(Invited) Using Intercalation Compounds for Electrochemical Water Desalination

Tuesday, 3 October 2017: 14:40
Chesapeake 12 (Gaylord National Resort and Convention Center)
K. C. Smith (University of Illinois at Urbana-Champaign)
Electrochemical desalination technologies offer benefits in both scalability and energy consumption relative to pressure-driven processes (e.g., reverse osmosis). Electrodialysis using gas-evolution electrodes that sandwich a series of ion-exchange membranes is the most used of such technologies to date, but alternative electrochemical technologies using electrosorptive interactions have gained significant research attention to date. For instance capacitive deionization (CDI) uses ion adsorption within solution-phase electric double layers to remove salt. Such technology has been applied readily in brackish water desalination. Alternatively, solid-state faradaic electrodes have also been used to remove salt from water through both intercalation and chemical conversion reactions. Specifically, the desalination battery used a Na-ion intercalation cathode paired with an Ag/AgCl conversion anode.1 Intercalation materials have also been used in conjunction with capacitive electrodes to produce a hybrid CDI configuration with enhanced salt adsorption capacity per unit mass of electrode material.2

Here we report on a desalination device in which cation intercalation host compounds (IHCs) of identical chemical composition are used in both electrodes, a concept originally referred to as Na-Ion Desalination (NID).3 While the use of the same IHC in both electrodes is impractical for battery use, we show through computational modeling3,4 and later confirm by experiments5 that desalination is possible with electrodes of identical composition. We show that this mode of desalination is most efficient if an anion exchange membrane (AEM) is used to suppress Na-ion transport between the electrodes, rather than an unselective porous separator used in Na-ion batteries.3 This concept, which we refer to presently as cation intercalation desalination (CID), can be employed with generic intercalation host compounds. In our original work we predicted the performance of CID using Na0.44MnO2 (NMO) and NaTi2(PO4)3 (NTP) intercalation electrodes, which exhibit sizable volumetric charge capacities and therefore produce substantial degree of desalination even for seawater level salt concentrations. Specifically, a Na0.44MnO2-based CID-cell with 0.5 mm-thick electrodes desalinated 700 mM NaCl influent by 63% while consuming only 50% more energy (0.74 kWh/m3) than thermodynamic minimum when cycled at C/2 rate. The high volumetric charge-capacity of Na-ion battery IHCs enabled 59–64% drop in influent salinity with water-recovery levels up to 80% and 95% for 700 mM and 70 mM influent, respectively.

Prussian Blue Analogues (PBAs) are an alternative class of intercalation host compounds that have shown facile intercalation and long cycle life in a variety of aqueous cation batteries. We have modeled a particular PBA, Na2NiFe(CN)6 (NiHCF), in CID operation and show that, despite its low charge capacity relative to NMO and NTP, efficient desalination of seawater-level concentrations is possible in a range of CID device configurations.4 In addition a variety of electrochemical cell architectures were explored ranging from the use of flow-through electrodes and extending to electrodialysis stacks using intercalation electrodes.4 We observe that the distribution of ionic current within flow channels is biased toward the inlet when concentrate and diluate streams flow in the same direction (i.e., parallel-flow configuration), but these effects can be mitigated by flowing concentrate and diluate streams in opposing directions (i.e., in counterflow configuration).

Finally, we report on experiments using porous electrodes containing redox-active NiHCF nanoparticles in a CID configuration.5 On the basis of active-material mass NiHCF-based CID cells are shown to achieve a salt adsorption capacity in excess of typical CDI cells. These results show potential for decreased energy consumption relative to CDI cells. Initial results with electrolyte influent containing mixtures of cations also reveal preferential adsorption of K+ over Na+.


1. M. Pasta, C. D. Wessells, Y. Cui, and F. La Mantia, Nano Lett., 12, 839–843 (2012).

2. J. Lee, S. Kim, C. Kim, and J. Yoon, Energy Env. Sci, 7, 3683–3689 (2014).

3. K. C. Smith and R. Dmello, J. Electrochem. Soc., 163, A530–A539 (2016).

4. K. C. Smith, Electrochimica Acta, 230, 333–341 (2017).

5. S. Porada, P. Bukowska, A. Shrivastava, P. M. Biesheuvel, and K. C. Smith, arXiv:1612.08293 (2017) http://arxiv.org/abs/1612.08293.