Membrane capacitive deionization model including fouling indexes obtained via real-time fouling layer measurements

Abstract

For the application of Membrane capacitive deionization (MCDI), it is essential to investigate how MCDI performance deteriorates owing to natural organic matter (NOM) fouling during operation. However, the difficulty in obtaining real-time fouling layer information has hindered this understanding. Herein, we investigated the effect of fouling with a model using humic acid (HA) as NOM with two background electrolytes (NaCl and CaCl2) in different ratios. To quantify the effect of the fouling, fouling thickness was monitored in real-time through optical coherence tomography. We found that feed water containing more Ca2+ quickly formed more neutrally charged complexes with HA, which increased the rate of fouling thickness and significantly decreased the salt removal rate. The reduction in salt adsorption capacity exhibited high correlation with the observed fouling thickness (≥0.89). Using the experimental values of fouling thickness and newly added parameters that specifically considered fouling, the MCDI transport model was revised to simulate the effluent salt concentration under HA fouling conditions. The modified model showed superior prediction performance than the original model: R2 improved from 0.728 to 0.977 when the ratio of Na+ to Ca2+ was 3:7. Therefore, the developed MCDI transport model will pave the way for the practical application of MCDI.