux and Cm suggested that Cm is a good structural parameter for the prediction of MD flux. The flux prediction errors for membranes with pore sizes from 0.1 to 0.9 m were generally smaller for the model developed with Cm than for the commonly used dusty gas model.Because high water flux of MD is desirable when treating feed waters with low fouling potentials, the fundamental principles of mass transfer mechanisms in MD were investigated to provide insight into methods for flux improvement. The performances of three DCMD systems were investigated: 1) traditional DCMD, 2) pressure-enhanced DCMD (PEDCMD), and 3) vacuum-enhanced DCMD (VEDCMD). VEDCMD was found experimentally to have the highest water flux, followed by DCMD and PEDCMD, which had similar water fluxes. The main factors leading to the higher water flux of VEDCMD were membrane compaction and the air pressure inside the membrane pores. The pressure gradient across the membrane was also found to have minimal effect on water flux.Membrane fouling has been recognized as one of the main obstacles inhibiting the full-scale implementation of DCMD, and can be an issue when treating feed waters with both high and low fouling potentials. The efficiency of several membrane cleaning solutions was investigated for removal of commonly observed scalants (e.g., CaCO3, CaSO4, SiO2, and NaCl) in DCMD. Experimental results suggested that a citric acid solution could effectively remove CaCO3 scalant, while de-ionized water alone could effectively remove CaSO4 scalant. SiO2 scaling was more difficult to remove, and a two-stage cleaning procedure using NaOH solution at 40 oC followed by Na2EDTA solution was necessary to fully clean SiO2 scaled membranes. An HCl solution was found to fully remove the scalants (mainly NaCl) from a hypersaline solution, however membrane wetting occurred after membrane cleaning leading to incomplete restoration of membrane performance. By drying the membranes after HCl cleaning, more than 90% restoration of the maximum water flux and batch recovery was achieved. The identification of effective membrane cleaning solutions in this study is applicable to the full-scale implementation of MD in the future.The development of the simplified flux prediction model will make a contribution for MD membrane selection since membranes with different fluxes are suitable for specific applications. Besides membrane properties, experimental conditions also affected water flux; therefore, the investigation of the fundamental principles of mass transfer mechanisms in MD has helped to clarify confusions and contradictions about factors affecting water flux in the literature and direct the way for flux improvement in future. The identification of the effective membrane cleaning solutions for typical scales removal in MD will improve the full-scale implementation of MD since membrane scaling is one of the crucial problems affecting water flux.