Recent work has predicted that Na-ion batteries with symmetric electrodes can be used for desalination with low energy consumption and high desalination degree from brackish to seawater salinity level.¹ This technology, referred to as cation intercalation desalination (CID), shows much promise, but its wide-scale application will be dependent on capital cost. Similar to other electrochemical desalination technologies,² initial embodiments of the CID concept used ion-exchange membranes (IEMs) to separate desalinated and concentrated streams in electrodes.^1,3,4 IEMs in capacitive deionization cells can consume 50% of their expense, and IEMs have a finite lifespan.⁵ One solution to these challenges is to replace IEMs with chemically inert separators. Since the 1920s, non-selective diaphragms have been used in brine electrolysis. Furthermore, environmental-friendly diaphragms have shown the property of limiting ionic transfer between electrodes.

Building on these concepts, we consider a novel CID cell design that omits IEMs by the use porous, fluid-blocking diaphragms as economical and stable alternatives to IEMs. In this study we quantify how the loss of selectivity incurred by using a diaphragm impacts the resultant salt removal and energy consumption levels. We consider two different flow arrangements for these devices, with or without electrolyte recycling, among which we compare energy consumption and salt removal. We also explore the effect of the operating conditions using a reduced-order analytical model. With our analytical model we predict salt removal efficiency (SRE, the ratio of the amount of salt removed relative to that of a perfect diaphragm) and degree of desalination (DoD, defined as the fraction influent concentration removal as a percentage) as a function of two dimensionless numbers: (1) the Peclet number Pe (the ratio of advective mass transfer rate over diffusive mass transfer rate) and (2) the Damköhler number Da (the ratio of reaction rate to the diffusive mass transfer rate)_. Figures 1a and 1b show contours of SRE in the space of Pe and Da for the non-recycling and recycling cell, respectively. Based on the definition of Pe and Da, these results reveal the influence of flow rate and current density on SRE, thus, operating conditions can be selected to reach a specific efficiency.

Using the optimized operating conditions determined from the analytical model, we simulate the two-dimensional charge transport processes within porous electrodes and the diaphragm as source water is desalinated. The two-dimensional model employed here is based on porous electrode theory. To produce 60% DoD and 90% SRE from influent with 500 mM NaCl we use the conditions marked by α and β in Figs. 1a and 1b, respectively, for our two-dimensional simulations: Da=6.31 and Pe=109 for the recycling cell and Da=6.31 and Pe=6.31 for the non-recycling cell. Simulations with these operating conditions and 90% SRE reveals that the diaphragm cell can desalinate 500 mol/m³ solution to 260 mol/m³ (with recycling) and 244 mol/m³ (without recycling) solution (see Figs. 1c and 1d). For these two cases both diaphragm cells consume approximately 4.7 kWh energy when 1 m³ of desalinated water is produced. Figure 1c shows effluent salt concentration as a function of time for the non-recycling cell. Effluent salinity within the negative electrode decreases to a steady value after the process starts For the recycling cell effluent salinity continuously decreases during the charging processes (Fig. 1d). Salt concentration is distributed differently within each cell depending on whether or not effluent recycling is used. This effect is shown in Fig. 1e, where the concentration distribution in the recycling cell (lower) is more uniform than that in the non-recycling cell (upper), because a faster flow speed is used in the recycling cell. The simulated DoD (50%) agrees well with the predicted 60% DoD, and hence the analytical model can be used as a tool for performance predictions. The CID cells simulated here using diaphragms demonstrate the potential to replace IEMs with diaphragms. These findings motivate the future application of membrane-less CID technology in ion separation and/or removal processes.

Acknowledgements: We are grateful for funding and support from the College of Engineering and the Mechanical Science and Engineering Department at UIUC.

References

1.) K. C. Smith and R. Dmello, J. Electrochem. Soc.,163, 2016

2.) L. F. Greenlee, D. F. Lawler, B. D. Freeman, B. Marrot, P. Moulin, Water Res., 2317, 2009

3.) K. C. Smith, Electrochim. Acta, 230, 2017

4.)S. Porada, A. Shrivastava, P. Bukowska, P. M. Biesheuvel, K. C. Smith, Electrochim. Acta, 2017

5.) HJ Lee, F Sarfert, H Strathmann, SH Moon, Desalination, 267, 2002