EDWARD MWANGI KIMANI (UT/WETSUS)
Supervisors: Walter van der Meer (UT, promotor), Antoine Kemperman (UT), Maarten Biesheuvel (Wetsus)
Funded by Wetsus/ H2020 / Marie Skłodowska-Curie Actions COFUND and Wetsus
Reverse Osmosis (RO) is a membrane-based technology for the desalination of seawater and brackish water. RO makes use of a pressure difference to push water through the ultra-thin polyamide “top-layer” ( 200 nm thick) of a thin-film composite RO membrane, while retaining the ions. To develop an optimal RO-processes, the availability of accurate theoretical models to describe energy consumption and the composition of the product water is essential. These models help in understanding the transport and separation processes better, the development of better membranes and module designs and help to design a better process layout for a given objective. Accurate models help to predict the exact retention of key undesired components.
Mass transport, modelling, multicomponent, desalination, reverse osmosis
However, precise models based on a detailed physico-chemical description of the processes that take place in and around the membrane are not yet developed to make the desired design possible. Though there exists prediction software that describes various aspects of the performance of RO membranes, this software uses a simple algorithm with neglects the various intricate couplings between the transport fluxes and the various ions when the feed water contains a large number of species (multicomponent mixtures). These couplings originate from the constraints of zero total current and local electroneutrality at each position in the membrane, while ions also interact via acid-base reactions. Therefore, this software is unable to accurately predict the retention of ions in a multicomponent mixture, certainly when many ions undergo acid-base reactions, such as ammonia (neutral) and ammonium (a cation), with the involvement of a hydronium ion.
This research therefore aims to develop and test a continuum transport model for RO which is based on physical principles and chemical information, and therefore includes all known interactions between all ions and water and membrane, which results in more robust predictions outside the (necessarily limited) window of experimental validation. The development and extension of the transport model which is based on the “extended Donnan Steric Partitioning Pore Model” (ext-DSPM) will be done to the level of describing a full module (two dimensions) and a small RO plant. A physico-chemical based model helps in a better understanding of the mechanisms involved in the transport and separation (e.g., what driving forces are the most important; what are concentration profiles of ions in the membrane, are ion-ion frictions in the membrane of importance?).
Develop a mass transport modelling approach based on the ext-DSPM to include all relevant chemical effects of ion-ion interactions, acid-base equilibria, membrane charge ionization, and electroneutrality and Donnan layers, in a multicomponent mixture.
Validate and upgrade the model with experiments.
Extend the model to a full RO module description (two dimensions).
Determine optimal RO operational conditions.