<title>Physical-Chemical Nitrogen Removal Wastewater Treatment</title>
<type>single page tiff</type>
<keyword>ammonia regenerant nitrogen stripping removal tower process breakpoint chlorine metre plant water clinoptilolite tahoe bed scaling exchange ion chlorination unit</keyword>
<author>Culp, Gordon L. United States. Environmental Protection Agency. Office of Technology Transfer. United States. Environmental Protection Agency. Office of Technology Transfer.</author>
<publisher>Environmental Protection Agency, Technology transfer,</publisher>
<subject>Nitrogen; Water--Purification </subject>
EPAlechndogy Transfer Seminar Publication
PHYSICAL-CHEMICAL NITROGEN REMOVAL
AWBERC LIBRARY U.S. EPA
ENVIRONMENTAL PROTECTION AGENCY* Technology Transfer
This seminar publication contains materials prepared for the
U.S. Environmental Protection Agency Technology Transfer Program
and has been presented at Technology Transfer design seminars
throughout the United States.
The information in this publication was prepared by Gordon
Gulp, representing Gulp, Wesner, Gulp—Clean Water Consultants,
Eldorado Hills, Calif.
The mention of trade names or commercial products in this publication
is for illustration purposes, and does not constitute endorsement or recom-
mendation for use by the U.S. Environmental Protection Agency.
Chapter I. Ammonia Stripping 3
Chapter II. Selective Ion Exchange 11
Chapter III. Breakpoint Chlorination 17
Chapter IV. Comparison of Processes 21
There are three basic physical-chemical nitrogen-removal techniques available for application
today. These three processes are
• Ammonia stripping (ch. I)
• Selective ion exchange (ch. II)
• Breakpoint chlorination (ch. Ill)
All of these approaches have the advantage that they are based on the removal of nitrogen in
ammonia form, which eliminates the costs of converting the ammonia to nitrate in the biologic-
treatment step. They also have the advantages that they are unaffected by toxic compounds that
can disrupt the performance of a biologic nitrogen-removal system, they are predictable in perform-
ance, and the space requirements for the treatment units are less than for biologic-treatment units.
The advantages and disadvantages of each of these physical-chemical processes are discussed
in detail and the processes are compared in the chapters that follow. Discussion of these processes
includes application at the following facilities, either in existence or under design:
• South Lake Tahoe, Calif.
• Orange County, Calif.
• Windhoek, South Africa
• Blue Plains, B.C.
• Upper Occoquan Sewage Authority, Va.
• Rosemount, Minn.
• North Lake Tahoe, Calif.
• Montgomery County, Md.
• Cortland, N.Y.
This page intentionally blank
The only nitrogen-removal process that actually has been used on a plant scale in wastewater
treatment is ammonia stripping. This process has been in use for ammonia nitrogen at the South
Lake Tahoe plant for about 4 years. Both the advantages and limitations of this process have been
The ammonia-stripping process itself consists of
• Raising the pH of the water to values in the range of 10.8 to 11.5, generally with the lime
used for phosphorus removal
• Formation and re-formation of water droplets in a stripping tower
• Providing air-water contact and droplet agitation by circulation of large quantities of air
through the tower
The towers used for ammonia stripping closely resemble conventional cooling towers.
Questions are sometimes raised concerning the fate of ammonia discharged to the atmosphere.
Are we merely converting a water-pollution problem to an air-pollution problem? Does the ''
ammonia stripped from the wastewater cause an air-pollution problem or find its way back to the
receiving stream owing to scavenging by precipitation?
The concentration of ammonia in the stripping-tower discharge is only about 6 mg/m3 for
domestic wastewaters (at an air flow of 500 ft3/gal and at an ammonia concentration of 23 mg/1
in the tower influent). As the odor threshold of ammonia is 35 mg/m3, the process does not
present a pollution problem in this respect. The ammonia discharged to the atmosphere is a stable
material that is not oxidized to nitrogen oxides in the atmosphere. The natural production and
release of ammonia as part of the natural nitrogen cycle is about 50 billion tons per year. Roughly
99.9 percent of the atmosphere's ammonia concentration is produced by natural biological
processes.1 There is a large turnover of ammonia in the atmosphere, with the total ammonia content
being displaced once a week on the average. Ammonia is returned to the earth through gaseous
deposition (60 percent), aerosol deposition (22 percent), and precipitation (18 percent). Ammonia
is not considered an air pollutant because there are no known public health implications, and
because it is a natural constituent of the atmosphere derived almost entirely from natural sources.
For example, a single cow releases as much nitrogen to the atmosphere in feces and urine as 12
people would contribute if all of their ammonia production were stripped to the atmosphere.
There are no standards in the United States for ammonia concentrations in the atmosphere.
Some foreign standards1 have been established.
• Czechoslovakia, 100 mg/m3 (24 hours)
• U.S.S.R., 200 mg/m3 (24 hours)
• Ontario, Canada, 3,500 mg/m3 (30 minutes)
All of these standards are far above the 6 mg/m3 that will occur right at the tower discharge. The
process cannot be dismissed from consideration because of air pollution.
A remaining question is the fate of the ammonia discharge to the air. Is it likely to find its
way into the receiving stream by being scavenged from the atmosphere by precipitation?
Ammonia may be washed from air by rainfall, but not by snowfall. The natural background
concentration of ammonia in the atmosphere is 5-7 ppb. In rainfall the natural background ranges
from 0.01 to 1 mg/1, with the most frequently reported values of 0.1 to 0.2 mg/1. The amount of
ammonia in rainfall is related directly to the concentration of ammonia in the atmosphere. Thus,
an increase in the ammonia in rainfall wuuld occur only in that area where the stripping-tower
discharge increases the natural background ammonia concentration in the atmosphere.
Calculations for the ammonia washout in a rainfall rate of 3 mm/h (0.12 in./h) have been
made for the Orange County, Calif., project. The ammonia concentrations of ammonia in the
rainfall would approach natural background levels within 16,000 feet of the tower. Of course, the
ammonia discharge during dry periods diffuses into the atmosphere quickly so that the background
concentration and resulting washout rate of ammonia at greater distances from the tower are not
affected during a subsequent storm. The ultimate fate of the ammonia that is washed out by rain-
fall within the 16,000-foot downwind distance depends on the nature of the surface upon which
it falls. Most soils will retain the ammonia. That portion which lands on paved areas or directly
on a stream surface will appear in the runoff from that area. Even though a portion of the ammonia
washed out by precipitation will find its way into surface runoff, the net discharge of ammonia to
the aquatic environment in the vicinity of the plant would be very substantially reduced.
One of the great advantages of this method of nitrogen removal is its extreme simplicity. Water
is merely pumped to the top of the tower at a high pH, air is drawn through the fill, and the am-
monia is stripped from the water droplets. The only control required is the proper pH in the
influent water. This simplicity of operation also enhances the reliability of the process.
Several factors affect the efficiency of the ammonia-stripping process.
• Type of stripping unit
• Loading rate
• Scale of deposition
There are three basic types of stripping units now being used in full-scale applications.
• Countercurrent towers
• Crossflow towers
• Stripping ponds
Countercurrent towers (the entire airflow enters at the bottom of the tower while the water enters
the top of the tower and falls to the bottom) have been found to be the most efficient. In the
crossflow towers, the air is pulled into the tower through its sides throughout the height of the
packing. This type of tower has been found to be more prone to scaling problems. The stripping-
pond approach will be discussed in more detail later.
The pH of the water has a major effect on the efficiency of the process. The pH must be
raised to the point that all of the ammonium ion is converted to ammonia gas. The pH required
varies somewhat with temperature,2 but is generally about 11.0.
Another critical factor is the air temperature. The water temperature has less effect on per-
formance because the water temperature reaches equilibrium with the air temperature in the top
few inches of the stripping tower. The efficiency of the process decreases as the temperature de-
creases. For example, at 20° C 90 percent removal of ammonia is typically achieved. At 10° C,
the maximum removal efficiency drops to about 75 percent. When air temperatures reach freezing,
the tower operation must generally be shut down owing to icing problems.
The hydraulic loading rate of the tower is also an important factor. This rate typically is ex-
pressed in terms of gallons per minute applied to each square foot of the plan area of the tower
packing. When the hydraulic loading rates become too high, good droplet formation is disrupted
and the water begins to flow in sheets. Tower loading rates of 2 gal/min/ft2 have been shown to be
compatible with optimum tower performance.2 It is critical that the water and air be uniformly
distributed over the tower area.
Another factor that may have an adverse effect on tower efficiency is scaling of the tower
packing resulting from deposition of calcium carbonate from the unstable, high-pH water flowing
through the tower. The original crossflow tower at the South Lake Tahoe plant has suffered a
severe scaling problem. The severity of the scaling problem was not anticipated from the pilot
studies in which a countercurrent tower was used. As a result, the full-scale crossflow-tower packing
was not designed with access for scale removal in mind. Thus, portions of the tower packing are
inaccessible for cleaning. Those portions that were accessible were readily cleaned by high-pressure
hosing. The potential scaling problem must be recognized in design. The use of countercurrent
towers and design of the packing with access for cleaning can adequately combat this problem.
An example of design for scale control is the 15-mgd tower now under construction at the
Orange County, Calif., Water District plant (fig. 1-1). There the tower packing has been designed
to be readily removable for cleaning as a precaution against scaling problems, although no signifi-
cant scaling problem has been observed in several months of pilot tests at Orange County.3 Scaling
has also been reported not to be a significant problem at the Windhoek, South Africa, plant where
only a soft, easily removed scale was encountered.4 On the other hand, tests at the Blue Plains
pilot plant encountered a hard scale that was extremely difficult to remove.5 The hardness of the
scale at Blue Plains was affected by operating pH, with a harder scale forming at pH 11.5 than at
Typical design criteria are
• Hydraulic loading, 1 to 3 gal/min/ft2
• Air-to-water ratio, 300 to 500 ft3/min per gal/min
• Air-pressure drop, 0.5 to 1.25 inches water
• Fan-tip speed, 9,000 to 12,000 ft/min
• Fan-motor speed, 1 or 2 speed
• Packing depth, 20 to 25 feet
Figure 1-1. Ammonia-stripping tower design. Orange County, Calif.
• Packing spacing, 2 to 4 inches horizontal and vertical
• Packing material, wood, plastic (Vfc-in. PVC pipe being used at Orange County)
A curve for estimating the costs of the ammonia-stripping process for various-size plants is
presented in figure 1-2. This curve is based on a loading rate of 2 gal/min/ft2. Because some applica-
tions may require ammonia removal only during warm weather months, operating costs are shown
for both 6-month and 12-month operation.
The South Tahoe system is being modified to reduce the impact of temperature and scaling
limitations encountered at this plant.6 Basically, the modified process will consist of three steps
(see figs. 1-3,1-4, and 1-5).
Operating and maintenance
12 months' operation
Operating and maintenance
6 months' operation
I I I I I
45678910 20 30 40 50
PLANT CAPACITY, mgd
Figure 1-2. Ammonia-stripping costs. (EPA STP Index = 200; includes engineering, legal, administrative,
construction financing, and contingencies.)
• Holding in high-pH, surface-agitated ponds
• Stripping in a modified, crossflow forced-draft tower through air sprays installed in the
• Breakpoint chlorination
This system was inspired by observations in Israel of ammonia nitrogen losses from high-pH
Pilot tests at South Tahoe indicated that the release of ammonia from high-pH ponds could be
accelerated by agitation of the pond surface. In the modified Tahoe system, the high-pH effluent
from the lime clarification process will flow to holding ponds. Holding pond detention times of
7-18 hours will be used in the modified South Tahoe plant. The pond contents will be agitated
and recycled 4-13 times by pumping the pond contents through vertical spray nozzles into the air
above the ponds. At least 37 percent ammonia removal is anticipated, even in cold weather condi-
tions, in the ponds. The pond contents then will be sprayed into the forced-draft tower. The pack-
ing will be removed from the tower and the entire area of the tower will be equipped with water
sprays. At least 42 percent removal of the ammonia in the pond effluent is anticipated, based on
pilot tests, from this added spraying in cold weather, which will include recycling of the pond
effluent through the tower to achieve 2-5 spraying cycles. The ammonia escaping this process then
,,R SPRAYING OF RECYCLED POND WATER
IN THE SECOND OF TWO PONDS
IN SECOND POND, TWO
Vh TO 1354 RECYCLES
TWO HIGH pH PONDS IN SERIES
7 TO 18 HOURS DETENTION TIME
FLOW VARIES, 2.5 TO 7.5 mgi
Figure 1-3. Proposed new and modified ammonia nitrogen removal processes. South Lake Tahoe:
New high-pH flow-equalization ponds.
EXISTING CROSS FLOW AMMONIA
Figure 1-4. Proposed new and modified ammonia nitrogen removal processes, South Lake Tahoe: Existing
stripping tower modified with new sprays.
pH - 7.0
,r*i n_(*i rtrf-i i
pH - 7.0
TO FILTERS AND
f EXISTING 2 STAGE NEW BREAKPOINT EXISTING 1 MG
* RECARBONATION CHLORINATION BALLAST POND
^^J BASIN CHAMBER FOR CHLORINE CONTACT
Figure 1-5. Proposed new and modified ammonia nitrogen removal processes. South Lake Tahoe:
Breakpoint chlorination (new).
will be removed by downstream breakpoint chlorination. The quantity of ammonia to be removed
by breakpoint chlorination will vary from 5 to 16 mg/1, depending on the plant flow and
Another approach to overcoming the limitations of the stripping process has been developed
by CH2M/HILL Consulting Engineers.8 Although the process is only in its initial stages of develop-
ment, preliminary tests indicate it may be a significant advance in the state of the art of nitrogen
removal. It appears that the new process overcomes most of the foregoing limitations and has the
advantage of recovery of ammonia as a byproduct.
The improved process, shown diagrammatically in figure 1-6, includes an ammonia-stripping unit
and an ammonia-absorption unit. Both of these units are essentially sealed from the outside air but
are connected by appropriate ducting. The stripping gas, which initially is air, is maintained in a
closed cycle. The stripping unit operates essentially in the same manner that is now being or has
been used in a number of systems, except that this system recycles the gas stream rather than using
single-pass outside air.
Most of the ammonia discharged to the gas stream from the stripping unit is removed in the
absorption unit. The absorbing liquid is maintained at a low pH to convert absorbed and dissolved
ammonia gas to ammonium ion. This technique effectively traps the ammonia and also has the
effect of maintaining the full driving force for absorbing the ammonia, since dissolved ammonia
GAS STREAM WITH
REDUCED BY ABSORPTION
OR SOLID), OR
DISCHARGE TO STEAM
STRIPPER FOR AMMONIA
GAS REMOVAL AND
WASTEWATER STRIPPED OF NEARLY
ALL OR PART OF AMMONIA (NH3)
Figure 1-6. Process for ammonia removal and recovery.
gas does not build up in the absorbent liquid. The absorption unit can be a slat tower, packed tower,
or sprays similar to the stripping unit, but will usually be smaller owing to kinetics of the absorption
The absorbent liquid initially is water with acid added to obtain low pH, usually below 7.0.
In the simplest case, as ammonia gas is dissolved in the absorbent and converted to ammonium ions,
acid is added to maintain the desired pH. If sulfuric acid is added, for example, an ammonium sul-
fate salt solution is formed. This salt solution continues to build up in concentration and the
ammonia is finally discharged from the absorption device as a liquid or solid (precipitate) blowdown
of the absorbent. With current shortages of ammonia-based fertilizers, a salable byproduct may
Other methods of removal of the ammonia from the absorbent may also be applicable, depend-
ing on the acid used and the desired byproduct. Ammonia gas or aqua ammonia could be produced,
for example, by steam stripping the absorbent. In this case, acid makeup would be unnecessary.
It is believed that the usual scaling problem associated with ammonia-stripping towers will be
eliminated by the improved process, since the carbon dioxide which normally reacts with the cal-
cium and hydroxide ions in the water to form the calcium carbonate scale is eliminated from the
stripping air during the first few passes. The freezing problem is eliminated owing to the exclusion
of nearly all outside air. The treatment system will normally operate at the temperature of the
SELECTIVE ION EXCHANGE
The selective ion exchange process derives its name from the use of zeolites that are selective
for ammonia relative to calcium, magnesium, and sodium. The zeolite currently favored for this use
is clinoptilolite, which occurs naturally in several extensive deposits in the Western United States.
Studies of the process have been conducted by Battelle Northwest9 and the University of Cali-
fornia.10 Clinoptilolite used in studies conducted by Battelle Northwest for EPA was obtained
from the Hector, Calif., leases of the Baroid Division of the National Lead Company, Houston, Tex.
The clinoptilolite is crushed and sieved to obtain a 20 by 50 mesh size. Ammonia is removed by
passing the waste water through a bed of clinoptilolite at a rate of about 10 bed volumes per hour.
The use of clinoptilolite was investigated at the University of California with the objective of
optimizing its application to ammonia removal from wastewaters. Pilot-plant operations were
carried out at three different municipal sewage-treatment plants. An average ammonia removal of
96 percent was obtained in these operations with influent ammonia nitrogen concentrations of
about 20 mg/1.
The ammonia capacity of the clinoptilolite was found to be nearly constant over the pH range
of 4.0 to 8.0, but diminished rapidly outside this range. The effect of wastewater composition on
the ammonia exchange capacity was analyzed by exhausting clinoptilolite beds with waters having
different chemical compositions. For relatively constant influent ammonia concentrations, the
ammonia exchange capacity was observed to decrease sharply with increasing competing action con-
centrations up to about 0.01 molar. Increases of cation concentrations above this value continued
to decrease the exchange capacity, but to a much lesser degree. Ammonia removal to residual levels
less than 0.5 mg/1 ammonia nitrogen is technically feasible, but only with shorter service cycles and
greater regeneration requirements. Flow rates in the range of 7.5 to 15 bed volumes per hour had
no effect on ammonia effluent values.
Battelle Northwest conducted pilot studies of the clinoptilolite process applied to secondary
effluents, advanced waste treatment effluents, and clarified raw sewage.9 -11 Ammonia removals
ranging from 93 to 97 percent were demonstrated using a 100,000-gal/d mobile pilot plant. These
studies were conducted at several different locations across the United States.
After about 150-200 bed volumes of normal-strength municipal waste have passed through the
bed, the capacity of the clinoptilolite has been used to the point that ammonia begins to leak through
the bed. At this point, the clinoptilolite must be regenerated so that its capacity to remove ammonia
The key to the applicability of this process is the method of handling the spent regenerant. The
resin is regenerated by passing concentrated salt solutions through the exchange bed when the am-
monia concentration has reached the maximum desirable level. Following regeneration, the
ammonia-laden spent-regenerant volume is about 2.5 to 5 percent of the throughput treated before
The original approach to recovering and reusing the regenerant was to use a lime slurry as the
regenerant so that the ammonium stripped from the bed during regeneration would be converted to
gaseous ammonia, which could then be removed from the regenerant by air stripping.9
Regeneration with lime alone was found to be a rather slow process; therefore, the ionic
strength of the regenerant solution was increased by the addition of salt (NaCl). The increased
ionic strength of the regenerant plus the presence of sodium ion accelerates the removal of ammonia
from the zeolite. Although most of the sodium chloride added to the regenerant is converted to
calcium chloride by continuous recycle of the regenerant, sufficient sodium ion remains under
steady state conditions to promote the elution of the ammonium ions. The sodium ion has a
higher diffusion coefficient than calcium ion, which is believed responsible for increasing the am-
monia elution rate. With the lime-slurry regenerant, the regenerant stripping tower handles only a
small fraction of the total plant throughput. Heating the stripping tower, even during cold weather
periods, is then practical.
The use of the high-pH regenerant is accompanied by an operational problem. Some plugging
of the bed with Mg(OH)2 and CaCO3 occurs when the high-pH regenerant is used. Attrition of the
zeolite is aggravated by the violent backwashing needed to remove these solids, and is 0.17-0.25
percent per cycle, making makeup clinoptilolite costs a significant factor. These problems make more
recently developed methods of regenerant recovery more attractive.
In one approach, ammonia in the regenerant solution may be converted to nitrogen gas by
reaction with chlorine which is generated electrolytically from the chlorides already present in the
regenerant solution. This process can be carried out with a regenerant of neutral pH so that the
problem of precipitation of Mg(OH)2 and CaCO3 within the bed during regeneration is eliminated.
Also, cold weather does not affect the regenerant recovery process. The regenerant solutions used
are rich in NaCl and CaCl2 which provide the chlorine produced at the anode of the electrolysis
cell. The reactions for the destruction of ammonia by chlorine are the same as for breakpoint
During regeneration of the ion exchange bed, a large amount of calcium is eluted from the
zeolite along with the ammonia. This calcium may be removed from the spent regenerant solution
by soda ash softening before passing the spent regenerant through the electrolytic cells. The soften-
ing step would lower the calcium concentration below the level that would cause calcium hydroxide
formation in the electrolytic cells. High flow velocities through the electrolysis cells are required in
addition to a low concentration of MgCl2 to minimize scaling of the cathode by calcium hydroxide
and calcium carbonate. Acid flushing of the cells would be necessary to remove this scale when the
cell resistance becomes too high for economical operation.
In pilot tests of the electrolytic treatment of the regenerant at Blue Plains, Battelle Northwest
found that about 50 Wh of power were required to destroy 1 gram of ammonia nitrogen (NH3-N).
When related to the treatment of water containing 25 mg/1 NH3-N, the energy consumed would be
4.7 kWh per 1,000 gallons. Tests at South Tahoe also indicated that a value of 50 Wh per gram is
reasonable for design.12 Preliminary capital and operating costs of $1.5 million and 9 cents per
1,000 gallons, respectively, were estimated by Battelle for a 10-mgd plant using electrolytic destruc-
tion of ammonia in recycled regenerant containing chloride salts of calcium, sodium, and magnesium.
Electrolytic treatment of the regenerant avoids the disposal of ammonia to the atmosphere or dis-
posal of aqueous ammonia concentrates. Total costs, including capital amortization, were estimated
at 12.7 cents per 1,000 gallons.11
A 22.5-mgd plant designed by CH2M/HILL for the Upper Occoquan Sewage Authority in the
State of Virginia will employ selective ion exchange with electrolytic treatment of the regenerant
for ammonia removal. This plant will utilize soda ash softening of the regenerant to avoid cathodic
scaling of the electrolysis cells. A simplified flow schematic of the regeneration system is illustrated
in figure II-l. The regeneration of the clinoptilolite beds will be accomplished with a 2-percent
solution of NaCl. The spent regenerant will be collected in a large holding tank to minimize varia-
tion in the calcium content before soda ash addition for calcium removal. After the soda ash addi-
tion, the regenerant will be clarified and transferred to another holding tank where the regenerant
will be recirculated through electrolysis cells for ammonia destruction.
Design criteria for the ammonia-removal plant for the Upper Occoquan District are summarized
in table II-l. The electrolysis cell to be used by this plant is a 500-Ampere unit manufactured by
Pacific Engineering and Production Company of Nevada, Henderson, Nev. The cell consists of a
lead dioxide coated graphite anode in a cylindrical stainless steel vessel which is the cathode. The
lead dioxide is highly resistant to attack by chlorine or oxychloroacids. The estimated total cost
for this plant is 12.6 cents per 1,000 gallons for the selective ion exchange process.
In order to develop the design criteria for the Occoquan plant, CH2M/HILL conducted pilot
tests of the process at the South Tahoe plant.12 The ammonia concentration in the wastewater at
South Tahoe ranged from 21 to 28 mg/1 during these pilot tests. After about 6 weeks of pilokplant
operation, the calcium concentration of the influent increased from about 55 mg/1 to about 80 mg/1.
This increased calcium concentration together with concurrently occurring lower influent tempera-
tures reduced the quantity of ammonia that could be loaded onto the clinoptilolite before a break-
through of 1 mg/1 of ammonia. The average loading to the clinoptilolite column before breakthrough
of 1 mg/1 of ammonia was 144 bed volumes with an influent containing 55 mg/1 calcium at 22° C.
When the influent calcium increased to 80 mg/1 and the temperature dropped to 14° C, the loading
capacity of the clinoptilolite column dropped to 104 bed volumes. Ammonia removals achieved
were in excess of 95 percent.
Figure 11-1. Simplified flow diagram of Upper Occoquan regenerant treatment system.
Table II -1 .-Design criteria for the Upper Occoquan ammonia removal plant at 22.5-mgd flow rate
Size and type
Service cycle loading:
Length of service cycle
Length of cycle
Regeneration rate . . .
NH3 destruction rate
Number of electrolytic cells in service
Total number of cells provided
10-foot-diameter X 50-foot-long horizontal
20 X 50 mesh
3.2 mgd per bed
5 mgd per bed
365 pounds NH3 per bed cycle
40 Wh per gram NH3-N destroyed
0.16 pound NH3-N per hour per cell
The pilot column was regenerated successfully with a 2-percent sodium chloride solution at
neutral pH. No loss of clinoptilolite by attrition was observed when using the neutral regenerant,
and no difficulties in backwashing were observed. Although the neutral regeneration scheme was
found to involve 30-40 bed volumes of regenerant rather than the 10 or less needed by others with
the high-pH schemes, the minimization of attrition losses is achieved without significant disadvan-
tage. The closed-loop regenerant-recovery system results only in added downtime for regeneration.
Scaling within the electrolytic cell used for regenerant recovery was the primary concern of the
Occoquan pilot-plant study; therefore, the electrolytic cell was routinely dismantled and inspected
for scaling. The flow rate through the cell was set initially at velocities of 0.13 to 0.16 ft/s, and a
thin buildup of scale was observed on the cathode at the bottom-cell-inlet end after 160 hours of
operation. After 230 hours of operation, the flow velocity was reduced to 0.06 ft/s, and very light
scale buildup was observed depositing over the entire cathode area.
Scale was removed from a 1-in.2 area of the cathode, and the flow velocity through the cell
was increased to 0.21 ft/s to determine the effect of scaling at higher cell velocities. At this in-
creased flow, which was maintained for most of the period of the pilot-plant study, no new scale
was deposited on the cathode. Visually, it appeared that from 25 to 50 percent of the previously
deposited scale was removed. These observations suggest that scaling within the cell can be con-
trolled by sufficient flow velocities. The average power requirements for regenerant recovery were
measured as 43.3 Wh per gram ammonia destroyed. To allow for normal system losses, a design
value of 50 Wh/g appears reasonable.
An alternative to air stripping or electrolysis of the regenerant is steam stripping. A 0.6-mgd
plant in Rosemount, Minn., which is now entering its startup period, utilizes this technique.13-14
At Rosemount ammonia is recovered from the spent ion exchange regenerant in an ammonia
stripper. Steam is injected into a distillation column countercurrent with the regenerant solution
to strip off the ammonia. An air-cooled plate-and-tube condenser then condenses the vapor for
collection in a covered tank as 1-percent aqueous ammonia for sale as a fertilizer, However, it is
a dilute (1 percent) ammonia solution, which reduces its potential for sale as a fertilizer, since
commercial fertilizers require handling of only 1/10 the volume of liquid for the same ammonia
No detailed data on the Rosemount design and anticipated operating parameters were available
at the time of this report. An EPA evaluation of the plant will be made in 1974 after the initial
shakedown problems are resolved. The steam-stripping process is based on the use of the high-pH
regenerant, which has the disadvantages noted earlier. Battelle Northwest's evaluation of steam
stripping51 indicates that it is economically feasible if the regenerant volume is held to 4 bed volumes
per cycle, which is achievable with high-pH regenerant. The steam requirements were estimated to
be 15 pounds per 1,000 gallons. At a steam cost of $2 per 1,000 pounds, the steam costs would
be only 0.03 cent per 1,000 gallons. Heat recovery by contacting the cold regenerant with stripped
regenerant and by contacting it with the condenser would be necessary to achieve economical
operation. Because of the unstable, high-pH regenerant, scaling problems on the heat exchanges
could be anticipated.
Another technique for regenerant recovery is the use of the stripping-recovery process (shown
in fig. 1-4) on the spent regenerant. A 6-mgd plant at North Lake Tahoe is being designed using
this approach. Tests to date indicate that ammonia sulfate concentrations of 50 percent are readily
achievable in the absorption tower. The estimated costs of the selective ion exchange approach based
on this technique of regenerant recovery are shown in figure II-2. No credit for potential sale of
ammonium sulfate has been included.
aB. W. Mercer, Battelle Northwest, personal communication, Dec. 14, 1973.
Operating and maintenance
I I I I I I
5 6 7 8 910 20
PLANT CAPACITY, mgd
30 40 50
Figure II-2. Ammonia removal by selective ion exchange. (EPA STP Index = 200; includes engineering, legal,
administrative, construction financing, and contingencies.)
When chlorine is added to a wastewater containing ammonia nitrogen, ammonia reacts with
the hypochlorous acid formed to produce chloramines. Further addition of chlorine to the break-
point converts the chloramines to nitrogen gas. The chlorine and ammonia reactions in dilute
NH4 + HOC1 -» NH2C1 (monochloramine) + H2O + H+
NH2C1 + HOC1 -> NHC12 (diochloramine) + H2O
NCH12 + HOC1 -* NC13 (nitrogen trichloride) + H2O
The reactions are dependent on pH, temperature, contact time, and initial chlorine-to-ammonia
ratio. Chlorine is added to the wastewater being treated until the chlorine residual has reached a
minimum (the breakpoint) and the ammonia is'removed. A typical breakpoint curve is shown in
figure III-l. The reaction with ammonia is very rapid. Less than 1 minute, in the pH range of 7.0
to 8.0, and all of the free chlorine is converted to monochloramine at a 5:1 weight ratio of
chlorine:ammonia nitrogen. As the weight ratio exceeds 5:1, the monochloramine breaks down and
forms dichloramine and ammonia,
Monochloramine is then oxidized by excess chlorine under slightly alkaline conditions to nitrogen
2NH2C1 + HOC1 -» N2t +3HC1 + H2O
Stoichiometrically, a weight ratio of 7.6:1 of chlorine to ammonia nitrogen is required to oxidize
ammonia to nitrogen gas.
Breakpoint chlorination tests on domestic wastewaters at the Blue Plains plant indicate that
95 to 99 percent of the ammonia is converted to nitrogen gas and that no. significant amount of
nitrous oxide is formed.15 The quantity of chlorine required to achieve breakpoint was found to
decrease with an increasing degree of treatment before the breakpoint process. The quantity of
chlorine required for breakpoint chlorination of raw wastewater was found to be 10 parts by weight
of C12 to 1 part of NH3 nitrogen. This ratio decreased to 9:1 C12 :NH3 nitrogen for secondary
effluents, and 8:1 C12 :NH3 nitrogen for lime-clarified and filtered secondary effluent. The Blue
Plains tests found that the chlorine dose was minimized at pH values between 6.0 and 7.0. The
minimum NO3 production (1.5 percent of the NH3-N) occurred at pH 5.0. At pH 8.0, the nitrate
production increased to 10 percent of the influent NH3 nitrogen. NC13 production at the break-
point decreased from 1.5 percent to the influent at pH 5.0 to 0.25 percent at pH 8.0. Temperature
did not affect the product distribution or the required chlorine dose in the range 5° to 40° C.
MOLE RATIO, CI2 : NH4-N
0.5 1 1.5
m T I
OC •» '
CHLORINE DOSAGE, mg/|
Figure 111-1. Typical breakpoint-chlorination curve.
10 11 12
The use of chlorine produces an equivalent weight of hydrochloric acid which may depress the
pH of the wastewater unless the natural alkalinity is adequate or a base such as sodium hydroxide is
added. If the pH is allowed to fall, highly odorous nitrogen trichloride (NC13) is formed, which is
an intolerable end product. If a base is used to prevent pH depression, the mixing of the wastewater,
chlorine, and base must be extremely violent to avoid local areas of low pH which would generate
NC13. Tests at Blue Plains showed that eductors do not give adequate chlorine-wastewater mixing,
which did result in localized low-pH regions in which objectionable quantities of NC13 formed.
Violent mechanical mixing is required. The use of sodium hypochlorite rather than chlorine does
not depress the pH and avoids the foregoing problem.
The use of chlorine gas may produce more acid than can be neutralized by the wastewater.
According to the EPA study reported by Pressley,15 14.3 mg/1 of alkalinity (as CaCO3) are required
to neutralize the acid produced by the oxidation of 1 mg/1 NH3-N to N2. Either sodium hydroxide
or lime may be used for pH control if the wastewater is deficient in alkalinity. A wastewater con-
taining 25 mg/1 NH3-N requires an alkalinity of about 357 mg/1 if chlorine gas is used.
A significant factor in considering this process for application in some cases is the addition of
dissolved solids inherent to the process. If, for example, chlorine gas were used and the influent
ammonia nitrogen concentration were 25 mg/1, the dissolved solids would be increased by 156 mg/1.
Neutralizing with lime would result in a total increase of 306 mg/1 of total solids. If the chlorinating
agent were sodium hypochlorite, the increase in dissolved solids would be 177 mg/1.16
The effects of breakpoint chlorination on organic nitrogen are somewhat uncertain. The Blue
Plains tests15 found only a "slight reduction in organic nitrogen within the two hour contact time."
Other tests17 observed a decrease in organic nitrogen content as the C12 :N ratio increased. Reduc-
tions from 3.2-3.5 mg/1 to 0.2-0.4 mg/1 organic nitrogen were reported for the breakpoint process.
The authors,17 however, felt that such apparent removals result from an analytical anomaly in
which the organochloramine formed is not measured as nitrogen in the Kjeldahl organic nitrogen
analysis. At higher chlorine dosages, however, their literature review indicated that organochloramines
will be oxidized to aldehydes and nitrogen gas. The breakpoint reactions of organochloramines pro-
ceed more slowly than the ammonia chloramines, and probably will not be complete in a 30-minute
Several recent studies16'17'18'19 have investigated the possibility of adding only enough
chlorine to form monochloramines and then removing the monochloramines on activated carbon.
Some advantages would be realized if monochloramine could be removed by activated carbon. The
theoretical C1:N ratio for 100 percent ammonia removal would drop from 7.6:1 for breakpoint to
about 5:1 for the formation of monochloramine. The dissolved solids added to the system and the
alkalinity requirements would be significantly reduced. Two studies16-17 found that ammonia
removals of about 50 percent could be achieved at C1:N ratio of 5:1 when the breakpoint process
was followed by activated-carbon adsorption. Complete removal still required dosages of about 9:1
in three studies.16'17-18 Carbon contact times of 10 minutes were found to be adequate for com-
plete dechlorination of the effluent.16
Experiences with the breakpoint process in South Africa20 confirm that automatic control of
the process is important. The African researchers concluded that monitoring of the ammonia
coupled with automatically controlled chlorine dosing is a necessity. A successful, automated-
computer-control system has been developed and demonstrated at the Blue Plains pilot plant.21
This system matches the quantity of chlorine fed to the quantity of incoming nitrogen, and also
controls the pH to 7.0 to minimize the formation of NC13 and NO3. (See fig. III-2.)
There are several projects in the design or construction stage utilizing the breakpoint-chlorina-
tion process. The 7.5-mgd South Lake Tahoe plant is adding facilities to provide breakpoint chlorina-
tion of the quantities of ammonia which escape the upstream nitrogen-removal processes (5-16 mg/1).6
The Orange County, Calif., 15-mgd wastewater reclamation plant now nearing completion will
include facilities to remove the 2-3 mg/1 of ammonia that will escape the upstream ammonia-stripping
process.22 Chlorine gas will be supplied from purchased 1-ton cylinders and by an on-site electrolytic
generator rated at 2,000 Ib/d. The chlorine generation system will utilize an electrochemical cell
to electrolyze sodium chloride brine to chlorine gas and sodium hydroxide solution. The sodium
hydroxide solution will be used in an adjacent sea water desalting plant.
A 60-mgd facility is under design for Montgomery County, Md., by CH2M/HILL, which will
utilize the breakpoint process as the primary nitrogen-removal process. In this plant, sodium hypo-
chlorite will be produced on site by electrolysis of a salt brine. The Cortland, N.Y., 10-mgd
physical-chemical plant design includes facilities for breakpoint chlorination of the portion of the
flow required to meet stream standards.
The costs of the process applied to the 309-mgd plant at Blue Plains were estimated at 6.7 cents
per 1,000 gallons, with chemical costs constituting 5.9 cents of this value. These costs were based
on a chlorine cost of only $75 per ton and a dose of only 120 mg/1. The control of pH was assumed
to be by lime addition (1 pound of lime per pound of chlorine) at a lime cost of $24 per ton. In any
case, the cost of the chlorine itself constitutes a large portion of the total project costs. Assuming a
chlorine cost of 0.07 cent per pound and a C1:N ratio of 8:1, the chlorine cost for removal of 25
mg/1 ammonia would be 11.8 cents per 1,000 gallons. The chlorine demand for this dose is equiva-
lent to 1,668 Ib/mg.
Figure 111-2. Breakpoint-chlorination control system.
The breakpoint process is useful for eliminating low concentrations of ammonia as a polishing
step following another nitrogen-removal process.
COMPARISON OF PROCESSES
Each of the processes discussed earlier has its advantages and disadvantages. Unfortunately,
no single process for nitrogen removal is superior to others both in terms of performance and
The ammonia-stripping process has the advantages of low cost, removal of ammonia with a
minimal addition of dissolved solids, simplicity, and reliability. However, it has the disadvantages
of poor efficiency in cold weather and the potential for scaling problems that may reduce its effi-
ciency, and it raises concerns, whether valid or not, over ammonia gas discharge. The new stripping-
recovery system overcomes many of these problems, but at the sacrifice of low process costs.
The selective ion exchange process has the advantages of high efficiency, insensitivity to tem-
perature fluctuations, removal of ammonia with a minimal addition of dissolved solids, and the
ability to eliminate any discharges of nitrogen to the atmosphere other than nitrogen gas. This
process has the disadvantage of relatively high cost, and process control and operation are relatively
The breakpoint chlorination process has the advantages of low capital cost, a high degree of
efficiency and reliability, insensitivity to cold weather, and the release of nitrogen as nitrogeii gas.
It has the disadvantage of adding a substantial quantity of dissolved solids to the effluent in the
process of removing the ammonia, it will raise public concerns over handling of chlorine gas, the
process controls required are relatively complex, and it requires a downstream dechlorination
The relative costs of the physical-chemical nitrogen processes for a 10-mgd plant are
• Ammonia stripping, 5 cents per 1,000 gallons
• Selective ion exchange, 10-13 cents per 1,000 gallons
• Breakpoint chlorination, 11 cents per 1,000 gallons
These costs all are based on the removal of 25 mg/1 ammonia nitrogen. The cost of biological
nitrogen removal by the three-stage activated-sludge process has been estimated23'24 at about 13
cents per 1,000 gallons. Preliminary estimates on the costs of the new ammonia-stripping/ammonia-
recovery process discussed earlier, which minimizes the seasonal restrictions on the ammonia-
stripping process, indicate that the cost will be 8-10 cents per 1,000 gallons. It can be seen from
the above costs that there is little economic incentive to select one process over another if faced
with a requirement for cold weather removal of ammonia. The choice must be made by weighing
the advantages and disadvantages of each approach in light of the circumstances applicable to a
This page intentionally blank
1S. Miner, "Preliminary Air Pollution Survey of Ammonia," U.S. Public Health Service,
Contract No. PH22-68-25, Oct. 1969.
2A. F. Slechta and G. L. Gulp, "Water Reclamation Studies at the South Tahoe Public Utility
District," J. Water Pollut. Cont. Fed., 39, 787, May 1967.
3G. M. Wesner and R. L. Gulp, "Wastewater Reclamation and Seawater Desalination," J. Water
Pollut. Cont. Fed., 44, 1932, Oct. 1972.
4R. B. Dean, ed., Nitrogen Removal from Wastewaters, Federal Water Quality Administration
Division of Research and Development, Advanced Waste Treatment Research Laboratory, Cincinnati,
Ohio, May 1970.
5T. P. O'Farrell et al., "Nitrogen Removal by Ammonia Stripping," J. Water Pollut. Cont.
Fed., 44, No. 8, 1527, Aug. 1972.
6J. G. Gonzales and R. L. Gulp, "New Developments in Ammonia Stripping,"Pub. Works,
May and June 1973.
7Y. Folkman and A. M. Wachs, "Nitrogen Removal Through Ammonia Release from Ponds,"
Proceedings, 6th Annual International Water Pollution Research Conference, 1972.
8L. G. Kepple, "New Ammonia Removal and Recovery Process," Water Waste, in press, 1974.
9Battelle Northwest, "Ammonia Removal From Agricultural Runoff and Secondary Effluents
by Selective Ion Exchange," Robert A. Taft Water Research Center Rep. No. TWRC-5, Mar. 1969.
10University of California, "Optimization of Ammonia Removal by Ion Exchange Using Clinop-
tilolite," U.S. Environmental Protection Agency Water Pollution Control Research Series No. 17080
DAR 09/71, Sept. 1971.
1:LBattelle Northwest and South Tahoe Public Utility District, "Wastewater Ammonia Removal
by Ion Exchange," U.S. Environmental Protection Agency Water Pollution Control Research Series
No. 17010 ECZ 02/71, Feb. 1971.
12R. Prettyman et al., "Ammonia Removal by Ion Exchange and Electrolytic Regeneration,"
unpublished report, CH2M/HILL Engineers, Dec. 1973.
13"Physical/Chemical Plant Treats Sewage Near the Twin Cities," Water Sewage Works, 120,
86, Sept. 1973.
14D. Larkman, "Physical/Chemical Treatment," Chem. Eng., Deskbook Issue, 87, June 18, 1973.
15T. A. Pressley et al., "Ammonia Removal by Breakpoint Chlonnation," Environ. Sci. Technol.,
6, No. 7, 622, July 1972.
16W. N. Stasuik, L. J. Hetling, and W. W. Shuster, "Removal of Ammonia Nitrogen by Break-
point Chlorination Using an Activated Carbon Catalyst," New York State Department of Environ-
mental Conservation Tech. Paper No. 26, Apr. 1973.
17 A. W. Lawrence et al., "Ammonia Nitrogen Removal from Wastewater Effluents by Chlorina-
tion," presented at 4th Mid-Atlanta Industrial Waste Conference, University of Delaware, Nov. 1970.
18P. F. Atkins, Jr., D. A. Scherger, and R. A. Barnes, "Ammonia Removal in a Physical Chemical
Wastewater Treatment Plant," presented at 27th Purdue Industrial Waste Conference, May 1972.
19R. C. Bauer and V. L. Snoeyink, "Reactions of Chloramines with Active Carbon," J. Water
Pollut. Cont. Fed., 45, 2990, Nov. 1973.
20L. R. J. Van Vuuren et al., "Stander Water Reclamation Plant: Chlorination Unit Process,"
Project Rep. 21, Pretoria, South Africa, Nov. 1972.
21D. F. Bishop et al., "Computer Control of Physical Chemical Wastewater Treatment,"Po//u-
tion Engineering and Scientific Solutions, vol. 2, Plenum Press, 1973.
22G. M. Wesner, "Water Factory 21—Waste Water Reclamation and Sea Water Barrier Facilities,"
Orange County Water District Rep., Feb. 1973.
23Bechtel, Inc., "A Guide to Selection of Cost Effective Wastewater Treatment Systems," draft
rep. for EPA U.S. Environmental Protection Agency, May 1973.
24R. Smith, "Updated Cost of Dispersed Floe Nitrification and Denitrification for Removal of
Nitrogen From Wastewater," U.S. Environmental Protection Agency Memorandum, Cincinnati,
Ohio, Apr. 13,1973.
METRIC CONVERSION TABLES
Flow in pipes.
nels, over weirs.
Usage of water
cubic metre per
litre per second
litre per person
Basic SI unit
The hectare (10 000
m2) is a recognized
multiple unit and
will remain in inter-
The litre is now
recognized as the
special name for
the cubic decimetre.
Basic SI unit
1 tonne = 1 000 kg
1 Mg = 1 000 kg
Basic SI unit
Neither the day nor
the year is an SI unit
but both are impor-
The newton is that
force that produces
an acceleration of
1 m/s2 in a mass
of 1 kg.
The metre is
lar to the line of
action of the force
N. Not a joule.
purposes it may be
convenient to meas-
ure precipitation in
terms of mass/unit
1 mm of rain -
1 l/s = 86.4 m3/d
The density of
water under stand-
ard conditions is
1 000 kg/m3 or
1 000 g/l or
39.37 in.=3.28 ft=
3.937 X 10'3=103A
1 0.764 sq ft
= 1.196 sq yd
6.384 sq mi =
0.001 55 sq in.
35.314 cu ft =
1. 057 qt = 0.264 gal
= 0.81 X 10^ acre-
0.035 oz = 1 5.43 gr
0.984 ton (long) «
1.1023 ton (short)
0.22481 Ib (weight)
* 7.233 poundals
0.02089 Ibf/sq ft
0.14465 Ibf/sq in
quantity of heat
litre per second
joule per second
Basic SI unit
The Kelvin and
The use of the
Celsius scale is
it is the former
1 joule - 1 N-m
where metres are
the line of
1 watt = 1 J/s
1 5,850 gpm
0.000145 Ib/sq in
0.145 Ib/sq in.
14.5 b/sq in.
2.778 X 10'7
3.725 X ID'7
hp-hr = 0.73756
ft-lb = 9.48 X
Application of Units
0.0624 Ib/cu ft
per unit area;
per unit volume;
per square metre
per cubic metre
cubic metre or
litre of free air
If this is con-
verted to a
should be ex-
pressed in mm/s
(1 mm/s = 86.4
3.28 cu ft/sq ft
39.37 in. =
U.S. ENVIRONMENTAL PROTECTION AGENCY • TECHNOLOGY TRANSFER EPA625/4-74-008
"M- *-.«£ ""* <3» t
- ->.?**•-** £ >* * «T« ,">
' % **^ *•"-" *• J5^ "^ « "^ n*
't ^ i?/"-
Vfy«*3fc "'" -V^ •*'
Vp.^,^ x" ;-^-:' -«
v^v* -'- • .<'-•*•&
•^\f"t\ % t ,' : i*i^1*J-'
;^. a, r* *•*, ^ <^ ^ ; ?*