Theoretical Models for Membrane Capacitive Deionization for the design of Modular Desalination Processes

Abstract

Due to climate change, water scarcity will be exacerbated around the globe. To increase the water availability in regions at risk, water desalination plants can be a solution. Especially in rural areas, energy e cient technologies are needed so that an operation with renewable energy as photovoltaic modules can be feasible. Recent publications showed that the novel technology membrane capacitive deionization (MCDI) can achieve a lower speci c energy consumption (SEC) than reverse osmosis (RO), for brackish water desalination with salt concentrations below 2.5 g L-1. There is still a gap in research between laboratory operation and applied commercial scaled desalination, regarding experimental but also theoretical model studies. Therefore the latter is elaborated in the present PhD thesis. Hereby, existing models are reviewed, adapted and further developed to t to applied MCDI operation for drinking water production. Two dimensional nite element methods (FEM) modelling of ion transport, according to the Gouy-Chapman-Stern theory for electrical double layers (EDL) as well as computational uid dynamics (CFD) is combined with an adjusted semi-analytical modi ed Donnan (mD) model, with a constant excess chemical potential att = 2:33 kT, for the electrosorption of ions into porous active carbon electrodes. It predicts the e uent salt concentration time-dependently for di erent inputs of applied electrical currents Icell and voltages as well as inlet concentrations and volume ows. Applied MCDI operation was optimized for drinking water production with practical experiments, which support the evaluation of the theoretical ndings. The model ts to experimental data for Icell = 20 A, however the equations for the voltage over the electrodes need to be re-assessed so that the model ts for further input parameters. A CFD model of the water ow through large scaled MCDI modules (> 50 pairs of electodes) shows the need of constructing spacer thicknesses Sp small enough, to ensure equal retention times of the water between the electrodes in the module, which is important for stable diluate concentrations. Furthermore, an analytical calculation tool is developed, by adjusting the mD model and introducing an e ective salt adsorption capacity ð€€€salt; , to predict the maximum e cient charging time tmax,ch, removal- and recovery rate as well as SEC values for optimized operation of applied MCDI processes. The model reaches an accuracy of 87% for the prediction of salt removal, 86% for tmax,ch and 75% for SEC values, compared with an experimental study and thus can be used to optimize the process design of applied MCDI desalination plants.