UPCORE System: Suspended Solids Removal For Countercurrent Ion Exchange Systems
UPCORE System: Suspended Solids Removal For Countercurrent Ion Exchange Systems
UPCORE System: Suspended Solids Removal For Countercurrent Ion Exchange Systems
Suspended Solids
Removal for Countercurrent
Ion Exchange Systems
Daniel B. Rice - The Dow Chemical Company, Midland, Michigan
Andre Medete - Dow Deutschland Inc., Rheinmuenster, Germany
Published in International Water - July 2002
IWC-97-13
Introduction
Ion exchange demineralizers are
designed and built to economically
remove ionic impurities from water.
Their design also makes them
unintentional filters for any suspend-
ed solids left in the feedwater after
pretreatment. The design and
operation of any demineralizer
system must include a mechanism
for periodically removing trapped
suspended solids from the ion
exchange resin bed. This paper
discusses the mechanism of
suspended solids removal with
traditional pressure filters and
cocurrent ion exchange systems.
It also reviews operational changes
required to compensate for suspend-
ed solids removal in the operation of
various countercurrent demineralizer
system designs.
Background
Ion exchange resin development
in the 1940s allowed the efficient
removal of ionic impurities from
water by reversibly exchanging
acceptable ions for undesirable
ions. The system design incorporat-
ed for these early cocurrent systems
used downflow (in the direction of
gravity) service and downflow regen-
eration. This design was simple and
it worked effectively from an ion
exchange standpoint. However, ion
exchange vessels are also effective
filters and will remove particulate
matter that is present in the feed-
water. To cleanse the bed of accu-
mulated suspended solids, the resin
was backwashed when it was taken
out of service for regeneration.
Incorporating a backwash step to
the ion exchange regeneration
procedure was a logical approach.
The designs of pressure sand filters
and cocurrent ion exchange vessels
have a number of similarities and a
backwash step was already being
used for pressure sand filters.
Raw Water Inlet
Top Baffle
Pressure Filter
Strainer
Heads
Filtered
Water
Outlet
Laterals
Concrete
Sub-Fill
3'
50%
Freeboard
Coarse Support Layer
Filter Media
Raw Water Inlet
Upper Lateral
Ion Exchange Softener
Strainer
Heads
Soft
Water
Laterals
3'
50%
Free
Graded Quartz
Ion Exchange
Resin
Comparison of Sand Filter &
Cocurrent IX System Operation
The rapid-sand filter used in the
early 1900s contained sand of
0.4 - 0.5 mm effective size in beds
roughly 3 feet deep, and operated at
a 2 - 4 gpm/ft
2
filtration rate.
1,2
Ion exchange systems, such as the
softener shown in Figure 1,
contained resin of 0.4 - 0.5 mm
effective size in beds 30 inches or
more in depth and operated at
flowrates of 2 gpm/ft
3
(5 gpm/ft
2
for
a 30 inch bed depth).
These sand filters were operated
to a P endpoint, typically around
12 psi. Run times normally ranged
from 12 - 36 hours.
1
The material
trapped in the filter was largely held
in the top 1 to 4 inches of the bed.
To clean this matter from the filter, a
reverse flow of water was used to
fluidize the filter media and wash the
accumulated solids from the bed.
The backwash time was 15 - 25
minutes at flowrates sufficient to
expand the bed up to 50%.
Ion exchange vessels were operat-
ed to an exchange capacity endpoint
but also were designed to have run
times of around 24 hours.
Suspended solids present in the
feedwater are also filtered out at the
Figure 1. Design Comparisons.
Pressure Filters and Ion Exchange Softeners
top of the bed. Backwash times
of 10 - 20 minutes at flowrates
sufficient to expand the bed 50%
or more were used to flush
accumulated solids from the bed.
Blocking Flow Countercurrent
Systems
The chemical dosage requirements
of conventional cocurrent ion
exchange demineralizers are typical-
ly 200 to 400% of stoichiometeric.
For every mole of ions captured, 2
to 4 moles of regenerant chemical is
required. The dosage rates are nor-
mally elevated to these levels to
meet the water quality constraints of
the system.
Although an ion exchange
system can be operated much more
efficiently if the regenerant chemical
is introduced to the ion exchange
bed in the opposite direction to the
service flow, countercurrently, it is
more difficult from an operational
standpoint. To achieve the advan-
tages of countercurrent regeneration,
the resin bed must stay classified.
The most highly regenerated resin
must stay at the service outlet end
of the vessel. The resin most
completely exhausted must remain
at the service inlet end of the vessel.
Suspended Solids Removal for Countercurrent Ion Exchange Systems
Blocking flow demineralizer
designs solve this challenge by
introducing a blocking flow of either
air or water during regeneration that
keeps the bed from expanding. An
intermediate lateral placed several
inches below the top of the resin bed
collects both the regenerant chemi-
cal and the blocking flow (see Figure
2). Although both external piping
and support equipment are different,
the basic vessel design is the same
as a cocurrent system. The direction
of service water is downflow through
a fixed bed of IX resins. Particulate
in the feedwater will be filtered out at
the top of the resin bed. When pres-
sure drop becomes elevated, typical-
ly every 10 - 20 cycles, the resin bed
is backwashed. The vessel contains
adequate freeboard to allow the
resin to be backwashed in-situ.
Resin backwash has been
shown to be effective in removing
particulate trapped in the resin bed if
allowed to expand 50% or more.
The backwash step also relieves bed
compaction that results from the
normal shrink-swell cycles that occur
when a resin is converted from one
ionic form to another. Regardless
of whether the increased pressure
drop across a resin bed is due to
compaction or accumulation of par-
ticulate matter, the backwash step is
effective at eliminating the cause and
preparing the bed for continued use.
However, performing a backwash
Figure 2. Blocking Flow Countercurrent Systems
Intermediate
Lateral
Freeboard
Service Feedwater
Product Water
Resin Bed
X
Regenerant
Waste
Blocking Flow
Regenerant
Resin Bed
disrupts bed classification. To pre-
vent a decrease in water quality in a
countercurrently regenerated ion
exchange system, a double regener-
ation is required before returning the
bed to service after every backwash.
Packed-Bed Countercurrent
Systems
Another approach to achieve the
advantages of countercurrent opera-
tion is to fill the entire vessel with
resin. If a bed is packed full of
resin, reversing the flow direction
cant disturb bed classification.
In operation it isnt practical to
keep a resin vessel filled at all times
because resins change size during
use. Another difficulty with packed
bed designs is the elimination of
freeboard space in the vessel makes
it impossible to backwash (expand)
the resin in situ. Cleaning particulate
and resin fines from the bed must be
addressed in some other way. The
solution to these difficulties has been
handled in two different approaches.
The difference between the two
approaches stems from the choice of
whether to run service water upflow
or downflow.
Upflow Service Fluidized
Packed Beds
A small amount of void space is left
in a packed bed to compensate for
volume changes that occur when a
resin is converted from one ionic
form to another. The change in the
size of the resin varies from a few
percent for strong acid cation resin
up to 20% or more for a strong base
anion resin. Actual volume change
depends on feedwater water compo-
sition, capacity utilization, etc. In all
cases, some freeboard will exist in
the vessel at some point in the ser-
vice cycle. This freeboard space will
move from one end of the vessel to
the other when flow direction
changes (see Figure 3). In a proper-
ly designed system, the movement
of the freeboard is not significant
enough to disrupt classification of the
resin. The small amount of shifting
in the position of individual resin
beads has negligible impact on the
ion exchange performance of the
vessel. The shifting of the resin is
actually a benefit in that it releases
any compaction that exists in the
resin bed.
Direction
of Flow
Resin Bed
Freeboard
Figure 3. Movement of Freeboard from Bottom to Top
These systems are commonly
referred to as fluidized packed
beds because the lower portion of
the bed (the service inlet end) is
partially fluidized during service.
The percentage of the bed fluidized
will vary dependent on a number of
factors including flow rate, particle
size, distribution of flow, water
temperature, and vessel freeboard.
Upflow, fluidized bed systems are
completely different from earlier
IX system designs in how they
filter/accumulate particulates in the
feedwater. The resin beads at the
inlet end of the vessel during service
are fluidized. They do not have the
ability to trap suspended solids in the
pinch points between resin beads as
earlier ion exchange systems had.
Particulates are able to pass through
the fluidized resins, deeper into the
bed. As with blocking flow systems,
the particulates will accumulate over
time and the pressure drop across
the bed will increase.
However, the increase in pressure
drop is more gradual in the fluidized
bed systems than in blocked flow
systems. One reason is that sus-
pended solids are partially washed
from the bed during regeneration.
Even at the relatively slow flowrates
used in regeneration, and even with
most of the particulates trapped
deep in the bed, some suspended
solids are removed during the initial
minutes of regeneration. A recent
IWC paper
3
reported the amount of
suspended solids removed during
regeneration was about 7% of the
suspended solids collected during
the service cycle. Most of the solids
were removed in the first three min-
utes of regeneration. The second
reason that pressure drop increases
more gradually in fluidized bed
systems is because deep bed filtra-
tion distributes suspended solids
throughout the entire vessel. In
blocking flow systems, most solids
accumulate within the top few inches
of resin, plugging the void space,
and causing a more rapid increase
in pressure drop.
However, the pressure drop does
continue to increase. After a period
that might be a few weeks or
perhaps several months or more, the
pressure drop reaches a point where
the vessel must be removed from
service. At this point, the resin is
sluiced from the service vessel to
the backwash vessel specifically
designed for cleaning the resins
(see Figure 4). The backwash step is
effectively the same as a backwash
in earlier system designs. The key
difference is the suspended solids
and resins fines are mixed through-
out instead of being concentrated at
the top of the resin bed. The back-
wash step is monitored (versus run
for a preset automated time period)
to determine when the bed is ade-
Service
Vessel
Transfer Port
Resin
Resin Transfer
Resin Return
Backwash
&
Sluice Water
Backwash
Tank
Particulates &
Fines to Drain
quately cleaned and ready to return
to service. The resin is then sluiced
back to the service
vessel.
As with the blocking flow system,
the amount of regenerant chemical
used is increased by 2 - 3 times
the normal dosage for the first
regeneration following a backwash.
The elevated chemical dosage is
required to return the polishing zone
of the bed to a high degree of regen-
eration and ensure expected water
quality is produced.
Figure 4. Fluidized Bed Backwash System
Downflow Service Packed Beds
The second approach for operation
of a packed bed system is to run the
service step downflow and regener-
ate the vessel upflow. This approach
insures the bed remains classified
during service. No fluidized resin
zone exists.
The challenge to operation with
this design is keeping the bed
compacted during regeneration.
As mentioned earlier, packed beds
all will have freeboard space left in
the vessel. Typical regeneration
flowrates are not sufficient to lift
the resin to the top of the vessel
and keep the resin classified.
This problem is resolved in the
UPCORE* system design, an upflow
countercurrent regeneration system
pioneered by The Dow Chemical
Company, by incorporating a com-
paction step in the regeneration
sequence.
Compaction
The first step in the regeneration
cycle of an UPCORE system uses
an upflow of water to compact the
resin bed against a layer of floating
inert beads at the top of the vessel.
The flow rate needed for compaction
is determined by the resin particle
size, particle density, amount of
freeboard, and water temperature.
It only takes a few minutes for
the bed to fully compact. Once
compacted, the resin bed remains
in place even if the flow rate is
reduced. This allows regeneration
to take place at the flow rate which
achieves the best results in terms
of regenerant contact time and
concentration.
As this design operates downflow,
through a fixed bed of resin,
particulates in the feedwater are
removed at the top of the bed as in
a conventional cocurrent system.
Penetration of solids into the bed is
mostly limited to the top few inches
of resin. During compaction, the
resin is pushed to the top of the bed
and compressed against a layer of
floating inert resin (see Figure 5).
This step prepares the bed for
chemical injection. Compaction also
flushes trapped particulate matter
from the bed.
The size of the inert beads and the
design of the upper laterals are
critical variables in the compaction
step. The inert beads must be small
enough to prevent the resin beads
from passing through, but large
enough to allow free passage of
particulates. The overall system is
designed so particulates and resin
fines can be freely flushed from the
vessel.
Figure 5. UPCORE System - Compaction
Resin Bed Resin Bed
Feedwater
Product Water Compaction Water
Particulates/Waste Water
Inert Resin
Freeboard
The depth of the inert layer is also
important. Too shallow of bed depth
may allow resin to penetrate, caus-
ing lateral blockage. Too deep a bed
depth could reduce the efficiency of
particulate removal. Surprisingly, the
proper inert resin depth is roughly
the same as the amount of freeboard
space left above the fluidized resin
bed in a cocurrent system during
backwash (see Figure 6). The
distance particulates must travel
from the top of the resin to the outlet
of the vessel is effectively the same
in both system designs.
Figure 6. Solids Removal Mechanisms
Backwash
Expansion
Compacted
Resin
CONVENTIONAL
Cocurrent Backwash
Flow = 1X
Flow = 3X
Freeboard
Inert Resin
Settled Bed
UPCORE
Compaction
Fines/
Contaminants
Freeboard
Linear Flow
1X
2.0X
2.5X
1.5 X
Linear Flow
9X
9X
*Trademark of The Dow Chemical Company
However, a key difference between
the backwash of a cocurrently
operated vessel and the compaction
step of an UPCORE system is the
flow rate of the cleansing flows.
The flowrate used during compaction
is significantly faster, as measured
by gpm/ft
2
, than used during a
conventional backwash. The higher
flowrates used during compaction
result in substantial increases in the
hydrodynamic shear forces across
the resin surfaces. The compaction
flow is effective at removing the
particulates attached to the surfaces
of the resin beads as well as the
matter held in pinch points between
resin beads.
Closer inspection of backwash and
compaction shows the difference in
hydrodynamic shear is more signifi-
cant than the ratio of their respective
flowrates. The void space between
the resin beads is roughly 33% of
the total bed volume. If the linear
flowrate through an empty vessel is
X gpm, the flowrate through a
packed bed of resin will be (1
0.33)X or 3X. The linear velocity
through a resin bed during regenera-
tion will have a linear velocity of 3X
at the start of backwash. However,
as the bed begins to expand, the
void volume increases, and the lin-
ear velocity decreases. At 100% bed
expansion, the linear velocity is 1.5X.
There is always some freeboard
space left above the expanded resin
to prevent resin loss. The freeboard
space above the resin will once
again have a linear flowrate of X.
The flowrate of X is the final velocity
for flushing particulate matter from
the vessel.
A packed bed vessel also has 33%
void space. The percent void space
remains constant during the com-
paction step because the bed is kept
packed. The flowrate throughout the
entire compaction step is three times
the linear velocity through an empty
vessel. It also has to be noted
compaction flowrates are faster than
used during backwash, up to three
times faster. So the actual flow
velocities at the top of the vessel
where the solids are flushed from the
vessel are up to nine times faster
during compaction than during
backwash (see Figure 6).
A profile of a backwash effluent
plotting concentration of suspended
solids versus time is shown in Figure
7.
4
This curve highlights two typical
characteristics. First, there is always
a time lag after the start of the
backwash step that varies with both
flowrate and percent freeboard.
The greater the freeboard, the
longer it will take to expand the
bed sufficiently to allow particulate to
begin to be flushed from the bed.
Second, fluidization of the bed during
backwash is responsible for the tail
of the curve The total time for back-
washing a cocurrent system where
the resin is backwashed every
regeneration ranges from 10 to 20
minutes. Suspended solids are not
totally removed at any time, so the
decision on when to stop the
backwash is somewhat arbitrary.
In practice, an occasional extended
backwash may be required to more
thoroughly cleanse the bed.
A profile of the effluent during com-
paction is shown in Figure 8. It plots
the same properties as Figure 7 but
the scale of units has been modified
by a factor of 10X for suspended
solids concentrations and by a factor
0.1X for time. The change in scale
was required to adjust for the more
rapid removal of suspended solids
achieved during compaction. This
profile differs in that the time lag
before suspended solids begin to be
washed from the vessel is much
shorter. It only takes a matter of
seconds for the resin bed to move to
the top of the vessel because there
is minimal freeboard space.
Figure 7. Backwash Effluent Profile Suspended Solids vs. Backwash Time
600
500
400
300
200
100
0
0 2 4 6 8 10 12 14 16 18 20
Backwash Time (min)
Suspended Solids
Concentration (mg/l)
Figure 8. Compaction Effluent Profile Suspended Solids
vs. Backwash Time
6000
5000
4000
3000
2000
1000
0
0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Compaction Time (min)
Suspended Solids
Concentration (mg/l)
Compaction has one additional
advantage over the conventional
backwash that warrants note. Most
systems are required to monitor the
backwash temperature and adjust
flowrate to compensate for the
change in bed expansion at different
water temperatures. In practice,
adjustments are generally belated or
ignored completely, resulting in inad-
equate backwash expansions and
poor removal of suspended solids, or
excessive backwash and loss of
resin that is flushed to drain. In con-
trast, only resin fragments, if any are
present, can be flushed from the
UPCORE system during compaction.
Settling
Following compaction, regenerant
injection, and slow rinse, the flow
through the UPCORE system is
stopped for a few minutes to allow
the resin to settle to the bottom of
the vessel. Settling does not occur
all at once with the entire bed drop-
ping to the bottom of the vessel as a
plug. Instead, the resin rains down
in a continuous fashion as a narrow
zone of fluidized resins moves up the
vessel. The movement of an individ-
ual resin bead, either up or down, is
not significant enough to detract from
the ionic stratification.
Settling has two critical affects on
system operation. First, it relieves
any compression of the resin created
during the ionic cycling of the resin
during regeneration. Secondly, it
helps in cleaning the bed. The
movement of the fluidized zone dur-
ing settling allows any resin frag-
ments or small particulates that are
located within the bed to move
upward. The movement in any one
cycle is small but the continuous
compaction and settling that occurs
with normal operation has a signifi-
cant affect. This has been demon-
strated by injecting dyed resin frag-
ments into a bed of resin several feet
from the top. Over a series of six
compaction/settling cycles, the
fragments moved to the top of the
bed where they were flushed from
the system.
5
Summary
The very nature of the way ion
exchange resins are most efficiently
used makes them effective filters of
suspended solids. Removal of
suspended solids from conventional
cocurrently regenerated demineraliz-
ers have proven effective at continu-
ously removing particulate matter.
The chemical efficiency and water
quality of countercurrent regenera-
tion is known to be superior over
cocurrent designs but the mechani-
cal aspects of operating countercur-
rently have certain barriers to over-
come. Countercurrent system
designs must compensate for bed
compaction that occurs with use.
They also must have some mecha-
nism for suspended solids removal.
Blocking flow system designs
operate downflow and rely on period-
ic in-situ backwashing to remove
trapped particulates and release
compaction. Stopping operation to
backwash is roughly required every
10 - 15 cycles. Upflow service
packed beds are regenerated down-
flow so using blocking flow during
regeneration is unnecessary. The
settling of the resin when direction of
flow changes releases bed com-
paction. Deep bed filtration reduces
the need for backwash to a period of
weeks or even months. When back-
wash is required, the system is shut
down and the resin is transferred to
a separate backwash vessel.
The UPCORE system is a down-
flow service, upflow regeneration
packed bed system. It utilizes a
compaction step to compress the
resin to the top of the bed as the first
step in every regeneration cycle.
Compaction effectively removes the
suspended solids trapped in the
upper part of the resin bed during
the service cycle. A separate settling
step relieves compaction of the bed.
It also lifts any resin fragments or
particulates located within the resin
bed toward the top of the vessel
where they can be removed during
compaction. Suspended solids are
continuously removed in every
regeneration so a separate back-
wash operation is not required.
References
1
Water Quality & Treatment,
Third Edition, The American
Water Works Association, Inc.,
McGraw-Hill Book Company, 1971.
2
Demineralization by Ion Exchange,
Samuel B. Applebaum, Academic
Press, 1968.
3
Williams, S., Cockerham, C.,
Economic Advantages of a Packed
Bed Retrofit, International Water
Conference, IWC-96-44, Oct. 21-23,
1996.
4
Pilot plant data shared with Dow
by an equipment manufacturer.
5
Internal laboratory test data,
The Dow Chemical Company.
*Trademark of The Dow Chemical Company Form No. 177-01749-702XQRP
Warning: Oxidizing agents such as nitric acid attack organic ion exchange resins under certain conditions. This could lead to anything from
slight resin degradation to a violent exothermic reaction (explosion). Before using strong oxidizing agents, consult sources knowledgeable in
handling such materials.
Notice: No freedom from any patent owned by Seller or others is to be inferred. Because use conditions and applicable laws may differ
from one location to another and may change with time, Customer is responsible for determining whether products and the information in this
document are appropriate for Customers use and for ensuring that Customers workplace and disposal practices are in compliance
with applicable laws and other governmental enactments. Seller assumes no obligation or liability for the information in this document. NO
WARRANTIES ARE GIVEN; ALL IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE ARE
EXPRESSLY EXCLUDED.
Published July 2002.
Dow Liquid Separations Offices.
For more information call Dow Liquid Separations:
Dow Europe
Dow Customer Information Group
Liquid Separations
Prins Boudewijnlaan 41
B-2650 Edegem
Belgium
Tel. +32 3 450 2240
Tel. +800 3 694 6367