F W R J

Reverse Osmosis Post-Treatment

Stabilization Utilizing Liquid Lime
Vaile Feemster and Jim Smith

T
he Dauphin Island Water & Sewer Author- prepared. The evaluation narrowed the prospective
ity (Authority), in Dauphin Island, Ala., treatment alternatives to two possible secondary Vaile Feemster is general manager with
owns and operates a reverse osmosis (RO) options of the primary CO2 /lime (slurry) dosing Dauphin Island Water & Sewer Authority in
drinking water treatment facility that first came on- option. The Authority pilot-tested the two selected Dauphin Island, Ala., and Jim Smith, P.E., is
line in May 2010. The facility serves Dauphine Is- methods and then selected the most advantageous
project manager with Constantine
land’s 1,200 permanent residents, and a seasonal option for poststabilization at its water treatment
Engineering in Fort, Payne, Ala.
tourist population of more than 20,000. The barrier facility.
island is located off the coast of the state, approxi- The Authority commissioned a new RO treat-
mately 30 mi south of Mobile. The facility treats ment facility in May 2011. From start-up, the facil-
water from a sand aquifer that is about 700 ft below ity has experienced problems with corrosive water feed water, including calcium and bicarbonate/car-
the ground’s surface. The water is relatively good and meeting the Alabama Department of Envi- bonate ions. The resulting RO permeate will typi-
quality and only requires treatment for chlorides in ronmental Management (ADEM) requirements cally have low levels of calcium hardness and
the 1,700-parts-per-mil (ppm) range. The facility for lead levels. The Authority initially implemented alkalinity and is “stabilized” to protect distribution
currently operates at a recovery rate of 75 to 80 per- a corrosion control program, held over from a pre- pipelines, pump stations, and storage tanks. The
cent (depending on the season) and can supply a vious iron removal plant, by treating the plant ef- Authority has attempted to provide stabilization
production of up to 1.2 mil gal per day (mgd). fluent with a proprietary blended with proprietary blended phosphates and pH ad-
The RO treatment of brackish water purifies zinc-orthophosphate. In August 2011, the Author- justments using sodium hydroxide. This approach
and significantly changes the mineral composition ity failed the ADEM lead sample limits; in January has provided adequate poststabilized water; how-
of the water. Pure water is considered a reactive 2012, it implemented a new corrosion control plan ever, the lead corrosion continues to bump the ex-
chemical, and water containing little to no hard- that included changing the corrosion inhibitor to ceedance level and a new approach should be
ness is often found to be aggressive towards distri- a blended orthopolyphosphate and increasing the implemented.
bution system components. Consequently, corrosion inhibitor dosage. The new plan also in- The chemical stability of potable water is typ-
poststabilization of RO-treated water is required cluded extensive testing at the water plant and in ically determined by three parameters:
prior to storage and distribution. the system. The sampling plan and performance  pH buffering capacity or alkalinity
The Authority re-evaluated its post-treatment monitoring program included corrosion test  Tendency of the water to precipitate calcium car-
and stabilization treatment after failing a lead cor- coupons located at various locations in the Au- bonate or scaling potential
rosion sample soon after the plant was commis- thority’s service area, frequent water sampling, and  Concentration of soluble calcium ions in the
sioned. It looked at four primary options for trending of historical data. water
improving its poststabilization treatment: In September 2013, the Authority again failed
 Chemical addition: minerals other than lime or the ADEM lead exceedance level. Since that time, The pH is relevant in the finished water, but it
calcite Constantine Engineering has worked with the Au- is dependent on the values of the three parameters
 Blending with a water containing high mineral thority’s operators to develop, evaluate, and imple- listed. Several calculated indices are used in the
content ment alternative water treatment processes that water industry for water stability control to deter-
 Carbon dioxide (CO2) addition, followed by cal- provide stable finish water chemistry and eliminate mine the scaling tendency of calcium carbonate.
cite or dolomite dissolution the permit violations. The most commonly accepted indices are calcium
 CO2 addition, followed by lime (slurry) dosing carbonate scaling potential (CCSP), Ryznar Stabil-
Study Objectives ity Index (RSI), and Langelier Saturation Index
Each of these four methods was reviewed and (LSI).
a cost evaluation for implementing each option was The RO process removes dissolved solids from The targeted post-treatment water quality ob-
jectives are as follows:
 40<alkalinity<80 mg/L as calcium carbonate
Table 1. Common Chemicals that Add Carbonate or Shift Carbonate Species (mg/L as CaCO3)
 LSI>0
 50<calcium (Ca)<120mg/L as CaCO3
 8.0<pH<8.5

The goal for the Authority was to increase al-

kalinity from the current level of 10 mg/L to above
40 mg/l and increase the LSI from the current -3.5
Continued on page 6

4 November 2015 • Florida Water Resources Journal

Continued from page 6 option; however, undesirable constituents in the mon at water treatment plants. The challenge with
to a positive number between 0 and 1. This can be blend water, such as color-causing agents and dis- alkalinity is to find a chemical that can shift the car-
accomplished by post-treatment remineralization. solved organic matter, prevent this option, which bonate species, add more carbonate to the system,
Generally, post-treatment remineralization was eliminated early in discussions with the Au- and remain cost effective. All of these goals cannot
can be achieved by four treatment processes: thority’s operators due to associated undesirable ef- be accomplished with one chemical, so treatment
 Chemical addition: minerals other than lime or fects and operation issues. requires the use of multiple chemicals that can add
calcite carbonate to the system, and the chemical that can
 Blending with a water containing high mineral Calcite Contactor shift the carbonate species toward carbonate ion.
content Acidification of permeate by the addition of Table 1 shows the most common water treat-
 Carbon dioxide (CO2) addition, followed by cal- CO2 that is followed by upflow calcite (limestone) ment plant chemicals that add carbonate or shift
cite or dolomite dissolution contacting is recognized in Europe and the the carbonate species. The approach to the chal-
 CO2 addition, followed by lime (slurry) dosing Caribbean to be a suitable method of post-treat- lenge of adding alkalinity is to use two of the chem-
ment of RO permeate. Although the process is used icals (one from each column) with the lowest costs
at plants in south Florida and Texas, the design cri- simultaneously.
Treatment Options teria used to develop these systems are not well es- An advanced lime feed system utilizes dis-
tablished in other parts of the United States. solved CO2 dosing systems to provide the carbon-
Chemical Additions
Dissolution of calcite is a dynamic process, ate. These systems dissolve CO2 into a carrier water
Chemicals, such as sodium bicarbonate, cal-
which may be enhanced or inhibited, depending solution to be added to the process stream. When
cium sulfate, or calcium chloride, can be used, but
on the contactor design and influent water quality. carbon dioxide solution is added to water with
there are challenges associated with chemical cost,
Constantine consulted with Tonka Water Treat- moderate pH changes, the required reaction time is
storage, and dosing. The addition of chemicals also
ment (Tonka) for its expertise in designing and op- approximately two minutes. The Authority is for-
introduces additional minerals in the finish water.
erating calcite filters for the U.S. military. tunate to have source water that has ample
In the case of calcium chloride, the resulting per-
Calcite design factors include loading rate, cal- amounts of naturally occurring CO2 dissolved into
meate chloride levels would increase to 110 to 180
cite particle size and purity, contactor bed height, the raw water, which eliminated the need for a CO2
mg/L above the current levels, which would put
and bed porosity. Influent water quality parame- feed system, thereby saving approximately $75,000
chloride levels close to or above the U.S. Environ-
ters that affect calcite dissolution include influent to $150,000 in capital costs.
mental Protection Agency (EPA) maximum con-
calcite saturation level, pH, temperature, ionic There are three options for feeding lime at the
taminant level ( MCL) of 250 mg/L . Due to these
strength, and feed water impurities. A calcite con- Dauphin Island Water Treatment Plant (WTP):
undesirable results and challenges, chemical addi-
tactor was included in the cost comparison, but was  Quicklime slaking
tions were not considered for post-treatment and
eliminated from discussion due to site constraints  Hydrated lime solution
they were eliminated from further consideration.
and the capital cost of the system.  Bulk-delivered hydrated lime solution
Blending Lime Feed Systems
At the Authority, blending with low saline feed As discussed earlier, the alkalinity of water can The main differences between hydrated lime
water from existing shallow wells is a cost-effective be increased by a variety of chemicals that are com- and quicklime are their reactivity, feed/dosing pro-
cedures, and chemical composition. Hydrated lime
and quicklime are both calcium compounds. In its
hydrated state, calcium is called calcium hydroxide,
and in its pure state, it is called calcium oxide, or
quicklime. Calcium oxide, the “natural” state of cal-
cium that comes out directly from the mine, has a
heavy density (65lb/ft³) and is more reactive than
hydrated lime.
Hydrated lime is the result of adding water to
powdered quicklime, putting it in a kiln or oven,
and then hydrating/pulverizing it with water. The
resulting lime has a density of 35lb/ft³, and is called
calcium hydroxide because it has been hydrated.
It is necessary for quicklime to be slaked in a
controlled environment because it can create heat
that reaches up to 120°F. Calcium hydroxide, or hy-
drated lime, is already neutralized, so it will not un-
dergo oxidation and can be used with water, for pH
control, lime slurry addition, and lime slurry mixes.
Quicklime’s hydrophobic reaction with
water requires a lime slaker to be used in the
process. The quicklime is generally received in
pebbles of about one-quarter to one-eighth of an
in., or in powder form (<300µ). The slaking of the
pebble lime and powdered quicklime has to be en-
gineered in respect to their exothermic reactions.

6 November 2015 • Florida Water Resources Journal

The lime slaker mixes quicklime with water to cre- Table 2. Reverse Osmosis Permeate Post-Lime Dosage
ate calcium hydroxide in a solution, which is
called lime slurry. Slakers are good for high-vol-
ume consumption or high demand of calcium.
However, at the Authority’s WTP, where a smaller
or medium lime solution is needed, hydrated lime
is more efficient because the equipment required
to use the hydrated lime is simpler and does not
need to be designed to handle an exothermic re-
action. In this case, the powder can be fed with
screw conveyors, or manually dumped directly
into the slurry tank equipped with a slurry mixer;
water is then added to create the required lime
slurry concentration. The lime slurry is dosed to
the permeate using peristaltic hose pumps. Table 3. Finished Water
Bulk liquid lime is simply hydrated lime that
has been mixed into a slurry off-site at a chemi-
cal plant where the process is closely monitored
and precisely controlled to provide a stable, con-
sistent product delivered to the water plant.

Pilot Testing
Bulk-Delivered Liquid Lime
Cal-Flo bulk-delivered liquid lime supplied
by Burnett Lime Company Inc., which was
pilot-tested in October 2013. Burnett Lime sup-
plied a complete liquid lime feed system that in-
cluded a bulk storage tank, feed pumps, mixers,
and a programmable logic control (PLC) con- $330,000, and the yearly operating cost is esti- Pros Cons
trol system. The Cal-Flo system consists of the mated to be $12,000. Precise application Higher operation
following major items: Cal-Flo presented an option to purchase a cost
 16,000-gal lime slurry tank used system that was approximately $100,000 Low maintenance Requires large
 Feed pump building less than the cost of a new system, stating that it bulk tank
 Feed pumps would provide a warranty and support the used No dust Single supplier
 Control panel and instrumentation system as if it were sold as new. Nonhazardous
 Tank mixer There could be some potential cost savings Predictable results
by designing and implementing a system other Dissolves on contact
The Cal-Flo system capital cost for equip- than that presented in the Cal-Flo proposal. The
ment and installation is estimated to be Continued on page 8
Authority can purchase an exterior tank and
mixer and utilize a transfer pump-to-pump liq-
uid lime to the existing chemical feed room; a
new day tank and mixer would be required,
along with an additional chemical feed pump.
It’s estimated that the cost for this used liquid
bulk lime alternate system would be $115,000,
which would be a savings of $100,000 over a
new system. A major disadvantage is that the
Cal-Flo feed system is patented, and dosing its
product with alternate equipment would elimi-
nate the operation guarantee from Burnett Lime
for the performance of the system.
The Cal-Flo system pilot-tested very well
and the operators found it to be easy to operate
and maintain. When the system was running,
the water quality was easy to maintain, pH was
stable, and alkalinity was easily adjusted by
changing the lime dosing rate (Tables 2 and 3).
The following shows the pros and cons of the
system, both subjective and quantitative: Hydrated Lime Bag Delivery

Florida Water Resources Journal • November 2015 7

Continued from page 7  Mixing tank with ergonomic height for dry
lime filling by operators
On-Site Liquid Lime Mixing
Pros Cons
Liquid lime can be produced on-site at the Lower operating cost Dusty
water treatment facility by mixing hydrated lime No bulk tank Increased opera-
and water to the required concentration per- tor attention
centage. For the Authority, dry hydrated lime Nonhazardous Clumping and
would be delivered to the WTP on pallets with clogging
50-lb bags; the product is delivered in 45 bags Dissolves on contact Varying consis-
per 48-in. x 40-in. pallets. The operator would tency
mix the product by manually dumping the bags Multiple suppliers Turbidity
of lime into a mixing tank and adding the ap-
propriate amount of water to create a 30 percent Plant operators have pilot-tested the on-site
solution. The lime solution would be fed to the lime mixing method and the results are extremely
RO permeate with a hose pump. good. The biggest drawbacks mentioned by oper-
This system would be best operated in a sep- ators are the dusty environment created by empty-
arate building from the existing WTP due to the ing the bags of lime and stabilizing the pH. The pH
heavy amount of dust that is created from filling may have been difficult to stabilize due to inconsis-
the mixing tank. The new building would need an tent mixing with the pilot mixer and tank; this can
area for lime pallet storage, an area for the mixing be improved with a full-scale system. It should be
tank, and a protected area for the control panel. noted that operators did not experience any tur-
Some of the recommended building amenities, bidity spikes or clogging during the pilot study. On-Site Lime Mixing and Dosing System
and their pros and cons, would include: The abundant amount of CO2 in the raw
 16-ft x 24-ft brick-and-siding building to water reacts to dissolve the lime almost instanta-
match the WTP building neously after injection. Another drawback that brings the operating cost closer than it would be
 Space for lime pallet storage should be noted is the higher feed rate that was if the dosage were equal. From the pilot study
 Loading dock for pallet offloading required to achieve the same water quality im- data, the estimated operating cost for the on-site
 Separate PLC panel room to protect the con- provements. This problem could be from the mixed liquid lime is approximately $9,000, which
trol system from dust same issues that caused the inconsistent pH sta- is about 25 percent less than liquid bulk lime.
 Roll-up doors for easy ingress/egress of bilization. This disparity in the solution feed rate Because the CO2 is naturally occurring, it
equipment and pallets between liquid bulk and on-site mixed lime should be noted that the operators have no control
over the CO2 concentration. While the concentra-
tion remains at its current level, there is plenty of
CO2 to react with the lime; however, if the CO2 con-
Table 4. Equipment Capital Cost Comparison centration should drop in the future, a supple-
mental CO2 system would be required. The CO2
levels have been high since start-up of the well in
2011, so the likelihood of a change should be small.

Pilot Study Conclusions

The pilot study tested two liquid lime feed
options as stabilization treatment for finished
water at the water plant. The study results show
that both options were able to sufficiently raise al-
kalinity and pH, and thereby stabilizing the RO
Table 5. Operation and Maintenance Cost Comparison
permeate to achieve the water quality targets. The
RO permeate treated with liquid lime, both mixed
and bulk, delivered yielded alkalinity and hard-
ness at levels above the target 40 mg/L as CaCO3
and an LSI above -0.3. No visible turbidity was ob-
served during the tests of either product and no
increase in chlorine gas was required for proper
chlorine (Cl2) residual. Operators had no prob-
lems meeting the pH, hardness, and alkalinity tar-
get levels; adjustments could be made to match
any pH level desired. The mixed liquid lime did
fluctuate more than the bulk product and some
factors that could cause this include inadequate

8 November 2015 • Florida Water Resources Journal

mixing, inaccurate measuring the dry product, or Table 6. 20-Year Present-Worth Analysis
water and/or changing consistency (i.e., changing
percent solids in the mix tank). The quality of dry
lime delivered to the site can fluctuate where bulk
delivered liquid lime is produced in a factory, with
precise formulation and quality control.
Operators have continued to work with
Carus Chemicals to select the appropriate cor-
rosion inhibitor and dosage. Carus recom-
mended targeting a pH of 8.0-8.2 and a
hardness of 25-35 mg/L, and continuing to pro-
vide a 1 ppm phosphate residual in the system.
The pilot study shows that both the bulk
purchase liquid and the dry mix on-site product
will work to stabilize finish water at the Author-
ity’s WTP. Tables 4 and 5 provide a cost compari-
son for installing and operating each system.

Capital and Operation and

Maintenance Cost Comparison
The conceptual capital cost in Table 4 is
based on equipment budget quotes and esti-
mated installation cost by a contractor provid-
ing a sealed bid. There may be some cost savings
for separating portions of the construction
and/or self-performing portions of the project.
The operation and maintenance (O&M)
cost comparison in Table 5 is based on the pilot-
study lime consumption and only takes into ac-
count lime usage; it was assumed for this
comparison that power cost and equipment
maintenance cost differences should be negligi-
ble. Prior to the new system, the WTP used ap-
proximately $3,000 per year of sodium hydroxide
solution that will no longer be required; this Figure 1. Lime Dosage Impact on pH and Alkalinity
amount can be deducted from the cost shown on
the table to achieve a net O&M value.
The 20-year present-worth analysis in
Table 6 is based upon the capital and O&M costs continuation of this program includes the fol-  pH
presented in Tables 4 and 5. lowing:  Hardness (mg/L as CaCO3)
1. Achieving a 1 mg/L of total phosphate resid-  Temperature
Improvement Summary ual in the distribution system.  TDS
2. Coupon testing in the service area and at the  Iron
The pilot study confirmed that liquid lime WTP. Coupon samples should be pulled for test-  Polyphosphates and orthophosphates
is the best solution for properly stabilizing the ing quarterly. Quarterly samples should indicate  Lead and copper
RO permeate water and eliminating the lead corrosion rates not greater than 10 mils/yr, with
permit limit excursions. a target of 5 mils/yr or less. Coupons should in- The sampling and performance monitoring
In January 2014, Burnett Lime proposed a clude mild steel, copper, and lead. has shown that the original goals are being met,
refurbished lime feed system. A site visit was 3. The sampling and performance monitoring and the plant operators have flexibility to adjust
conducted to assess the condition of the pro- program. This program provides historical water quality to suit their specific treatment
posed equipment and to allow Authority per- data that can be used to adjust chemical rates, goals. Figure 1 shows the pH and alkalinity prior
sonnel to inquire about O&M procedures. From change chemical types, and alert department to and after the lime system was placed on-line.
the site-visit findings and the results of the pilot personnel to changes in water quality within The targeted post-treatment water quality
study, the Authority’s board selected Burnett the distribution system; samples were taken objectives are as follows:
Lime to provide the refurbished Cal-Flo lime weekly for the first quarter and monthly  40<alkalinity<80 mg/L as calcium carbonate
feed equipment. This equipment was commis- thereafter. Samples are taken from the same (mg/L as CaCO3)
sioned in June 2014. locations each time (at coupon testing sites).  LSI>0
The Authority continues the sampling and The following tests are recorded:  50<calcium (Ca)<120mg/L as CaCO3
monitoring program put into place in 2012. The  Alkalinity (mg/L as CaCO3)  8.0<pH<8.5 

Florida Water Resources Journal • November 2015 9