ELECTRO-DEIONIZATION (EDI)

Text may appear fuzzy or blocky due to file compression. You may be able to change the
appearence (Acrobat Reader 3.x) by going to the menu item File and selecting Preferences,
then General. Click the box next to the item “Smooth text and monochrome images” and
then the OK box at the bottom.

Contents
8.1 Fundamentals ........................................................................................................... 2
8.1.1 EDI Theory of Operation ........................................................................ 9
8.1.1.1 Ion Exchange.................................................................................. 10
8.1.1.2 Ion Removal ................................................................................... 11
8.1.1.3 Regeneration .................................................................................. 11
8.1.2 EDI Performance ..................................................................................... 12
8.1.2.1 EDI Compared to Conventional Mixed-Bed Deionization .. 12

8.2 Process and Operations Overview ..................................................................... 15

8.2.1 Process Flow ............................................................................................ 16

8.3 Equipment Design and Options ......................................................................... 23

8.3.1 E-Cell Stack™ .......................................................................................... 24
8.3.2 Control Panel ........................................................................................... 25
8.3.2.1 PLC .................................................................................................... 25
8.3.2.2 Operator Interface .......................................................................... 26
8.3.3 Brine Injection System (Optional) ...................................................... 27
8.3.4 Diverter Valves (Optional) .................................................................... 28
8.3.5 Concentrate Pump .................................................................................. 28
8.3.6 Piping ........................................................................................................ 29
8.3.7 Valves ........................................................................................................ 29
8.3.8 Electrical Power Rectifier ...................................................................... 30
8.3.9 Instrumentation ...................................................................................... 31

8.4 Applications and Design ..................................................................................... 32

8.4.1 E-Cell™ Process Design ........................................................................ 33
8.4.2 Feedwater Analysis and E-Cell™ Feedwater Specification ........... 33
8.4.2.1 Anion and Cation Loading .......................................................... 34
8.4.2.2 Calculation of Maximum Allowable Recovery ....................... 34
8.4.3 Hydraulic sizing ..................................................................................... 35
8.4.4 Brine Injection For Concentrate Loop Management ....................... 35
8.4.5 Current Determination and Rectifier Sizing ..................................... 36
8.4.6 Concentrate Pump Sizing ..................................................................... 38

34a_m8_r0 8-1 Electro-Deionization

 Module 8
ELECTRO-DEIONIZATION (EDI)

8.1 Fundamentals

Electro-Deionization, or EDI, is a process that evolved from conventional ion

exchange technology. EDI provides continuous demineralization at recovery rates of
90% or more. In EDI, just as in conventional ion exchange, cations and anions in the
feedwater are exchanged for hydrogen and hydroxyl ions in the ion exchange resins,
producing demineralized water. The key operational difference is that with EDI, the
ion exchange resin is regenerated continuously, while with conventional ion ex-
change, chemical regeneration is performed intermittently.

Continuous regeneration in EDI is achieved electrochemically, by means of ion

conducting membranes and an imposed electric current. The hydrogen and hy-
droxyl ions necessary for regeneration are formed in-situ, without addition of chemi-
cal reagents, by means of the familiar water dissociation reaction, sometimes called
water splitting. Figure 8.1-1 is an illustration representing a water molecule splitting
into hydrogen and hydroxyl ions.

H+

OH-
H2O

Figure 8.1-1: Water Molecule Splitting

34a_m8_r0 8-2 Electro-Deionization

table of contents
 To understand how Electro-Deionization (EDI) removes ions from water, we need to
understand the components of an EDI stack and the function of each. An EDI stack
consists of multiple beds of ion-exchange material sandwiched with membrane walls
and open channels between two electrodes.

The two electrodes are located at opposite ends of the stack, as seen in Figure 8.1-2.
These electrodes supply an electric current to the water flowing inside the cells. One
of these electrodes is the cathode. It is negatively charged and is a source of
electrons. The cathode attracts cations (positively charged ions).

The second electrode is the anode. This electrode is positively charged and attracts
anions (negatively charged ions). This attraction occurs because opposite charges
attract and like charges repel. Therefore, a negatively charged cathode attracts posi-
tively charged ions, and a positively charged anode attracts negatively charged ions.

Ultrapure Concentrate Ultrapure Concentrate

Anion Membrane
Product Recycle Product Recycle
Anion Membrane

Cation Membrane

Cathode Anode
Electrode Na+ Electrode
Na+
Na+ Cl
-

CATHODE ANODE
Cl- Na+

H+
Cl- Cl-

Na+
OH- OH-

Anion
Resin
Cation
Na+ Cl- Resin
Na+ Cl-

Electrolyte Feed Diluting Concentrating Feed Diluting Concentrating Electrolyte

Compartment Compartment Compartment Compartment Compartment Compartment
(E Chamber) (D Chamber) (C Chamber) (D Chamber) (C Chamber) (E Chamber)

Multiple Cells

Figure 8.1-2: View of an EDI Cell with Electrodes Indicated

34a_m8_r0 8-3 Electro-Deionization

table of contents
 The next items we need to notice are the membranes that separate the chambers.
Figure 8.1-3 below shows an EDI stack with the membranes highlighted. There are
two types of membranes: anion membranes and cation membranes. Using the same
reasoning used for the electrodes, the anion membranes allow only negatively
charged ions (anions) to permeate, and the cation membranes allow only the posi-
tively charged ions (cations) to permeate. Water does not permeate these
membranes.

Anion Membranes

Ultrapure Concentrate Ultrapure Concentrate

Anion Membrane
Product Recycle Product Recycle
Anion Membrane

Cation Membrane

Na+
Cl- Na+
Na+ ANODE
CATHODE Cl- Na+

H+
Cl- Cl-

Na+
OH- OH-
Anion
Resin
Cation
Na+ Cl- Na+ Resin
Cl-

Electrolyte Electrolyte
Feed Diluting Concentrating Feed Diluting Concentrating
Compartment Compartment
Compartment Compartment Compartment Compartment
(E Chamber) (E Chamber)
(D Chamber) (C Chamber) (D Chamber) (C Chamber)

Multiple Cells

Cation Membrane

Figure 8.1-3: View of an EDI Cell with Membranes Indicated

34a_m8_r0 8-4 Electro-Deionization

table of contents
 The membranes are situated so that they become the walls that separate chambers.
There are three types of chambers; the Dilute or “D” chambers, the Concentrate or
“C” chambers, and the Electrolyte or “E” chambers. Figure 8.1-4 shows an EDI stack
with the three different types of chambers indicated.
Concentrate Chambers

Ultrapure Concentrate Ultrapure Concentrate

Anion Membrane
Product Recycle Product Recycle
Anion Membrane

Cation Membrane

Na+
Na+
Cl-
+
CATHODE
Na ANODE
Cl- Na+

H+
Cl- Cl-

Na+
OH- OH-

Anion
Resin
Cation
Na+ Cl- Na+ Resin
Cl -

Electrolyte Feed Diluting Concentrating Feed Diluting Concentrating Electrolyte

Compartment Compartment Compartment Compartment Compartment Compartment
(E Chamber) (D Chamber) (C Chamber) (D Chamber) (C Chamber) (E Chamber)

Dilute Chambers

Electrolyte Chambers

Figure 8.1-4: View of an EDI Cell with Chambers Indicated

34a_m8_r0 8-5 Electro-Deionization

table of contents
 Water to be treated is fed into the Dilute or “D” chamber. This chamber contains a
resin bed that consists of cation and anion beads compressed between an anion
membrane and a cation membrane. These resin beads allow the ions to adsorb into
the respective beads and travel through the beads toward the membranes.

Once contaminant ions pass through the membranes, they are in the
Concentrate or “C” chamber. Ions are swept away by a recirculating flow
(the concentrate loop) which has a bleed to prevent an excessive buildup of ions, and
a corresponding makeup flow to introduce more fresh water. Figure 8.1-5, below,
shows a graphic of an EDI stack with the concentrate loop highlighted.

Concentrate
Loop

Concentrate
Loop

Concentrate
Loop

Figure 8.1-5: Graphic of an EDI Stack

with the Concentrate Loop Highlighted

34a_m8_r0 8-6 Electro-Deionization

table of contents
 One D chamber, one cation membrane, one C chamber and one anion membrane
together are called one cell (or one cell pair). An EDI stack is made up of multiple
cells operating in parallel. Figure 8.1-6, below, is the EDI Stack with one of its cells
highlighted.

Ultrapure Concentrate Ultrapure Concentrate

Recycle

Anion Membrane
Product Product Recycle
Anion Membrane

Cation Membrane

Na+
Cl- Na+
CATHODE Na+ ANODE
Cl- Na+

H+
Cl- Cl-

Na+
OH- OH-
Anion
Resin
Cation
Na+ Cl- Resin
Na+ Cl-

Electrolyte Feed Diluting Concentrating Feed Diluting Concentrating Electrolyte

Compartment Compartment Compartment Compartment Compartment Compartment
(E Chamber) (D Chamber) (C Chamber) (D Chamber) (C Chamber) (E Chamber)

One Cell
Figure 8.1-6: Graphic of an EDI stack with a Cell Highlighted

34a_m8_r0 8-7 Electro-Deionization

table of contents
 The final chambers are the Electrolyte or “E” chambers. These two chambers contain
the electrodes and are fed water from the concentrate loop. Like the Concentrate
chambers, the E chambers receive contaminant ions from the nearest membrane.
The cathode E chamber also receives small amounts of hydrogen gas generated by
the reduction of H+ ions, which have migrated to the cathode:

2H+ + 2e- à H2

The anode E chamber similarly receives small amounts of oxygen and chlorine gas
generated by the oxidation of OH- and Cl- ions:

4OH- à 4e- + 2H20 + O2

2Cl- à 2e- + Cl2

The flow from the E chambers is generally sent to waste to prevent membrane attack
by any chlorine gas generated. Figure 8.1-7 is a graphic illustrating Hydrogen,
Oxygen and Chlorine gas production at the electrodes.

+ -

Battery

Anode Cathode
H2 Cl2 O2
Cl-
H+
Cl-
H+
OH-
H+
+
OH-
H

Figure 8.1-7: Production of Hydrogen, Oxygen,

and Chlorine Gas at the Electrodes

34a_m8_r0 8-8 Electro-Deionization

table of contents
 H+ OH-

Resin Na+ Cl- Resin

H+ OH-

Resin Resin

Na+ Cl-

Figure 8.1-8: Graphic Illustrating Ion Exchange

Cl- Na+
OH-
Na+ Cl- OH-
OH - Cl-
Cl-
H+ Cl-

Na+ OH-
+
Na

OH-
H+
H+
OH-
Na+
Cl-
H+

Figure 8.1-9: Graphic Illustrating Ion Removal

34a_m8_r0 8-9a Electro-Deionization

table of contents
 8.1.1 EDI Theory of Operation

Continuous demineralization in an EDI stack consists of three coupled processes:

a) Ion exchange, in which the feedwater is passed through a bed of ion exchange
resin. The ions in the feedwater are taken up by the resin just as in conventional ion
exchange, subject to the usual mass transfer and thermodynamic equilibrium
considerations. See Figure 8.1-8 on the opposite page.

b) Continuous ion removal by transport through the ion exchange resin and
membranes into the concentrate stream. This process is particular to EDI and is
driven by the applied direct current. See Figure 8.1-9 on the opposite page.

c) Continuous regeneration by hydrogen and hydroxyl ions derived from the water
splitting reaction and driven by the applied direct current. This process is particular
to EDI and can proceed even in the absence of ions in the feedwater. See
Figure 8.1-10 below.

H+ OH-

H+
OH-

H2O

Figure 8.1-10: Graphic Illustrating Continuous Regeneration

34a_m8_r0 8-9 Electro-Deionization

table of contents
 Ultrapure Concentrate Ultrapure Concentrate

Anion Membrane
Product Recycle Product Recycle

Anion Membrane

Cation Membrane
Na+
Cl- Na+
Na+ ANODE
CATHODE Cl- Na+

H+
Cl- Cl-

Na+
OH- OH-
Anion
Resin
Cation
Na+ Cl- Na+ Resin
Cl-

Electrolyte Feed Diluting Concentrating Feed Diluting Concentrating Electrolyte

Compartment Compartment Compartment Compartment Compartment Compartment
(E Chamber) (D Chamber) (C Chamber) (D Chamber) (C Chamber) (E Chamber)

Multiple Cells

Figure 8.1-11: EDI Stack

Refer to the representation of an EDI stack in Figure 8.1-11 for the following
discussions.

8.1.1.1 Ion Exchange

The water to be treated enters the EDI stack through the Dilute or “D” chambers. As
ions encounter the resin beads the usual ion exchange reactions take place as they
would in a regular ion exchange bed (Refer to module 5 section ). Ions are absorbed
into the resin beads, in exchange for release of hydrogen and hydroxyl ions. As the
water flows through the resin bed more ions are exchanged until high resistivity
product water is produced at the bed outlet.

34a_m8_r0 8-10 Electro-Deionization

table of contents
 8.1.1.2 Ion Removal

Ions in the resin bed move through the beads under the influence of the applied
electric field. The positive charge of a sodium ion (for example) causes it to
migrate towards the cathode. The sodium ion travels through the cation resin beads
until it reaches the cation membrane. Since the cation membrane is permeable to
cations, it allows the positive sodium ion to pass through the membrane and enter
the Concentrate (or C) chamber. Once in the C chamber, the sodium ion that is still
migrating towards the cathode (but now through water, not a resin bead), is trapped
inside the C chamber by the anion membrane. The positive sodium ion cannot
permeate the anion membrane. At this point the recirculating Concentrate flow
sweeps the sodium ion out of the chamber.

At the same time, a negatively charged chloride ion (for example) migrates similarly
but in the opposite direction. The negative chloride ion migrates towards the anode
under the effect of the electric field. The negative chloride ion attaches to and travels
through the anion resin beads and permeates the anion membrane into the Concen-
trate chamber on the other side of the D chamber. Once in the concentrate chamber
the ion is swept out of the cell by the Concentrate flow.

8.1.1.3 Regeneration

The electric field applied across the stack is sufficient to break apart a significant
quantity of water molecules into hydrogen and hydroxyl ions (a process called water
splitting). The hydrogen ions are adsorbed into the cation resin, displacing other
cations. Similarly, the hydroxyl ions are adsorbed into the anion resin, displacing
other anions. The displaced ions are free to migrate to the next ion exchange site
under the influence of the applied electric field (see discussion above). This process
of displacing cations and anions with hydrogen and hydroxyl ions is an ongoing
regeneration.

The key difference between EDI and conventional ion exchange systems lies in the
method used for resin regeneration. Instead of a batch regeneration with acid and
caustic, EDI regeneration occurs continuously with ions made by splitting water
molecules in place.

34a_m8_r0 8-11 Electro-Deionization

table of contents
 Figure 8.1-12: Typical view of a Mixed Bed

34a_m8_r0 8-12a Electro-Deionization

table of contents
 8.1.2 EDI Performance

The ion exchange resin closest to the dilute feed is in the unregenerated (partially
exhausted) state. Ions are taken up by the resin and removed by being transferred
towards the membrane. At the outlet end of the resin bed there are insufficient ions
remaining in the water to sustain the applied current. Here the DC electrical energy
enables the splitting of water into hydrogen (H+) and hydroxyl (OH-) ions,
maintaining the resin in a regenerated state. For removal of weakly charged ions,
such as silica, it is essential to use sufficient electric current to maintain the resin in a
highly regenerated state.

8.1.2.1 EDI Compared to Conventional Mixed-Bed Deionization

EDI and mixed bed vessels normally occupy the same position in a process: after a
reverse osmosis unit. A comparison of some of the major differences provides some
direction on which one is applicable in any given circumstance.

Mixed bed ion exchange technology uses resins that act like a sponge to remove
ionizable contaminants, and has been the industry standard in deionization for more
than 30 years. Figure 8.1-12 on the opposite page shows an illustration of a typical
mixed bed. As these resins become saturated with ions, they are restored in a batch
regeneration process using chemicals (strong acids and bases). Regeneration pro-
duces significant volumes of wastewater and requires operator attention to ensure a
successful result.

Continuous electrical regeneration of the resin in an EDI system eliminates the need
for the handling or consumption of aggressive regeneration chemicals. EDI also
eliminates the capital cost of the chemical regeneration equipment including special
alloy valves, pipe, pumps, lined day tanks, instruments, bulk storage tanks, forward-
ing pumps, interconnecting pipe, neutralization tank and related components, as
well as the costs of installation, continual refilling and ongoing maintenance of these
components.

EDI may benefit from periodic chemical cleaning if the upstream RO system is not
operated properly or if feedwater quality is outside of design limits. If an upset
occurs, EDI internal components are designed to withstand periodic chemical
cleaning.

EDI may also require a small chemical feeder for brine injection if the feedwater
quality is too good.

34a_m8_r0 8-12 Electro-Deionization

table of contents
 Regeneration

Continuous regeneration of EDI stacks replaces batch regeneration of mixed bed

vessels. Individual stacks can also be removed for cleaning, while the remaining
stacks accept short-term higher flow rates.

Startup and Operation

The continuous regeneration process requires minimal operator intervention, com-

pared with the batch regeneration process of mixed bed systems.

EDI operation requires very few automated valves or complex control sequences that
might need the attention of an operator. Instrumentation and controls are relatively
simple.

Product Quality

EDI product quality approaches 18 Mohm/cm resistivity at design conditions. The

product quality is not sensitive to inlet feed fluctuations or spikes, nor does it decay
over time as in the exhaustion of traditional deionization units.

Reduction of both SiO2 and ionizable TOC in an EDI system is comparable to results
achieved with primary mixed-bed operation. However, conventional mixed-bed
demineralizers currently provide a slightly higher product quality and have less
stringent feedwater requirements.

Feedwater Requirements

Mixed bed vessels generally have no restrictions on feedwater quality, except those
imposed by service cycle run time (premature exhaustion).

EDI system performance is subject to limitations on the feedwater TDS and on

hardness and weakly ionized compounds (carbon dioxide and silica).

34a_m8_r0 8-13 Electro-Deionization

table of contents
 Waste Generation

EDI recovery rates from 90 to 95% are typical. This is limited by the onset of scaling
in the C chamber. The “C” waste stream, however, can be reused in other unit pro-
cesses (recycle to reverse osmosis feed for example), providing an effective recovery
rate approaching 99%.

Mixed-bed vessels have recovery rates that cover a wide range, depending on the
regeneration frequency. A vessel regenerated once every 24 hours has an effective
recovery rate in the range of 95 to 98%. Although no continuous waste stream is
generated, a large peak in waste is produced during the regeneration process. Beds
with low feed TDS have higher effective recoveries due to less frequent
regenerations.

Footprint

EDI Systems have a similar footprint to conventional mixed-bed ion exchange

vessels of similar flow handling capacity. Considerable space savings are realized
with EDI systems since regeneration, storage, and neutralization equipment are not
required.

34a_m8_r0 8-14 Electro-Deionization

table of contents
 8.2 Process and Operations Overview

The basic principle of Electro-Deionization (EDI) is to treat water by passing it

through a small chamber of both cation and anion resins. The ions in the water are
removed as in conventional mixed bed ion exchange technology. However, the resin
in an EDI system is continuously regenerated using a current supplied by an applied
electric field.

An EDI system is designed to be used at the “clean” end of a typical process. Water
to be treated by EDI is typically reverse osmosis product water. The main
application of EDI equipment is in place of primary mixed beds.

The E-Cell™ EDI system has overcome some of the limitations of previous
incarnations of the technology. It is a system consisting of an arrangement of E-Cell
Stacks™ on a modular frame.

The E-Cell Stack™ is a compact robust module designed to create high purity water
using EDI technology. Figure 8.2-1, below, is a model of a 150 gallons per minute
E-Cell System™.

Figure 8.2-1: 150 GPM E-Cell System™

34a_m8_r0 8-15 Electro-Deionization

table of contents
 8.2.1 Process Flow

To understand the process flow of a typical E-Cell System™, we must look at how
the water to be treated flows through the system. A piping and instrumentation
diagram (P&ID) of a base model E-Cell System™ is shown in Figure 8.2-2 to help
illustrate a basic E-Cell System™ layout.

Looking at the specific system illustrated in this P&ID, we can see that it contains
eight E-Cell Stacks™. Each E-Cell Stack™ is rated to produce up to 12.5 GPM,
therefore this system has a capacity of 100 GPM. To aid in visualizing what the
system in this P&ID might actually look like, a picture of a 100 GPM E-Cell
System™ is shown in Figure 8.2-3, below.

Figure 8.2-3: Picture of a 100 GPM E-Cell System™

34a_m8_r0 8-16 Electro-Deionization

table of contents
 The E-Cell System™ is fed by product water. This feedwater is also known as the
Dilute feed. It enters the E-Cell System™ through the dilute feed inlet valve.

A pipe manifold distributes the Dilute feed flow to each of the E-Cell Stacks™.
Figure 8.2-4, below, is a view of the rear of the system with the Dilute feed pipe
manifolds highlighted.

Figure 8.2-4: Dilute Feed Pipe Manifolds

34a_m8_r0 8-17 Electro-Deionization

table of contents
 The deionized water is labeled Dilute product. As the Dilute product exits the stack
Dilute outlet connection, it enters the Dilute manifold. This manifold collects the
Dilute product from every stack and combines them into one common pipe.
Figure 8.2-5 shows a model of the rear of the system with the Dilute product
manifold indicated.

Figure 8.2-5: Dilute Product Manifold

Before the Dilute product leaves the system, a pressure gauge checks for proper
pressure, and a flow switch checks for proper flow. The Dilute product leaves the
skid through the Dilute Product Outlet Valve to be distributed to the desired points.

Now that it is understood how the Dilute water is fed into the system and flows out
of the system as ultrapure water, let’s look at what happens to the rejected ions in
the Stack. As mentioned before, as the Dilute water feeds through the Dilute
chamber, the Electro-Deionization process causes unwanted ions to flow through
the resin and permeate the membranes. Once these unwanted ions pass through
the membranes, they enter the Concentrate or “C” chamber. This resulting unwanted
product is known as the concentrate.

34a_m8_r0 8-18 Electro-Deionization

table of contents
 As seen in Figure 8.2-6, below, the concentrate is constantly recirculated through the
Stack by means of a recirculation pump.

Concentrate
Loop

Concentrate
Loop

Concentrate
Loop
Figure 8.2-6: Concentrate Loop

Remember that Dilute water enters the Stack through the dilute chamber, and the
contaminant ions migrate through the resin, permeate the membrane, and enter the
Concentrate chamber. The first flow path these ions follow after they enter the
Concentrate chamber is to exit the Stack. Therefore, let’s start from the point where
the concentrate flows out of the Stack.

34a_m8_r0 8-19 Electro-Deionization

table of contents
 Concentrate flows out of the top of the Stack through the concentrate outlet located
behind the stack. Once it leaves the concentrate outlet, it enters the concentrate
outlet manifold. Figure 8.2-7 on the opposite page is a picture of the rear of the
E-Cell System™ with the concentrate outlet manifold indicated.

Figure 8.2-7: Concentrate Outlet Manifold

34a_m8_r0 8-20 Electro-Deionization

table of contents
 This manifold receives all of the concentrate flows from each stack and combines
them into one common pipe. This pipe, or concentrate outlet, begins a loop that
allows the concentrate to flow back into the Stacks. A pressure gauge indicates the
pressure inside the concentrate outlet pipe. Included in the concentrate loop is a
recirculating pump, which is located on the rear of the system. This pump is used to
recirculate the concentrate flow back through the stacks. Figure 8.2-8, below, is a
model of the recirculating pump, also known as a concentrate pump.

Figure 8.2-8: Concentrate Pump

The concentrate pump boosts the concentrate pressure and sends it back to the
stacks. A percentage of this concentrate flow is allowed to bleed through the con-
centrate bleed valve and drain to waste. The purpose of this bleed is to prevent a
build up of ions that could lead to scaling in the Concentrate chambers. A mass
balance dictates when the ions removed from the feedwater must be discharged
from the system through the concentrate (and electrolyte) bleed streams. A flow
switch located on the concentrate bleed line measures the flow.

34a_m8_r0 8-21 Electro-Deionization

table of contents
 A small amount of Concentrate make-up water is added to the concentrate loop to
restore the drained concentrate. The make-up water comes from the Dilute feed
line.

Finally, the remaining concentrate flow is distributed through the concentrate inlet
manifold to each stack’s concentrate inlet.

The concentrate is recirculated, so that a sufficiently high conductivity can be kept in

the C chambers in the stack. If the conductivity is low, the resistance is high, and it is
difficult to maintain a sufficient current through the stack. However, too little con-
centrate bleed may allow too high a concentration of scaling ions in the stack.

The remaining outlet on the E-Cell Stack™ is the Electrolyte flow to waste, as seen
in Figure 8.2-9, below. The electrode chambers are fed by the concentrate feed
stream. These chambers receive the unwanted ions as they migrate and pass
through the membranes, just as the concentrate chambers do. The electrode cham-
bers contain the electrodes that supply the current to drive the EDI process. The
electrodes produce gases that escape the electrolyte chamber through the electrolyte
outlet to a vented drain. The electrolyte outlet must be sent to a vented drain to
avoid possible chlorine attack on the membranes and resin.

Electrolyte to
Waste

Figure 8.2-9: Electrolyte Flow

34a_m8_r0 8-22 Electro-Deionization

table of contents
 8.3 Equipment Design and Options

Each E-Cell™ EDI skid is designed as a stand-alone unit. The small, built-in PLC
performs the necessary control functions and, if desired, can tie into a main control
system. Figure 8.3-1 below is a photo of an E-Cell System™.

E-Cell Systems™ are modular in nature so that a system can be easily designed in
12.5 gpm increments. Consideration should be given to the future build-out of each
system to facilitate timely and cost effective growth. Areas to review include the
frame design for the addition of E-Cell Stacks™, floor space and headroom
availability, instrument ranges, pipe and valve sizes, pump size, motor starter size
and rectifier size.

Figure 8.3-1: E-Cell System™

34a_m8_r0 8-23 Electro-Deionization

table of contents
 8.3.1 E-Cell Stack™

The stack is the heart of the system. Each E-Cell Stack™ is designed to handle a
maximum product flow. The number of stacks needed for a system depends on the
desired total flow for the system. Systems are modular in nature and can be de-
signed in one stack increments.

The maximum working pressure for the E-Cell Stacks™ is 100 psig. If there is a risk
of over-pressure, a pressure relief valve should be installed between the EDI equip-
ment and any supply pumps.

Figure 8.3-2 below shows the rear of an E-Cell Stack™ and indicates the different
connections available.

Electrolyte Outlet

Concentrate
Outlet

Dilute Outlet
Power
Supply

Concentrate Dilute Inlet

Inlet

Figure 8.3-2: E-Cell Stack™

34a_m8_r0 8-24 Electro-Deionization

table of contents
 8.3.2 Control Panel

The Control Panel houses the dedicated PLC, Control Buttons/Switches, Meters,
Indicating Lights and the Operator Interface. Figure 8.3-3, below, is a picture of the
front of the control panel for an E-Cell System™.

Figure 8.3-3: Control Panel for an E-Cell System™

8.3.2.1 PLC

All automatic control functions associated with the operation of the equipment are
carried out by a dedicated programmable logic controller (PLC) complete with local
I/O, logic, and arithmetic controller functions capable of monitoring all digital signals
and controlling the process. Operating and alarm status and remote start/stop
signals are available for interface to the main water treatment control system.

34a_m8_r0 8-25 Electro-Deionization

table of contents
 8.3.2.2 Operator Interface

The Operator Interface is directly connected to the PLC. This Interface is designed to
allow the Operator to review data, system status, alarms, and other information.
Making all of this information available to the Operator allows safer and more
efficient operation of the system. Operators can also set parameters from this Inter-
face. Figure 8.3-4 is a picture of an Allen Bradley Microview Interface, commonly
used for the E-Cell Systems™.

Figure 8.3-4: Allen-Bradley Microview Interface

34a_m8_r0 8-26 Electro-Deionization

table of contents
 8.3.3 Brine Injection System (Optional)

The Brine Injection System consists of a chemical pump that draws concentrated
brine from a measuring tank and injects it into the Concentrate loop. Low
conductivity feedwater, which is normally advantageous, may not be conductive
enough to be used for makeup of the concentrate loop. The low conductivity
feedwater does not allow the stack to operate at the required current. Adding Brine
to the concentrate maintains a suitable Concentrate loop conductivity, enabling
higher currents to be used and resulting in better quality water. Figure 8.3-5, below,
is a picture of a commonly used chemical metering pump.

Figure 8.3-5: Brine Pump

When making the brine, care must be taken to use salt without a significant amount
of hardness minerals. Hardness minerals cause scaling in the E-CellTM concentrate
chambers.

34a_m8_r0 8-27 Electro-Deionization

table of contents
 8.3.4 Diverter Valves (Optional)

Automatic diversion of poor quality product water can be provided by adding

diverter valves to the Product Outlet and Rinse Outlet lines. If the product resistivity
is within specification, the Product Outlet Valve is open and the Rinse Outlet Valve is
closed: product water continues to the service outlet. If the product resistivity is low,
the Product Outlet Valve closes and the Rinse Outlet Valve opens. This action sends
the unacceptable water to drain rather than out through the service outlet, prevent-
ing the use of unacceptable product water.

The Product Outlet Valve is Fail Closed and the Rinse Outlet Valve is Fail Open,
preventing possibly unacceptable product water from leaving the system in the event
of a system failure.

The solenoids controlling the valves have

manual over-rides that allow the operator to
“energize” the solenoid and force a valve.

8.3.5 Concentrate Pump

The Concentrate Pump recirculates concentrate

water through the stacks. A high velocity is
necessary in the concentrate chambers to
prevent laminar flow and resultant high ion
concentrations at the membrane surface. High
ion concentrations cause mineral scaling.

The Concentrate pump is constructed of

austentic stainless steel or non-metallic parts.
It must keep the concentrate loop flow within
the design flow range for the number of stacks
present. A multistage turbine design is both
space and energy efficient in this application.
It also has a flat enough curve to avoid over
pressurizing the system in the event of a dead-
head type failure.

Figure 8.3-6, on the right page, is a picture of a

typical pump used as a Concentrate Pump on
an E-Cell System™. Figure 8.3-6: Typical View
of a Concentrate Pump

34a_m8_r0 8-28 Electro-Deionization

table of contents
 8.3.6 Piping

The piping used for E-Cell Systems™ is non-metallic. Non-metallic materials

provide electrical insulation, which enhances system safety. Ten bar Polypropylene is
available for all sizes. Schedule 80 PVC is available for 200 to 300 gpm systems.

8.3.7 Valves

Valves used for E-Cell Systems™ are made of non-metallic materials that match the
piping in which the valve is installed (Polypropylene or PVC).

Several different types of valves are used in E-Cell Systems™. The decision of what
valve to use depends on the size of the system and what the valve is used for. For
valves that are used to control a flow rate, a diaphragm valve is needed. When these
valves are sized accurately, flow rates may be controlled by simply opening or closing
the valve. Other valves used are ball valves and butterfly valves. These valves are not
used to control flow rates; they are only used to completely open or close a line.
Table 8.3-1 below is a Valve Selection Table used for E-Cell Systems™.

Application Type Sizing

Ball
Dilute Feed Inlet Line Size
Butterfly >2"

Dilute Product Outlet Diaphragm Line Size

Dilute Rins e Outlet Diaphragm Line Size

Concentrate Bleed Diaphragm M us t be s ized accurately to control the flowrate.

Electrode Outlet Diaphragm M us t be s ized accurately to control the flowrate.

Pump Suction Ball Line Size

Size to accurately regulate the pump dis charge pres s ure

Concentrate Inlet Diaphragm
for the range of flows required for the s ys tem

Concentrate M ake Up Ball Line Size

Concentrate By-pas s Globe Line Size

Table 8.3-1: E-Cell™ Valve Selection Table

34a_m8_r0 8-29 Electro-Deionization

table of contents
 8.3.8 Electrical Power Rectifier

The Electrical Power Rectifier is included with the E-Cell System™. It takes three
phase AC power from the plant and converts it to a settable DC voltage. The unit
includes appropriate power filtering and distribution to all stacks.

The rectifier has a control panel with an analog volt meter, a current amp meter, a
voltage adjustment knob, and a current adjustment knob. The adjustment knobs
allow the operator to increase or decrease the voltage or current as required for the
process.

Figure 8.3-7 shows a picture of the inside of a power rectifier cabinet. Figure 8.3-8
is a picture of the rectifier control panel.

Figure 8.3-7: Power Rectifier Figure 8.3-8: Power Rectifier

Internal View Control Panel

34a_m8_r0 8-30 Electro-Deionization

table of contents
 8.3.9 Instrumentation

Pressure gauges are used with the E-Cell Systems™ to verify pressures present in
the system. Stack inlet and outlet pressures for dilute and concentrate streams must
be set within acceptable ranges in order for the system to operate properly.

Flow meters are necessary for operators to properly set the required flows for the
system.

Flow switches are used to interlock the operation of the system to ensure that
incorrect operation does not damage components.

Figure 8.3-9 is a picture of flow meters with flow switch used for an E-Cell
System™.

Figure 8.3-9: Flow Meters Used for the E-Cell System™

34a_m8_r0 8-31 Electro-Deionization

table of contents
 8.4 Applications and Design

The E-Cell™ EDI process is normally employed in place of traditional mixed-bed ion
exchange. A typical system consists of feedwater pretreatment followed by Reverse
Osmosis (RO) and EDI.

Upstream Design Considerations

1. E-Cell™ feedwater pretreated by Reverse Osmosis, is fed to the E-Cell System™ in

one of two ways – either directly by the RO discharge or by a pump downstream of
an RO product storage tank.

2. Care must be taken to ensure that the feed pressure cannot exceed design pressure
under any conditions (including pump dead head).

3. A UV system can be installed upstream of the E-Cell™ to minimize the possibility of

biofouling.

4. Care must be taken to ensure that chemical dosing systems upstream of the
E- Cell™ do not introduce unacceptable chemicals into the E-Cell™ feed.

5. A softener may be installed upstream of the RO to reduce the E-Cell™ feedwater

hardness to specified levels.

6. The E-Cell System™ is not designed to handle suspended solids in the feed stream.
Appropriate filtration (preferably membrane) must be provided.

Downstream Design Considerations

The E-Cell™ unit can send product water directly into a tank, into other treatment
equipment before a tank, or into a pressurized distribution loop. Care must be taken
to ensure that operating pressure requirements are not violated at any time.

For applications requiring 18 Mohm-cm, water a polishing mixed-bed deionizer can

be used downstream of the E-Cell™. This type of deionizer typically uses
non-regenerable resin, with resin replacement cycles exceeding one year.

34a_m8_r0 8-32 Electro-Deionization

table of contents
 8.4.1 E-Cell™ Process Design

The process design procedure is used for systems producing product resistivity
greater than 16 Megohm-cm based on the feedwater parameters given in
Table 8.4-1 below. While the use of lower quality feedwater to yield lesser product
resistivity may be of interest in some circumstances, E-Cell™ technology is not
recommended for such applications. Operation on the knee of the performance
curve may produce product quality that is very sensitive to other parameters. Lower
quality feedwater is defined as water with higher levels of Total Exchangeable Anions
(TEA) / Total Exchangeable Cations (TEC), but strictly meeting the other E-Cell™
feedwater requirements for hardness, silica etc.

Constituent Units Feed

TEA ppm CaCO3 <25
TEC ppm CaCO3 <25
Conductivity uS/cm <65
pH 5 to 9
Hardness ppm CaCO3 <1.0
Silica (reactive) ppm <0.5
TOC ppm <0.5
Free Chlorine ppm <0.05
Fe, Mn, H2S ppm <0.01
SDI 15min. <1.0
Oil & Grease None detectable
Turbidity NTU <1.0
Oxidizing agents None detectable

Table 8.4-1: Feedwater Operating Parameters

8.4.2 Feedwater Analysis and E-Cell™ Feedwater Specification

A representative E-Cell™ feedwater analysis is necessary. This is usually an RO

projection based on the raw water. Analysis for some contaminants such as carbon
dioxide can be performed in the field or calculated from bicarbonate concentration
and pH. The usual care must be taken to account for seasonal or other variations in
feedwater quality.

The worst-case water analysis is compared with the E-Cell™ feedwater specifica-
tions given above. All of the stated specifications must be met at all times for the
specified product water quality to be achieved and to avoid maintenance problems.

34a_m8_r0 8-33 Electro-Deionization

table of contents
 8.4.2.1 Anion and Cation Loading

Calculate the separate anion (including carbon dioxide and silica) and cation
(including NH3) loadings expressed as ppm CaC03 (see Appendix J for conversion
to ppm as CaCO3). Cation and anion loading must both be less than 25 ppm as
CaCO3.

8.4.2.2 Calculation of Maximum Allowable Recovery

The maximum rated recovery of an E-Cell™ unit must be strictly adhered to, in order
to avoid the possibility of scaling by hardness and silica. Upsets in feedwater quality
must also be monitored to prevent scaling. Although moderate hardness scaling can
be successfully treated with a cleaning procedure, it should nonetheless be avoided.
Precipitated silica cannot be removed and must also be strictly avoided.

The recovery must never be increased in order to increase the conductivity in the
Concentrate loop. The Concentrate Bleed flow is set to achieve the required recovery
rate. If the Concentrate Bleed flow is decreased, the maximum allowable recovery
may be exceeded, and scaling of the Concentrate chamber will begin. Scaling occurs
because the Concentrate loop becomes enriched in scale-forming ions.

The formula to calculate the recovery rate is:

Product + Concentrate Bleed + Electrode Flows

Recovery = Product Flow X 100%

The maximum allowable recovery (amax), can have one of two values, depending on
the feedwater hardness:

amax = 90% recovery if feed hardness in less than 1.0 ppm as CaCO3
amax = 95% recovery if feed hardness is less than 0.1 ppm as CaCO3

34a_m8_r0 8-34 Electro-Deionization

table of contents
 Parameter Value Range Status
Ion ppm as CaCO3
Target Resistivity 17 0.02-17 MOhm cm O.K.
Design Flow (gpm) 100
Ca 0.01 Normal Flow (gpm) 75 50%(Max)-Design O.K.
Temperature (C) 20 5.0-40.0 (deg. C) O.K.
Mg 0.01
Na 5.4
K 0.0
NH4 0.0
Ba 0.0
Sr 0.0 Calc.
Parameter Range Status
Value
CO3 0.0
HCO3 0.0 TEA ( ppm as CaCO3 ) 5.5 0.0-25.0 O.K.
SO4 0.0 TEC ( ppm as CaCO3 ) 5.4 0.0-25.0 O.K.
Cl 5.4 CO2 ( ppm as CaCO3 ) 0.0 0.0-25.0 O.K.
Alk. ( ppm as CaCO3 ) 0.0 0.0-20.0 O.K.
F 0.0
Hardness ( ppm as CaCO3 ) 0.0 0.0-1.0 O.K.
NO3 0.0 Silica ( ppm SiO2 ) 0.1 0.0-1.0 O.K.
SiO2(ppm) 0.1 Cond ( ppm uS/cm ) 13.7 0.0-62.0 O.K.
pH 7.0 Charge Balance 0.23% (+/-)5% O.K.

Product Flow Rate (gpm) 100 75 Comments

Product Resistivity (MOhm cm) 17 17 - 16 MOhm cm guaranteed

Silica in the Product (ppm SiO2) 0.015 0.015
Maximum Allowable Recovery 95% 95% -based on feed hardness andCO2 levels
#of E-cell Stacks 8 8 -individual stack design flow = 12.5 gpm
Current (Amps) 12 12 -Estimated values!
Voltage (Volts) 360 360 -Rectifier must be capable of 600V!
AC Power (kW) 4 4
AC energy (kWh/1000 gal) 0.7 0.7
Salt Injection (YES or NO)? NO
NO NO
NO -If C loop cond.< 150 or Voltage =550 then salt injection.
Salt Inject tion Rate (gpd) 0.00 0.00 -Saturated NaCl brine is to be used as salt injection feed.
C-loop conductivity (uS/cm) 274 274
C-loop TDS (ppm as NaCl) 137 137
Feed-inlet (gpm) 105.3 105.3
Concentrate Bleed (gpm) 3.3 3.3
E-outlet (trace Cl2) (gpm) 2.0 2.0

Table 8.4-2: E-Cell™ Calc Spreadsheet Sample

34a_m8_r0 8-35a Electro-Deionization

table of contents
 8.4.3 Hydraulic Sizing

Each E-Cell StackTM has a normal capacity of 12.5 gpm of product flow. Due to the
robustness of the unit design, this flow rate can be exceeded for short periods but
should not be exceeded on a continuous basis. The number of stacks required is
calculated by dividing the desired product flow rate by the normal stack capacity.

The feed flow is calculated by dividing the product flow by the recovery (as selected
in the previous section). The electrolyte flow is fixed at .25 gpm per stack. The total
electrolyte flow is the flow per stack, times the number of stacks. The concentrate
bleed flow is calculated using:

Bleed = Feed flow – Product flow – Electrolyte flow

Or
Bleed = (100/recovery –1) x Product flow – Electrolyte flow
(Where recovery is expressed as a percentage)

8.4.4 Brine Injection For Concentrate Loop Management

To achieve the target effluent water quality, the conductivity of the water in the
Concentrate Loop must be sufficient to pass the required current. This may require
the injection of small amounts of NaCl brine into the Concentrate Loop. This is
specifically necessary when there is low feed conductivity and low recovery, and
particularly if silica removal is important.

NOTE: In determining the need for salt injection, the lowest expected feedwater
conductivity must be used. It is under these conditions that the amount of ions
transferred from the feedwater in the Dilute Chambers, through the membranes and
into the Concentrate Loop, is at a minimum.

Salt injection IS NOT required if:

· The feedwater conductivity > 5mS/cm AND the recovery is 95%

and no CO2 is present

· When CO2 is present, use E-Cell™ Calc to determine the need for
salt injection.

E-Cell™ Calc is a spreadsheet that performs the design calculations given inlet
conditions. Table 8.4-2 on the opposite page is a sample E-Cell™ Calc spreadsheet.

34a_m8_r0 8-35 Electro-Deionization

table of contents
 8.4.5 Current Determination and Rectifier Sizing

Product resistivity as a function of feed composition and applied current is shown

in Figure 8.4-3 below. For each curve corresponding to a particular applied
current, increasing the feed concentration eventually results in a decrease in product.
resistivity. Determination of the required operating current and sizing of the rectifier
are key steps in E-Cell™ process design.

18

16

14
Product Resistivity (Mohms-cm)

12

10

8
1.0 Amps

6 2.0 Amps

4 3.0 Amps

4.0 Amps
2

5 10 15 20 25 30 35 40
Total Exchangeable Anions (ppm CaCO3)

Table 8.4-3: Product Resistivity vs. Feed at Varying Currents (NaCl Feed, 12.5 gpm)

34a_m8_r0 8-36 Electro-Deionization

table of contents
 4.5

4
Current (Amps/stack)

3.5

2.5

1.5
5 10 15 20 25
Total Exchangeable Anions (ppm CaCO3)

>17MO.cm 16MO.cm 14MO.cm 12MO.cm

10MO.cm 8MO.cm 6MO.cm
(Rectifier Must Be Capable of 1.5 Amps Minimum)

Table 8.4-4: Current Requirements for Various Feed and Product Waters

34a_m8_r0 8-37a Electro-Deionization

table of contents
 Figure 8.4-4 on the opposite page shows the relationship between the Total
Exchangeable Anion (TEA) concentration in the feedwater and the minimum
operating current. Each curve corresponds to a target product resistivity.

Follow the steps below in using Figure 8.4-4 to size the rectifier:

1. Determine feedwater TEA:

Use the projected “worst case”TEA value for the feedwater (e.g., the 3-year
R.O. projection).

2. Select the correct current-TEA curve:

If effective silica removal is required, then the curve in Figure 8.4-4 labeled “>17
mohm-cm” must be selected. If silica removal is not important, then select the curve
in Figure 8.4-4 corresponding to the target product resistivity. If a curve is not
shown for the specific resistivity of interest, use the curve for the next highest resis-
tivity.

3. Current determination:
Using the feedwater TEA and the curve determined in steps 1 and 2, read off the
minimum required current from the vertical axis in Figure 8.4-4. This minimum
operating current includes a safety factor. A maximum of 4.5 Amps per stack is
permitted without damage.

4. Calculate total current:

Multiply the current from step 3 by the number of stacks to get the total system
current. A 600 Volt rectifier is required.

Note: For effective silica removal, highly regenerated anion resin is required. This
requires the selection of an E-Cell™ current that results in high resistivity product
water, regardless of the minimum target product resistivity.

Note: Gases (hydrogen, oxygen, trace chlorine) from water electrolysis are produced
at the electrodes and are carried away in the electrode outlet stream. A 2:1 mixture
of hydrogen and oxygen gases is produced at a rate of approximately 4 cm3/Amp/
minute (dry gases at 25°C, 1.0 atm). Provisions should be made for venting this gas.

34a_m8_r0 8-37 Electro-Deionization

table of contents
 8.4.6 Concentrate Pump Sizing

The Concentrate Pump is sized to deliver 4.5 U.S. gpm per Stack and to increase the
Concentrate pressure by 40 psig plus the line losses from the pump discharge to the
Stack inlet.

For example, for a 100 gpm system, where eight (8) E-Cell Stacks™ are
required:

Concentrate Flow = 8 x 4.5 gpm = 36 gpm

Concentrate Boost Pressure = 40psig

+ line losses from the pump discharge
to the Stack inlet

34a_m8_r0 8-38 Electro-Deionization

table of contents