1. 1
Applied Water Solutions, Inc.
P.O. Box 50, Burlington, MA 01803
Phone (781) 750-8682 Fax: (781) 791-5769
www.appliedwatersolutions.com
Electrodeionization, Revolutionary or
Evolutionary Technology?
Abstract
Electrodeionization (EDI) technology has matured since its commercial
introduction over 16 years ago. Acceptance of the technology has been driven
by lower costs, higher water quality demands, and improvements in the
manufacturing techniques of the product. EDI has not been revolutionary but
evolutionary. This paper will review the advancements over the product life cycle
and explain the different product offerings available today. EDI, once a
proprietary technology, is available to original equipment manufactures (OEMs)
and has been accepted not only in the pharmaceutical industry, but
semiconductor, power, and all industries that have a need for pure water
production.
Author: Felice DiMascio, Ph.D
Director of Research and Development
Halox Technologies, Inc.
Co-Authors Chris Gallagher
Principal
Applied Water Solutions, Inc.
Carla Haschelevici
Director of Marketing
Applied Water Solutions, Inc.
2. 2
Applied Water Solutions, Inc.
P.O. Box 50, Burlington, MA 01803
Phone (781) 750-8682 Fax: (781) 791-5769
www.appliedwatersolutions.com
Advent of Electrodeionization
Introduction: The first use of electrodialysis in mass separation dates back to the early
1900’s. The cloth membranes used in the process only allowed a maximum of 50 %
deionization, due to considerable water seepage across the cloth membrane. The
development of highly ion permeable membranes in the 1940's was the first major
improvement in the electrodialysis technology. This enabled the electrodialysis process
to achieve greater than 50 % deionization when treating concentrated liquids. However,
the process was unable to achieve adequate deionization with dilute liquids due to an
effect called concentration polarization. In the 1950's, filling the spaces between the ion-
permeable membranes with ion exchange resins was found to be a very effective
method to overcome this negative polarization effect, allowing dilute liquids to be treated
effectively and economically. This modification resulted in a new technology, called
electrodeionization (EDI).
Principle of Electrodialysis: Electrodialysis (ED) is an electrochemical separation
process that is widely used in the water treatment industry for the deionization of
aqueous solutions. There are many plants currently operating worldwide, deionizing
millions of gallons of aqueous solution each day. The principle of ED is illustrated in
Figure 1, which shows a schematic diagram of a typical ED module arrangement,
consisting of a series of anion- and cation-permeable membranes arranged in an
alternating pattern between an anode and a cathode to form individual compartments.
+
+
-
+
+
+
+
+
+
+
+
+
+
+
CathodeAnode
C A
Concentrate
Diluate
Feed Solution
C A
+ + + + +
+ +
Figure 1. Schematic Diagram Illustrating the Principles of Electrodialysis.
If a solution containing dissolved ions is pumped through these compartments, and an
electrical potential established between anode and cathode, the positively charged
cations migrate towards the cathode, and the negatively charged anions towards the
anode. The cations pass through the negatively charged cation-permeable membranes,
but are retained by the positively charged anion-permeable membrane. Likewise, the
anions pass through the positively charged anion-permeable membranes, and are
3. 3
Applied Water Solutions, Inc.
P.O. Box 50, Burlington, MA 01803
Phone (781) 750-8682 Fax: (781) 791-5769
www.appliedwatersolutions.com
retained by the negatively charged cation-permeable membranes. The overall result is
an increase in the ion concentration in alternate compartments, while the other
compartments simultaneously become depleted.
History of the Electrodialysis: The first use of electrodialysis in mass separation dates
back to 1903, when Morse and Pierce1
introduced electrodes into two solutions
separated by a dialysis membrane. They found that electrolytes could be removed more
rapidly from a feed solution when applying an electrical potential. In the 1930’s Billiter et
al.2
and Streicher3
investigated early electrodialysis processes for industrial water
treatment. They investigated two- and three-compartment cells for the deionization of
salt solutions containing 1.0 g/L dissolved solids. The three-compartment cells
consisted of an anode compartment at one side of the module, a middle compartment,
and a cathode compartment at the other side of the module, being separated by cloth
membranes. When electrical potential was applied across these modules, cations of the
salt solution migrated from the middle of the module to the cathode compartment and
anions of the salt solution migrated from the middle of the module to the anode
compartment. Hydrogen and hydroxyl ions are produced at the anode and cathode,
respectively, by water splitting reactions. The hydrogen ions migrated from the anode
compartment into the middle compartment, and the hydroxyl ions migrated from the
cathode compartment into the middle compartment. Since the hydrogen and hydroxyl
ions have much higher mobilities that the salt ions, most of the ions migrating from the
outside to the middle of the module were hydrogen and hydroxyl ions, which combined
to produce water. The ions migrating from the middle to the outside of the module were
salt ions, thus reducing the salt content in the middle of the module. In the 1940’s,
Meyer and Strauss4
suggested a multi-cell arrangement, using alternating series of
anion- and cation-permeable membranes between two electrodes.
These type modules were only able to achieve a maximum of 50 percent (%)
deionization. The cloth membranes allowed considerable seepage of water carrying
undesirable ions back into the middle compartment, especially when the catholyte and
anolytes became more concentrated, and offered no resistance to the passage of
cations from the anode compartment to the middle compartment and anions from the
cathode compartment to the middle compartment.
The Advent of Ion Selective Membranes: The study of electrically charged
membranes in biological systems started in 1920’s by Michaelis5
. Teorell6
and Meyer
and Sievers7
were the firsts to theoretically describe permeability of charged membranes
to ions. The development of highly ion permeable membranes occurred in the late
1940's by Juda and NcRae8
. These membranes allow for the first major improvement in
the electrodialysis technology. Ion permeable membranes were manufactured in sheet
form using ion exchange resins; the two primary types therefore comprise of anion or
cation exchange resins. The structure of the membrane consisted of an integrated
network of copolymer chains, which had channels or pores of molecular dimension.
Attached to the wall of these pores were fixed ion exchange sites, or functional groups.
These membranes were permeable to counter-ions, impermeable to co-ions, and
relatively impermeable to water. The exchange sites on cation-permeable membranes
are negatively charged, which allowed positive charged species (cations) to freely
4. 4
Applied Water Solutions, Inc.
P.O. Box 50, Burlington, MA 01803
Phone (781) 750-8682 Fax: (781) 791-5769
www.appliedwatersolutions.com
permeate, while preventing negatively charge ions (anions) to permeate because of
electrical repulsion by the fixed charge. The opposite occurs with anion-permeable
membranes, where as anions are able to permeate while cations are not. The property
of water impermeability allowed these membranes to be very effective as boundaries of
a liquid flow conduit.
Upon their commercial availability, many new possibilities opened for the design of
processes requiring these type separators and the interest in electrodialysis process
increased. The cloth membranes were replaced by mechanically sturdy, highly
conductive cation- and anion-permeable membranes, preventing seepage of water into
the middle compartment and minimizing ion migration from the cathode and anode
compartments to the middle compartment. This enabled the electrodialysis process to
achieve greater than 50 % deionization. Langelier9
and Wiechers and Van Hoek10
discussed the production of potable water from seawater or brackish water by means of
multi-compartment ion-permeable membrane stacks.
Concentration Polarization: Deionization greater than 50 % was only possible when
treating concentrated liquids, such as seawater or brackish water, where the dissolved
ion concentration is on the order of 0.5 equivalents per liter (N). The process was
unable to achieve adequate deionization with less concentrated liquids due to a
limitation called concentration polarization. Concentration polarization at the membrane-
liquid interface set a limit on the flux rate. As the voltage across the membrane
increased, the current and hence the rate of transport of electrolyte ions across the
membranes increased until there was significant depletion of electrolyte ions in the
boundary layer next to the membrane. Any further increase in voltage resulted in water
splitting, since there were insufficient electrolyte ions near the membrane to carry the
current. This caused hydrogen and hydroxyl ions to participate in the conduction
mechanism, reducing the electrical efficiency. Also, water splitting changed the pH next
to the membranes and led to membrane fouling and precipitate formation, which
shortened the life of the membrane.
The Use Ion Exchange Resins: In the 1950's, the use of ion exchange resins was
investigated for improving the electrical efficiency for mass transfer by maintaining ohmic
conductivity of the liquid spaces between the ion-permeable membranes. The active
functional groups of ion exchange resins make the resins highly conductive when in
equilibrium with relatively low conductivity solution. In low conductivity solutions, resin
beads are orders of magnitude more conductive than the solution in which they are
immersed.
Modules were assembled using ion exchange resins in the compartments between the
ion selective membranes. These modules achieved greater than 90 % deionization with
less concentrated liquids, maintaining good electrical efficiency. The process was
efficient and economical for the deionization of liquids with less than 0.01 N dissolved
solids, offering considerable possibilities for the deionization of dilute liquids. This
discovery resulted in the advent of a new technology, called electrodeionization (EDI).
Development History of Electrodeionization: EDI theory and practice have been
advanced by a large number of researchers throughout the world. It is believed that the
5. 5
Applied Water Solutions, Inc.
P.O. Box 50, Burlington, MA 01803
Phone (781) 750-8682 Fax: (781) 791-5769
www.appliedwatersolutions.com
concept of the EDI process has been extensively investigated since the mid-1950’s, for
various purposes. Today, EDI is considered the latest generation in the water treatment
technology. An excellent comprehensive review of the technical literature on EDI is
given by Ganzi et al 11
.
Experimental work on the electrodeionization of solutions by passage through a mixed-
bed ion exchange resin was first described in publication by Walters et al12
and at
Argonne Laboratory in January 1955 as a method for removing trace radioactive
materials from water. They introduced an intimate mixture of cation and anion exchange
resin between ion selective membranes and achieved electro-regeneration of the ion
exchange resin when DC electrical current was applied. They passed water containing
ions through the resins and produced deionized water, proposing an ionic conduction
mechanism through mixed-bed ion exchange resin in contact with dilute solutions. They
also investigated a batch electrodeionization process for the concentration of radioactive
aqueous wastes.
It is believed that Permutit Company in the United Kingdom developed the first pilot
device incorporating intimately mixed cation and anion resins in the late 1950’s for the
Harwell Atomic Energy Authority. Sammons and Watts13
from the Harwell Atomic
Energy Authority studied the deionization of a sodium chloride solution using multi-cell
electrodeionization modules, quantifying the relationships between solution
concentration, flow rates, and applied DC current. They also investigated the
deionization of solutions containing calcium, iron, phosphate, and anionic detergents,
and discussed the application to decontaminate low-level aqueous radioactive waste.
Sammons and Watts demonstrated the feasibility of the EDI process, but they did not
define in great detail the effects of parameters such as liquid flow velocity, cell width,
particle size, and type of resin filling. Gittens and Watts14
extended the work by
investigating these parameters by studying the deionization of sodium nitrate solutions
using a five-cell electrodeionization module. They expressed their results in terms of a
theoretical model. Their work also extended into the basic principles of
electrodeionization, giving a detailed understanding of the nature of the underlying
processes, and the relationships between parameters of practical interest.
Kollsman applied for the first patent for an electrodeionization device in 1953. He
proposed a continuous method of separation between different ions but gave very little
data on effective operation. His apparatus was used for the purification of acetone. A
patent was awarded to Kollsman15
in 1957. In this time period, numerous patents were
granted for various types of electrodeionization devices, including those to Tye16
,
Pearson17
, Kressman18
, and Parsi.19
In 1959, Glueckauf discussed the design, and operating conditions of the EDI process
on a theoretical level20
. He proposed a theoretical model based on a two-stage removal
of ions; diffusive transfer of ions from flowing solution to ion exchange resin beads and
the transfer of ions along the chain of ion exchange beads.
Shimokawa21
, Lischke22
, Matejka23
, Shaposhnik24
, Korngold25
, and Kedem26
extended
the investigation into the operating conditions and performance of the EDI process in the
1960’s through 1980’s. In 1971, Matejka investigated the EDI process for the
6. 6
Applied Water Solutions, Inc.
P.O. Box 50, Burlington, MA 01803
Phone (781) 750-8682 Fax: (781) 791-5769
www.appliedwatersolutions.com
deionization of brackish or tap water to produce high-purity water. He described the
mechanism of the process as three simultaneous steps: the exchange of ions between
the solution and the ion exchange resin beads; the transport of ions along the chain of
ion exchange resin beads by direct electric current; and electrolytic regeneration of the
resins. His work advanced the theory of ionic transport within an electrodeionization
device. Kedem introduced ion exchange spacers between membranes and found that
the mass transfer of monovalent ions was increased. EDI using media comprising
separate and sometimes alternating ion exchange layers, each layer substantially
comprising of resin of same polarity, e.g. each layer either anion or cation resin, was
described by Kunz27
in the patent literature in the early 1980’s. During this time, new
devices were being proposed and patents being granted to a number of others, including
those to Selegny28
, Davis29
, Tejeda30
, O’Hare31
, and Giuffrida et al32
.
Millipore Corporation (Bedford, MA) introduced the first commercially available EDI
modules and systems in 1987 under the trade name Ionpure. The expansion of the
technology has continued since 1989 by U.S. Filter (Palm Desert, CA), who acquired
Ionpure in 1993.
The evolution of EDI technology is shown in Figure 2. It also includes the individuals
that were instrumental in the advancement of the EDI technology.
Figure 2 Evolution of Electrodeionization
1980s –
Kunz,
Selgny,
Davis,
Tejeda,
O’Hare and
Giuffrida et
al.
1992 -
Ganzi, Wood
and Griffin
1962-
Shimokawa,
Lischke
1971 -
Matejka
1973 –
Shaposhnik
1975-
Korngold,
Kedem
1940 – Juda
and NcRae
1953 -
Kollsman
1955 –
Argonne Lab
–Walters et al
1959 –
Glueckauf
1930 –
Billitier et
al. and
Streicher
1903 –
Morse &
Pierce
Industry
Icons
2020200019801960194019201900Year
1980s –
Kunz,
Selgny,
Davis,
Tejeda,
O’Hare and
Giuffrida et
al.
1992 -
Ganzi, Wood
and Griffin
1962-
Shimokawa,
Lischke
1971 -
Matejka
1973 –
Shaposhnik
1975-
Korngold,
Kedem
1940 – Juda
and NcRae
1953 -
Kollsman
1955 –
Argonne Lab
–Walters et al
1959 –
Glueckauf
1930 –
Billitier et
al. and
Streicher
1903 –
Morse &
Pierce
Industry
Icons
2020200019801960194019201900Year
First use
of ED in
mass
separation
First use
of ED in
mass
separation
Highly ion
permeable
membranes
Highly ion
permeable
membranes
Use of Ion
Exchange resin
between
membranes
Use of Ion
Exchange resin
between
membranes
Walters - Use of
mixed bed ion
exchange and
electro-
regeneration
Walters - Use of
mixed bed ion
exchange and
electro-
regeneration
First patents
granted for EDI
First patents
granted for EDI
EDI for
brackish or
tap water to
produce
High Purity
water
EDI for
brackish or
tap water to
produce
High Purity
water
First
commercial
EDI systems
introduced
First
commercial
EDI systems
introduced
Reverse
Osmosis
Reverse
Osmosis
Spiral
Wound
module
offered to
market
Spiral
Wound
module
offered to
market
Thick cell
EDI offered
to market
Thick cell
EDI offered
to market
First Pilot:
Mixed cation
and anion
First Pilot:
Mixed cation
and anion
Thin cell EDI
offered to
market
Thin cell EDI
offered to
market
7. 7
Applied Water Solutions, Inc.
P.O. Box 50, Burlington, MA 01803
Phone (781) 750-8682 Fax: (781) 791-5769
www.appliedwatersolutions.com
Today, a number of companies have become involved with EDI:
• Christ AG (Aesch, Switzerland), a member of the BWT Group,33
has developed
and constructed a spiral wound EDI module called Septron™. In February 2004,
Christ AG purchased a 26 % interest in the water technology company Tenergy
Water (New Britain, CT) at which time this company changed its name to Tenergy
Christ Water LLC.
• A new EDI process was developed by Research Centre Jülich (KFA) and
presented the Achema fair in Frankfurt in 1994. SG Water Treatment &
Regeneration (Barsbuttel, Germany) manufactures combined reverse osmosis
(RO) and EDI system configurations.
• Glegg Water Conditioning Inc. (Ontario, Canada) launched its own version, called
the E-Cell™, which was introduced in December 1996. In October 1999, General
Electric acquired 82 % of Glegg Industries, Inc. In February 2005, General Electric
acquired Ionics (Waterford, MA), who manufactures a horizontal plate-and-frame
EDI module configuration
• In 1999, Vivendi Universal (Paris, France) acquired U.S. Filter, placing it under its
utilities unit, Vivendi Environnement which changed its name to Veolia
Environnement. The Ionpure name was re-established in 2003, offering a unique
module configuration called VNX EDI Module™. In 2004, Siemens (Frankfurt,
Germany) acquired U.S. Filter from Veolia Environnement.
• Electropure Inc. (Laguna Hills, CA) manufactures an EDI module, called
Electropure XL™. In October 2005, SnowPure LLC (San Clemente, CA)
purchased the Electropure EDI division from Electropure Inc.
• Omex Environmental Engineering Co. Ltd. (Zhejiang, China) manufactures a spiral
wound EDI, called the Omexell, and EDI components. In June 2006, the Dow
Chemical Company (Midland, MI) acquired Omex Environmental Engineering.
The commercial introduction of EDI technology came in a complete system and was a
propriety product through its first few years. The system design, including pretreatment,
was critical to the successful operation because the limitations of the technology were
not fully understood. The pharmaceutical industry was the first to recognize its benefits.
Some of the first EDI installations were operated on conventional media filtration, i.e.,
softeners, organic ion exchange, and media filters. The earlier pretreatment designs
were found to be unreliable and today suppliers recommend the use of RO as
pretreatment.
The EDI technology today has advanced to where it is available from suppliers as a
component to OEMs that design complete systems similar to the availability of RO
membranes. This was first initiated by Electropure, Inc. in the early 1990’s and has
been continued by others. However, because of the competitive nature in the industry
and a continuation of new products every two to three years, it is unlikely that there will
be harmonization of a standard EDI in the near future.
8. 8
Applied Water Solutions, Inc.
P.O. Box 50, Burlington, MA 01803
Phone (781) 750-8682 Fax: (781) 791-5769
www.appliedwatersolutions.com
The intense competition among suppliers has also simplified the design and significantly
decreased the system and module cost. The advancements include removal of the
concentrate recirculation system, removal of a brine injection system, and ‘leak free’
designs. Improvements in manufacturing have allowed for higher quality products,
improved plastics allowing for heat sanitizable systems, and ion exchange resin
configurations allowing for a wider range of applications.
It is this author’s opinion that the technology will continue to evolve with new offerings,
new applications, larger capacity systems, and systems that can allow for less strict feed
water requirements.
9. 9
Applied Water Solutions, Inc.
P.O. Box 50, Burlington, MA 01803
Phone (781) 750-8682 Fax: (781) 791-5769
www.appliedwatersolutions.com
References:
1
Morse, H.N. and Pierce, J.A., Z. physik. Chem., 45, 589 (1903)
2
Billiter, J., Trans. Electrocheical Society, 60, 217 (1931)
3
Streicher, L., Bowers, A.E., and Briggs, R.E., Ind. Eng. and Chem, 45, 2394 (1953)
4
Meyers, K.H. and Strauss, Helv. Chem. Acta, 23, 795 (1940)
5
Michaelis, L., Colloid Symp. Monogr., 5, 135 (1928)
6
Teorell, T., Trans. Faraday Soc., 33, 1053 (1937)
7
Meyers, K.H. and Sievers, J.F., Helv. chim. Acta, 20, 634 (1937)
Meyers, K.H. and Sievers, J.F., Helv. chim. Acta, 23, 795 (1940)
8
Juda, W. and McRae, W.A., J. Am. Chem. Soc., 72, 1044 (1950)
9
Langelier, W. F., Journal of American Water Works Association, 44, 845 (1952)
10
Wiechers, S. G., and Van Hoek, C., Research, 6, 192 (1953)
11
Ganzi, G.C., Wood, J.H. and Griffin, C.S., "Water Purification and Recycling Using
the CDI Process", Environmental Progress, Feb., 11, (1992)
12
Walters, W.R., Weiser, D.W., and Marek, J.L., "Concentration of radioactive aqueous
wastes - Electromigration through ion-exchange membranes", Ind. Eng. Chem., 47,
61 (1955).
13
Sammon, D.C., and Watts, R.E., "An Experimental Study of Electrodeionisation and
its Application to the Treatment of Radioactive Wastes", AERE-R3137, Chemistry
Division, U.K.A.E.A. Research Group, Atomic Energy Research Establishment,
Harwell, June 1960.
14
Gittens, G.J., and Watts, R.E., "Some Experimental Studies of Electrodeionisation
Through Resin Packed Beds", AERE-R4517, Chemistry Division, U.K.A.E.A.
Research Group, Atomic Energy Research Establishment, Harwell, March 1964.
15
Kollsman - U.S. Patent No. 2,815,320
16
Tye - U.K. Patent # 815,154
17
Pearson - U.S. Patent # 2,794,777
18
Kressman - U.S. Patent #2,923,674
19
Parsi - U.S. Patent # 3,149,061
20
Glueckauf, E., "Electro-deionisation Through a Packed Bed", British Chemical
Engineering, Dec., 646 (1959)
21
Shimokawa, J., JAERI (rep. Jap. Atom. Energy Res. Inst.), 1962, No. 1038
22
Lischke, P., Thesis, Institute of Chemical Technology, Prague. 1962
10. 10
Applied Water Solutions, Inc.
P.O. Box 50, Burlington, MA 01803
Phone (781) 750-8682 Fax: (781) 791-5769
www.appliedwatersolutions.com
23
Matejka, Z., "Continuous Production of High-Purity Water by Electro-deionisation, J.
Appl. Chem. Biotechnol., 21, 117 (1971)
24
Shaposhnik, V.A., Reshetnikova, A.K., Zolotareva, R.I., I.V. Drobysheva, and N.I.
Isaev, "Demineralization of Water by Electrodialysis with Ion-exchanger Packing
Between the Membranes", Zhurnal Prikladnoi Khimii, Vol. 46, #12, pp. 2659-2663,
December, 1973, translated by Consultants Bureau, Plenum Publishing Corp., 227
W. 17th St., New York.
25
Korngold, E., "Electrodialysis processes using ion exchange resins between
membranes", Desal., 16, 223 (1975)
26
Kedem, O., "Reduction of polarization in electrodialysis by ion-conducting spacers",
Desal., 16, 105 (1975)
27
Kunz - U.S. Patent # 4,636,296
28
Selegny - U.S. Patent # 3,686,089
29
Davis - U.S. Patent # 4,032,452
30
Tejeda - U.S. Patent # 3,869,376
31
O’Hare - U.S. Patent # 4,465,573
32
Giuffrida - U.S. Patent # 4,632,745
33
Rychen - U.S. Patent # 5,376,253