Scaling Of Calcium Sulfate In Membrane Distillation For Water Treatment

Membrane separations have been playing a potential role in addressing the global issue of water scarcity.In comparison with conventional membrane processes,membrane distillation（MD）is emerging as a powerful tool for dealing with high salinity water or wastewater.However,MD-based applications are suffering from fouling phenomena that result in flux decline and other negative effects to significantly decrease the efficiency of MD in various ways.As a special fouling phenomenon,scaling of sparingly soluble salts is dominant when the feed can be highly concentrated by an MD process.In particular,the scaling of calcium sulfate（Ca SO₄）is recalcitrant and deserves particular attention in the context of elucidating the underlying mechanisms,which entails novel characterization techniques to resolve the formation and evolution of a scaling layer on the hydrophobic membrane surface.Therefore,this study was aimed at investigating the scaling behavior of Ca SO₄ in MD by exploring a series of characterization methods that were based on optical coherence tomography（OCT）.The developed methods were further employed to investigate the scaling-induced flux decline and wetting phenomenon.Moreover,the evolution of a scaling layer in MD was analyzed by comparing the roles of bulk precipitation and surface growth.Taking advantage of the ability to optically section a semitransparent medium,an OCT system was integrated with a setup of direct contact membrane distillation（DCMD）in an effort to in-situ observe and analyze the scaling layer of Ca SO₄ gradually developed at the feed-membrane interface.In addition to creating tomographic images,a variety of numerical algorithms were exploited to analyze the OCT datasets.The numerical analysis was based on a coordinate system whose zeroth coordinate surface was determined by the feed-membrane interface.On the one hand,the scaling-induced variations in the intensity were evaluated on each of the coordinate surfaces to create profiles of surface-averaged intensity（SAI）and fraction of positive anomalies（FPAs）,which offered a tool for examining the scaling process in a statistical sense.On the other hand,the feed-membrane interface was accurately identified such that the scaling layer could be digitalized for numerically evaluating the surface coverage,the mean of local thicknesses,and the local growth rates.When mapping the local growth rates,it was revealed that the boundary layer could be hydrodynamically instable owing to the coupled heat and mass transfer,thereby giving rise to the striping phenomenon.The OCT-based characterization was first combined with a mathematical model to identify the relative importance of various mechanisms accounting for the scaling-induced flux decline in an MD process.When determining the cake-dominated regime（i.e.,the period when the cake thickness was uniformly increased）,the modeling and experimental results were compared to assess contributions of the different mechanisms.It was inferred by this comparison that densification at the cake-membrane interface could play a key role in substantially reducing the vapor flux.In order to provide deeper insights into the scaling-induced wetting phenomenon,the crystal-membrane interaction was resolved by numerically tracking the shift of the membrane surface in the framework of Stoney’s equation.It was indicated that the interfacial membrane structures could be stretched owing to the initial scaling,whereas the crystal-containing layer would be dominated by compressive stress as the surface coverage was substantially increased.The analysis also established a strong correlation between the occurrence of wetting and a relatively high rate of concentrating the feed,which could be essential for maintaining continuous growth of the confined crystals and creating sufficiently large crystallization pressure to irreversibly expand the membrane pores.Moreover,the morphological evolution of a scaling layer is a more fundamental concern to be addressed for better understanding the scaling in MD.The OCT-based characterization and conventional measurements were implemented in a synergistical way to compare the effects of bulk precipitation and surface growth on developing the scaling layer in MD processes with a varied transmembrane temperature.When determining the threshold of bulk crystallization,the different effects were decoupled to provide evidence for a mechanistic transition.This study not only confirmed the potential of OCT for in-situ characterizing the scaling phenomenon in MD,but also demonstrated the diversity of approaches that can be used to numerically analyze the OCT datasets.In addition,this study highlighted the advantages of implementing modeling-based analysis and conventional measurements in a complementary way to cover the downsides of the OCT-based characterization.All the characterization results in this study refined the mechanistic picture of scaling in MD and would shed light on the development of MD-based applications with enhanced performance for desalination and water/wastewater treatment.