Water. Desalination + reuse
Issue link: https://fhpublishing.uberflip.com/i/1056083
December 2018 Water.desalination+reuse Far Site 33 Berkeley seeks to predict membrane performance The project is working out how to computationally describe transport of water on and through membranes, and nucleation of crystals We have very little predictive ability for scaling in RO membranes, organic fouling in UF, or biofouling. At Lawrence Berkeley Na- tional Laboratory in the US, an intriguing project to model the flow of water on and through water purification membranes is underway. The project is col- laborating with scientists in the geosciences field at the Lab. The geoscientists have developed so•ware that enables them to computationally describe the flow of water through geo - logical media—porous rocks or other sedimentary media un- derground. The new project is applying that modelling to the flow of water on and through membranes. "The geoscientists have over many years been looking at the flow of water underground, water that is contaminated with salts and organics. The idea now is to apply that thinking to describe the transport of water in membranes, and also some of the fouling phenomena, the crystallisation phenomena," explains Daniel Miller, staff scientist, chemical sciences division, Lawrence Berkeley National Laboratory. "The same nucleation of crystals that occur on membranes, the same sorts of phenomena occur under - ground with these salty waters that the geoscientist have been looking at." Miller, whose research interests include the structures and property relationships that govern the transport of small solutes through polymeric membranes, believes that the project could lead to cost-saving predictive capabilities and to more rational membrane system design. "We have very little predic - tive ability for scaling in reverse osmosis (RO) membranes, organic fouling in ultrafiltra- tion membranes, or biofoul- ing," says Miller. "If we have some contaminated water and we want to use a membrane to purify it, we have no way of predicting if fouling will be a significant concern. We have no way of predicting which membrane we should use. There is no way to know what pre-treatment may work best, or what the permeate will be like." At the moment, plant opera - tors draw on their experience and empirical knowledge when installing a new membrane system. The industry is relying on "very educated guesswork", Miller says. Modelling crystal growth The computational modelling could result in a predictive ca- pability that provides insight up front about which membrane to use, how to operate it, and how effectively it's likely to be. The first step — which is what the team is currently focusing on — is to model what's going on at the membrane surface, including scaling, organic foul - ing, and biofouling. "For about a year, we have been using that work done by the geoscientists and working out how to apply it to describe the flow of water in and on the surface of the membranes. We are looking at flow of water through the channels in the membrane," says Miller. "We have started looking at scaling in an RO system, and we intend to push that also to other mem - brane types such as ultrafiltra- tion and microfiltration, and to other types of fouling such as organic fouling, or biological fouling, which is the growth of microorganisms into biofilms on the membrane surface." The geometries modelled inside the system are consist- ent with those used in a typical commercial RO module, and can include a feed-spacer, and other components and attributes found inside a real RO module. "We can right now reasonably accurately model the flow of salty water across the surface of the membrane. We can reasonably accurately model the flow of purified water through the membrane, and we are now beginning to look at the nucleation and growth of crystals on the membrane surface and within the bulk water above the membrane," Miller says. The ultimate project goals are to reduce the cost of operating a membrane-based water purification system, by providing a more quantitative understanding of how it works, or fails. "At the most basic, it's looking at feed-water and mem - brane characteristics, and op- erating conditions, and tuning those parameters to maximise the performance of the system: minimise energy use, mini - mise cleaning requirements, minimise downtime; and be able to choose the membrane best suited to that particular ap - plication," explains Miller. The so•ware will help to provide answers about what to do when something goes wrong, whether a fouled membrane, overly high energy use, not enough perme - ate, or if the permeate fails to meet quality control criteria. "Membrane installations could be much better tuned to the particular application that the customer has, so that this would allow for really rational design of membrane systems and ultimately perhaps, mem - brane materials, membrane modules," Miller says. Miller completed his PhD at Uni- versity of Texas, with a project de- veloping a hydrophobic coating for polymer membranes. The coating was the bio-inspired material poly- dopamine, originally found in the adhesive proteins of sessile marine organisms, such as mussels, which adhere to surfaces under water.