By Avijit Dey

This article is based on an education session presented by the author at WQA Aquatech USA 2005. The complete work is available exclusively online at

Electrodeionization (EDI) is a separation process combining electrodialysis and conventional ion exchange; the resulting hybrid process entails zero regenerant chemicals. One benefit of this technology includes the elimination of secondary hazardous waste associated with the chemical regeneration of ion exchange tanks. For fundamentally the same reason that cation and anion units are cost-effectively replaced by reverse osmosis (RO) units, EDI units in many cases can cost-effectively replace mixed bed deionizers. RO and EDI technologies enjoy a symbiotic relationship with each contributing complimentary characteristics that enhance the performance of overall treatment process. EDI technology in combination with RO offers a chemical-free option for the removal of ionic impurities from raw water. Industrial scale EDI devices are available in two major configurations: plate-and-frame or spiral wound. A new, more advanced Spiral Wound Electrodeionization (SWEDI) system was created about five years ago. The proven reliability of SWEDI modules in the high purity water generation application is substantiated by more than 40,000 gallons per minute (gpm) capacity of successful installations. These systems have now been in operation at several micro-electronics and power plants for over three years.

SWEDI is a robust and forgiving technology. SWEDI represents a breakthrough in desalination of RO permeate water as compared to plate-and-frame configuration. Associated benefits include leak-free operation up to 100 psig feed pressure because of the ease of sealing cylindrical pressure vessels; higher feed water hardness tolerance due to unique cross flow operation; ion exchange resins and/or membranes replacement; excellent electrical insulation due to the use of fiberglass reinforced polymer (FRP) housing; light module weight and simple system integration.

Cross flow is the term given to the mode of EDI operation where the direction of flow in the concentrate chambers is perpendicular to that in the diluting chambers. Conventional plate-and-frame devices are essentially operating in co-flow configuration; co-flow is the term given to the mode of EDI operation where the direction of flow in the concentrate chambers is the same as in the diluting chambers. The unique cross-flow design of the SWEDI modules results in higher feed water hardness tolerance compared to the conventional plate-and-frame devices. Vendors of plate-and-frame EDI devices suggest that the maximum total hardness in the feed to the EDI be limited to 0.5 to 1 ppm as CaCO3 whereas SWEDI modules can accept a maximum total hardness of 2 ppm as CaCO3 in the feed to the EDI. It sometimes translates to a reduction in capital expenditures for water treatment systems by eliminating the need for softeners and/or double pass RO ahead of the EDI systems.

The purification of aqueous streams by reducing the concentration of ionic impurities has been an area of considerable technical interest. Many techniques have been developed for this purpose. The most well known processes include electrodialysis, electrodeionization (EDI), liquid chromatography, membrane filtration and ion exchange. EDI is a separation process combining electrodialysis and conventional ion exchange; the resulting hybrid process does not entail regenerant chemicals. EDI technology has been reported in the literature since teh 1950s1, 2. The first patent for EDI was granted to a Dutch company in 19573. A patent for the purification of acetone with EDI was also granted to Kollsman in 19574. For fundamentally the same reason that cation and anion units are cost-effectively replaced by RO units, EDI units in many cases can cost-effectively supplant mixed bed deionizers5. EDI process is preferred in many systems primarily because of the environmental benefit of not requiring hazardous regenerant chemicals and inherent superiority of a continuous process over a batch process6. Although it may be possible to regenerate mixed bed resins using regenerants of high purity to achieve quality of water required even in electronics industry, this is inept due to the cost of producing and maintaining ultrapure regenerant. Additionally, it is critically important for mixed bed deionizer operation that the resins settle properly and the division between them occurs at the middle collector. Furthermore, the particle size range is limited by the separation requirement. The separation occurs because of the difference in the density of the two types of resin. The cation resin, being heavier, settles on the bottom, while the anion resin, being lighter, settles on top of the cation resin. The degree of separation can be determined using the terminal settling velocities (TSV) of the cation and anion resins respectively. The TSV is calculated using Stokes’ Law7:

2 x g x a2 x (d1 – d2)
V  =
9 x η
V    =    Resin terminal settling velocity, cm/s
g    =    Gravitational constant (= 981 cm/s2)
a    =    Radius of resin bead, cm
d1    =    Density of resin, g/cm3
d2    =    Density of water, g/cm3
η    =    Viscosity of water, poise

In the case of regenerable mixed beds, it is essential that distinct separation of cation and anion resins occurs upon backwashing to assure complete regeneration of the resins and avoid contamination by the other regenerant. However, the very nature of a regenerable mixed bed requires compromise: the resins must be able to separate for regeneration and to remain intimately mixed during exhaustion8. Obviously these are  contradictory requirements. If cation resin remains in the anion zone, it will absorb sodium from the caustic regenerant. Conversely, anion resin remaining in the cation zone will attract either sulphate or chloride depending on whether sulphuric or hydrochloric acid regenerant is used for the regeneration of cation resins. Consequently, EDI systems provide a technically superior alternative in a cost-effective manner.

One key factor stimulating the EDI market growth is the desire to move to a water treatment approach that consumes fewer chemicals. A main benefit of the EDI system is the elimination of the need for on-site bulk storage of concentrated acid and caustic regenerants and associated operator exposure. Moreover, no hazardous waste stream is generated by this technology. In combination with RO, EDI provides a continuous, chemical-free system. On the other hand, purifying water through regenerated resin beds may lead to organic release onto the downstream treatment steps. When newly regenerated resin beds are installed, high total organic carbon (TOC) leakage may occur. It is interesting to note that EDI systems can readily remove charged organics9. Moreover, EDI systems use a continuous process, which does not require any redundancy like the conventional mixed bed deionizers. The main consumable in EDI is electricity. The electrical cost of running an EDI system is usually between 0.5 to 3.0 kW-h/1,000 gallons of product water, depending on feedwater quality and product water specifications10. Recent developments in EDI technology have resulted in significant cost reductions arising from an expandable, modular system approach built upon a base module. EDI systems require minimal civil works for installation.

Some EDI modules utilize a spiral-wound membrane and ion exchange resins, sealed in a high strength FRP pressure vessel11. Such an EDI system is composed of an arrangement of flow-directing spacers separated by anion and cation semipermeable membranes. Spacers are provided between alternating cation and anion exchange membranes to maintain separation of associated membranes. Spacers, usually manufactured from thermoplastic materials, support even distribution of liquid through the chamber. Permselective ion exchange membranes are arranged to form parallel flow chambers. In EDI systems, ions are removed from water by a combination of ion exchange resin, ion exchange membranes and an electric current (DC). The feedwater entering the apparatus is divided into at least three parts. A small percentage flows over the electrodes, a majority of the feed passes through the diluting chambers and the remainder passes along the concentrating chambers. Diluting chambers are filled with specially graded ion exchange resins. A conductive path is developed through the resin beads that are much lower in electrical resistance compared to the path through the bulk solution. The chambers bounded by the anion membrane facing the anode and the cation membrane facing the cathode become exhausted of ions and are hence called diluting chambers. The chambers bounded by the cathode-facing anion membrane and anode-facing cation membrane are called concentrating chambers. The concentrating chambers will then trap ions that have electrically migrated in from the diluting chambers. Consequently, the ionic concentration of the water will decrease in the diluting chamber with a corresponding increase in the ionic concentration of the water in the concentrating chamber. These flow chambers are hydraulically in parallel but electrically in series.

In the present SWEDI systems, feed water enters the modules from below and is diverted into vertically spiraled dilute chambers. Concentrate enters the module through the central pipe from below and is diverted into spirally flowing concentrate chambers (See Figure 1). The patented cross-flow design of the concentrate stream in the present SWEDI modules is unlike that of the co-flow design used in conventional plate-and-frame devices. This imparts a higher hardness tolerance for the present SWEDI devices. The present SWEDI module offers the only serviceable EDI system in the market as the replacement of resins and/or membranes is possible at any time while reusing the housings, end caps and titanium anode. This unique feature dramatically reduces the operating cost by extending the useful life of the present SWEDI modules. Therefore, operating costs are lowered by reducing the high replacement cost for the conventional EDI modules. The influence of EDI module life is far greater on the overall operating cost than the influence of power consumption by the EDI module and associated electrical equipment in the EDI system. As these SWEDI modules are serviceable on-site, it is not necessary to send the modules back to the manufacturing facility.

Higher conductivity in the concentrate stream accomplished by concentrate recirculation and occasional brine injection in the concentrate stream facilitates the transfer of current while back-diffusion can limit effluent quality. Consequently, a certain percentage of the concentrate flow must go to the drain. The increase in the electric current flow due to higher electrical conductivity of the concentrate water lessens the power consumption of the device because the EDI module resistance and applied voltage decreases as a result. Further improvement in the power economy was accomplished by passing the current through a series of cell pairs with electrodes only at the terminal points of the series. A cell pair consists of a cation membrane, spacing material through which the feed and the treated water flows in the diluting chamber, an anion membrane and the spacing material through which the concentrate water flows. Consequently, the loss of energy at the electrodes per equivalent of salt removal from the RO permeate water will reduce to a minimum.

The current also splits water molecules into hydrogen and hydroxyl ions. The EDI systems operate in two different modes: Electrodeionization and Electroregeneration12. The system is working in the first mode when the feedwater salinity is high13. On the contrary, the system is working in the second mode when the feedwater salinity is very low due to the passage of strongly ionized species to the concentrating chamber in the upstream flow path. This allows a portion of the resins in the EDI to always be in the fully regenerated state. This will form a highly conductive path through the resin beads. Moreover, higher degree of regeneration of resin beads will reduce the leakage of all the ionic species including boron and silica.


  1. Walters, W. R., Weiser, D. W., and Marek, L. J. “Concentration of radioactive aqueous wastes—Electromigration through ion exchange membranes”, Industrial Engineering Chemistry, 47 (1), pp 61-67, 1955.
  2. Kollsman, P. “Method and apparatus for treating ionic fluids by dialysis”, U. S. Patent number 2,815,320, December, 1957.
  3. Verkeer, H. E. et al., U.K. Patent, 776,469, 1957.
  4. Kollsman, P. U.S. Patent, 2,815,320, 1957.
  5. Paul, D. H. “Electrodeionization in Pharmaceutical Water Treatment”, Pharmaceutical Technology, pp 36-42, July 2002.
  6. Hernon, B., Zanapalidou, H., Prato, T., and Zhang, L. “Removal of weakly ionized species by EDI”, presented at the 59th Annual International Water Conference, October 19-22, Pittsburgh, Penn., 1998.
  7. Mulligan, R., and Stahlbush, J. Personal Communication, 2000.
  8. Dey, A., and Thomas, G., “Electronics Grade Water Preparation” ISBN 0-927188-10-4, Tall Oaks Publishing Inc., Littleton, Colo., 2003.
  9. Wilkins, F. C., and McConnelee, P. A. , Solid State Technology, August 1988.
  10.  Tate, J. “Add polish to high-purity water with EDI: System combine RO and ion electrochemistry”, Water Technology, August, 2000.
  11.  Xiang, L., and Gou-Lin, L. “Helical Electrodeionization apparatus”, U.S. Patent, 6, 190, 528, 2001.
  12.  Ganzi, G. C. “Electrodeionization for high-purity water production”, AIChE Symposium series, Number 261, Volume 84, “New membrane materials and processes for separation”, edited by Sirkar, K. K., and Lloyd, D. R., New York, N.Y., pp 73-83, 1988.
  13. Ganzi, G. C., and Parise, P. L. “The production of pharmaceutical grades of water using continuous deionization post-reverse osmosis”, Journal of Parenteral Science and Technology, Parenteral Drug Association, July/ August, 1990.

About the author
Dr. Avijit Dey received his Ph.D. in chemical engineering from Jadavpur University, Calcutta, India. He is currently Technical Manager at Omexell Inc., a research and development, design, engineering manufacturer specializing in Integrated Membrane Technology (IMT) equipment used in the process of water treatment. Omexell has been developing membrane technologies since 1998 and has now emerged as a leader in IMT with its patented spiral wound EDI and unique hollow fiber UF technologies. Dey can be reached at (713) 973-9731 or via email at [email protected]


Comments are closed.