ROBERT A. TAFT WATER RESEARCH CENTER
REPORT NO. TWRC-5
AMMONIA REMOVAL
FROM AGRICULTURAL RUNOFF
AND SECONDARY EFFLUENTS
BY SELECTED ION EXCHANGE
ADVANCED WASTE TREATMENT RESEARCH LABORATORY -V
U.S. DEPARTMENT OF THE INTERIOR
FEDERAL WATER POLLUTION CONTROL ADMINISTRATION
OHIO BASIN REGION
Cincinnofi, Ohio
-------
AMMONIA REMOVAL FROM AGRICULTURAL RUNOFF
AND
SECONDARY EFFLUENTS BY SELECTED ION EXCHANGE
by
PACIFIC NORTHWEST LABORATORIES
a division of
BATTELLE MEMORIAL INSTITUTE
P. 0. Box 999
Richland, Washington 99352
for
The Advanced Waste Treatment Research Laboratory
Robert A. Taft Water Research Center
This report is submitted in fulfillment of
Grant No. WPRD 26-01 between the Federal
Water Pollution Control Administration and
Battelle-Northwest.
U. S. Department of the Interior
Federal Water Pollution Control Administration
Cincinnati, Ohio
March 1969
-------
TABLE OF CONTENTS
LIST OF FIGURES iii
LIST OF TABLES iv
FOREWORD v
ABSTRACT vi
INTRODUCTION 1
The Need for Ammonia Removal 1
Sources of Ammonia in Wastewater U
Water Quality Standards-Ammonia 6
Remedies-Ammonia Removal from Wastewater 11
CONCLUSIONS AND RECOMMENDATIONS 13
LABORATORY STUDIES 15
Zeolite Ion Exchange Equilibria 15
Small Column Studies 18
Characterization of Secondary Effluent 22
Two Inch Column Studies 2k
PILOT PLANT STUDIES 3k
Demonstration Plant Description 31*
Wastewater Pretreatment 39
Ion Exchange Processing Ul
Ammonia Stripping Operation 1+5
REFERENCES 50
APPENDIX 53
ii
-------
FOREWORD
In its assigned function as the Nation's principal natural resource
agency, the United States Department of the Interior bears a special
obligation to ensure that our expendable resources are conserved, that
renewable resources are managed to produce optimum yields, and that all
resources contribute their full measure to the progress, prosperity, and
security of America — now and in the future.
This series of reports has been established to present the results of
intramural and contract research studies carried out under the guidance of
the technical staff of the FWPCA Robert A. Taft Water Research Center for
the purpose of developing new or improved wastewater treatment methods. In-
cluded is work conducted under cooperative and contractual agreements with
Federal, state, and local agencies, research institutions, and industrial
organizations. The reports are published essentially as submitted by the
investigators. The ideas and conclusions presented are, therefore, those of
the investigators and not necessarily those of the FWPCA.
Reports in this series will be distributed as supplies permit. Requests
should be sent to the Office of Information, Ohio Basin Region, Federal Water
Pollution Control Administration, 4676 Columbia Parkway, Cincinnati, Ohio
45226.
iii
-------
ABSTRACT
A selective ion exchange process was developed for the removal of
ammonia nitrogen from wastewater. The process employs a natural zeolite,
clinoptilolite, which is selective for ammonium ions in the presence of
sodium, magnesium, and calcium ions. Regeneration of the exhausted clinop-
tilolite is accomplished with solutions or slurries containing lime. Lime
provides hydroxyl ions which react with the ammonium ions to yield an al-
kaline aqueous ammonia solution. This ammonia solution is processed through
an air stripping tower to remove the ammonia which is exhausted harmlessly
to the atmosphere. The spent regenerant is then fortified with more lime and
recycled to the zeolite bed to remove more ammonia. Since the regenerant
is not discarded, the process generates no liquid wastes.
The ion exchange equilibria of four zeolites was investigated and
clinoptilolite was selected for further study on the basis of its ammonium
ion selectively and low cost. A cubic foot of granular clinoptilolite,
regenerated with lime, was found capable of removing ammonia from more than
2000 gallons of secondary effluent. Ammonia removals exceeding 99% vere
obtained for two clinoptilolite columns in series during laboratory studies.
A mobile demonstration plant having a capacity of 100,000 gallons per
day was designed and constructed to remove ammonia from wastewater. The
plant contains facilities for flocculation, sedimentation, powdered
activated carbon adsorption, disinfection, and mixed media filtration followed
by ion exchange and associated regeneration equipment. Most of the equipment
is housed in a forty foot trailer van which can be transported to different
locations as needed. Other portable equipment includes, a 3500 gallon
swimming pool which serves as a backwash water storage tank and ion exchange
feed tank; two 500 gallon polyethylene regenerant storage tanks; and appro-
priate flexible rubber hose connections.
Operation of the mobile plant with secondary effluent resulted in ammonia
removals of 97 and 93 percent at 70,000 and 100,000 gallons per day respec-
tively; thus demonstrating that selective ion exchange provides a highly
effective means for removing ammonia from wastewater.
iv
-------
RESEARCH REPORT
on
AMMONIA REMOVAL FROM
AGRICULTURAL RUNOFF AND SECONDARY
EFFLUENTS BY SELECTIVE ION EXCHANGE
to
FEDERAL WATER POLLUTION CONTROL ADMINISTRATION
WASHINGTON, D.C.
INTRODUCTION
The Need for Ammonia Removal
Early in the developing science of water pollution control, the presence
of ammonia in surface and ground water supplies was regarded as a strong
indication of recent pollution . Klein has stated that ammonia con-
centrations of greater than 0.2 mg/1 are a strong indication of pollution
by sewage. More explicitly, McKee and Wolf state that: "The generally
accepted limit for free ammonia for sanitary purity of water supplies is
between 0.05 and 0.10 mg/1. Excess of this value renders the water suspect
of recent pollution." While ammonia-nitrogen is commonly observed in sur-
face waters it is not found in ground waters except in small amounts under
CO
anaerobic conditions
As water quality science progressed, it became apparent that the
presence of ammonia in water has far more serious implications than merely
serving as an index of recent pollution. It was demonstrated that:
l) Ammonia can be toxic to fish and aquatic life.
2) Ammonia can contribute to explosive algae growths, thereby pro-
moting eutrophication.
3) Ammonia can restrict wastewater renovation and water reuse.
k) Ammonia can have detrimental effects on disinfection of water
supplies.
5) Ammonia can be corrosive to certain metals and materials of
construction.
-------
Where states have established water quality standards specifically
for ammonia, it has been done primarily to protect the fisheries re-
sources. Jones pointed out that ammonia is one of the most important
inorganic gases in regard to fish toxicity. Actually NH_ is more toxic
than the NHi + ion. The ratio of NHo to NHj.+ ion in an aqueous solution
f 6)
varies according to pH as shown in Figure 1 . Thus, toxicity is greatly
dependent upon pH. The toxicity to fish is said to increase by 200 per
cent or more between pH 7.^ and 8.0 . Likewise, toxicity increases as
the dissolved oxygen concentration in water decreases, and as the hard-
ness increases . The presence of C02 can only reduce toxicity to a point
where COg then becomes a toxicant itself.
McKee and Wolf have reported typical lethal ammonia exposures
to fish. Some examples of these are: 0.3 to O.k mg/1 for trout fry,
and 0.7 to 5.0 mg/l for rainbow trout. They further state that the most
widely used value for maximum recommended ammonia concentration is 2.5
mg/1 in the pH range 7.U to 8.5. The FWPCA National Technical Advisory
(7)
Committee has noted that other investigators have concluded that
"... at pH levels of 8.0 and above total ammonia expressed as N should not
exceed 1.5 mg/1."
The widely-publicized decline in Lake Erie water quality has focused
national attention on the problem of accelerated eutrophication. One of
the major manifestations of eutrophic conditions in lakes is the proli-
feration of algal blooms. These blooms result from excessive fertiliza-
/ Q \
tion of the water body. Sawyer has said that: "Ammonia nitrogen is
the most important nitrogen stimulant to explosive algal growths (compared
with nitrate nitrogen) and may be a factor in determining the type of bloom
produced."
Ammonia can cause explosive algal growths due to the fact that it can
be rapidly utilized by plant cells. Although nitrate can be stored in plant
cells, it is only utilized following its reduction through the stages of
nitrite and ammonia . Dugdale and Dugdale report that in the spring,
in particular, ammonia nitrogen is more strongly assimilated by phytoplank-
ton than any other nitrogen form.
Recent studies of future water demand have shown that unless water is
used and reused judiciously the nation faces critical shortages in virtually
all areas of the country by the year 2000. One potential solution to this
-------
100
12.0
Figure 1. Percent NHi^OH in Water as a Function of pH,
-------
problem is wastewater renovation and water reuse. The FWPCA has had a vigor-
ous program in research and development in this area for some years. In
summarizing this work Weinberger, et al. had this to say: "In tests to
date, long-term ion removal has been largely non-selective. The concentra-
tion of each ion present is reduced by roughly the same fraction. This
is fortunate, because generally uniform removal is what is required to
provide water of satisfactory quality for most purposes of reuse. One
exception exists. Only a few parts per million of ammonia can be tolerated
in municipal water supplies and in many industrial supplies. Since typical
municipal waste water may contain 30 mg/1 NH^+, the removal of 90-95 percent
would be required. To achieve this by conventional electrodialysis would
be prohibitively costly."
One reason why the need exists for ammonia removal from renovated
water, and from public water supplies in general, is due to the fact that
ammonia reacts with chlorine, the most commonly used disinfectant, to form
chloramines. Mono- and dichloramine, while still bactericidal, are slower
(12)
acting and less effective . While the presence of ammonia during
chlorination is not favored, it is frequently added in small amounts after-
ward to maintain a chlorine residual. As a matter of interest, Jones
reports that concentrations of chloramines as low as O.h mg/1 are toxic to
trout.
It is well-known that ammonia can be corrosive to certain metals and
materials of construction, notably copper and zinc alloys . In addi-
tion to being corrosive to copper and zinc, ammonium salts can also be
(3)
destructive to concrete made from portland cement
Sources of Ammonia in Wastewater
Potential sources of ammonia in wastewater which could cause concen-
trations requiring remedial action are shown in Table 1 .
Table 1
Potential Sources of Ammonia in Wastewater
Agricultural Wastewater
Animal feedlot drainage
Overflows from livestock waste lagoons
Irrigation return flows
Drainage from land fertilized with animal manure
Natural drainage from forested and agricultural land
-------
Table 1 (Continued)
Municipal Wastewater
Sewage
Urban runoff
Industrial Wastewater (lM
Basic chemicals
Fertilizer
Petroleum refining
Steel rolling and finishing
Coke and gas
The decomposition of nitrogenous organic matter represents the princi-
pal source of ammonia pollution. In sewage and in animal feedlot
drainage, ammonia is a product of the hydrolysis of urea. Typical examples
of ammonia concentrations encountered in sewage and animal feedlot drainage
are 30 mg/1 and up to 200 mg/1, respectively.
Concentrations of ammonia in irrigation return flows are generally less
than 2.0 mg/1. Overflows from livestock waste lagoons produce variable
concentrations of ammonia depending on lagoon design and loading factors.
Sylvester observed that urban street runoff has a much higher potential
ammonia nitrogen content than forested or agricultural areas. He has reported
some values as high as 7 mg/1 as H for urban street drainage.
In the steel rolling and finishing industry it has been stated
that in 1963, the ammonia waste load was 8,660,000 pounds. The
1969 and 1977 waste loads are estimated to be 11,310,000 and 1^,530,000
pounds, respectively. In steel mill coke oven gas, ammonium sulfate is
produced by removing ammonia with sulfuric acid . The major ammonia
(17)
effluent streams in steel mill wastes are identified by Nemerow as
ammonia still wastes(187 mg/1 as NH^-N), and pure still wastes (10 mg/1 as
NH3-N). Steam stripping is presently used for treatment of these effluents.
Fundamental petroleum refining processes which contribute major
quantities of ammonia to process streams are crude oil distillation, cata-
/ -, Q \
lytic cracking, and crude oil desalting . Sour water contains the most
significant amount of ammonia from these process streams, but some also
appears in certain desalter effluent waters, phenolic condensates, and
(19)
sour condensates
-------
In addition to the significant amounts of ammonia in sour water,
high hydrogen sulfide concentrations are also evident. Steam stripping
of sour water is normally carried out to remove sulfides; however, the
removal of hydrogen sulfide requires an acidic pH, while ammonia removal
occurs in the alkaline range. Hence, stripper effluent may contain from
100 to 2000 mg/1 of ammonia depending on the influent ammonia concen-
(18)
tration . Because of the drastic effect of pH, ammonia removal
efficiencies in sour water strippers can vary from 0 to 95 percent.
Because of the amounts of ammonia in sour waters, by-product recovery
e
(19)
(19)
is commonly practiced . Steam stripping is also used to treat the
desalter effluent waters, phenolic condensates, and sour condensates
Air flotation, often used as one of the unit processes to treat
the major refinery wastewater effluent, can "be effective in stripping
some ammonia if the pH is alkaline.
The coke and gas industry also produces waste streams containing
(20)
ammonia . Although the principal potential pollutant from the ammonia
still is phenol, significant quantities of cyanide and ammonia are evident.
The ammonia-containing effluent stream most likely to be directly dis-
charged is the water layer from decanters following condensation. This
stream contains from 5 to 30 mg/1 of ammonia.
Stockyard wastes from meat packing contain 8 mg/1 NH-v-N, and beet
sugar wastes contain about 15 mg/1 NHo-N. Tannery effluents sometimes
(17)
have NHl^Cl concentrations of 10 mg/1 . Other wastewaters streams in
which ammonia might appear are those from ice plants, and from scouring
(3)
and cleaning operation using "ammonia water"
It is interesting to note that ammonia is added purposely to some
(21)
wastewaters as a nutrient to aid in biological secondary treatment
The pulp and paper and food processing industries commonly employ this
practice.
Water Quality Standards - Ammonia
It has been shown in the previous discussion that considerable
ammonia finds its way into environmental waters, and that ammonia can
indeed cause degradation of water quality. Hence, it is no surprise that
many standards have been promulgated to control the potential detrimental
-------
effects of ammonia. Perhaps the standard of longest standing is the WHO
European Drinking Water Standard. The recommended limit for ammonia
is 0.5 mg/1 as NH^ . The Subcommittee for Public Water Supplies of
the FWPCA National Technical Advisory Committee on Water Quality Criteria
has suggested that the permissible ammonia criteria for water supplies
(7)
should be 0.5 mg/1 as N . They further suggest that the desirable
criteria should be less than 0.01 ir.g/1 as N.
The Subcommittee for Fish, Other Aquatic Life, and Wildlife of the
Advisory Committee has likewise made recommendations regarding desirable
(7)
levels of ammonia . Their recommendation for fresh water organisms is
as follows: "Permissible concentrations of ammonia should be determined
by the flow-through bioassay with the pH of the test solution maintained
at 8.5, DO concentrations between k and 5 mg/1, and temperatures near
the upper allowable levels." For marine and estuarine organisms the re-
commendation is: "Allowable concentrations of .... ammonia .... should
be determined by the use of 96-hour TL values and appropriate application
factors. Preferably, the TLm values should be determined by flow-through
bioassays in which environmental factors are maintained at levels under
which these materials are most toxic. Tests should utilize the most sensi-
tive life sta.ges of species of ecological or economic importance in the
area. Tentatively, it is suggested that application factors should be ...
1/20 for ammonia."
(U)
In their report to the State of California, Pomeroy and Orlob state
that for waters percolated to the ground the limit for ammonium ion should
be less than 0.25 me/1 average (U.5 mg/1 as NH^+ or 3.5 mg/1 as N).
In order to ascertain whether individual states have specifically
designated ammonia standards in their water quality standards, all 50
states plus the District of Columbia, Guam, Puerto Rico, and the Virgin
Islands were contacted*. As of the date of this writing (December 1, 1968),
the following state water quality standards had not been approved by the
Secretary of the Interior: California, Florida, Iowa, Kansas, Kentucky,
Nebraska, Utah, Virginia, and Wyoming.
*Copies of standards were not received from the following states: Arkansa.s,
California, Connecticut, District of Columbia, Iowa, Kentucky, Missouri,
New Jersey, Texas, and West Virginia.
-------
Of the hh standards received, only 13 had specific standards for
ammonia. In most of the states which did not have ammonia standards,
this substance can be included in those sections on "Toxic Substances"
by implication. Generally, these states recommend that all toxic sub-
stances be excluded from surface waters at least in those concentra-
tions which may be regarded as being detrimental to specified bene-
ficial uses. Presumably, these detrimental effects will be determined
by bioassay or observable effects. Delaware and Idaho have stated
(3)
that they will use McKee and Wolf's Water Quality Criteria as a
guide in defining toxic substance concentrations.
Michigan, while not specifically designating an ammonia standard,
made a general statement on nutrients, including phosphorus, ammonia,
nitrates, and sugars. They state: "Nutrients originating from indus-
trial, municipal or domestic animal sources shall be limited to the extent
necessary to prevent adverse effects on water treatment processes or the
stimulation of growths of algae, weeds, or slimes which are or may become
injurious to the designated use."
Three of the states which were identified as having "ammonia
standards" really have total nitrogen or nitrate standards. However,
in two of the three cases the standards are so rigorous for these sub-
stances that it was believed reasonable to call attention to them. The
states are Hawaii, Nevada, and South Dakota.
Hawaii - The total nitrogen content for the following classes of water
shall not be exceeded: Class AA - 0.10 mg/1, Class A - 0.15 mg/1, and
Class B - 0.20 mg/1.
Nevada - The value for nitrates shall be not more than 1.0 to 7.0 mg/1
depending upon the designated water course. In Lake Tahoe, the annual
average total soluble inorganic nitrogen concentration shall not exceed
0.025 mg/1.
South Dakota - For domestic water supply the nitrate concentration shall
not exceed 10 mg/1 as N, or U5 mg/1 as NOg". For wildlife propagation
and stock watering the nitrate concentration shall not exceed 50 mg/1 as
N03~.
-------
The specific ammonia standards for the remaining 10 states will
be described below.
Arizona - For those waters designated for the propagation of aquatic and
wildlife resources ammonia concentrations shall not exceed 1.5 mg/1.
Guam - For public or domestic water supply ammonia, nitrogen shall be
less than 0.01 mg/1 as N.
Illinois - In general, ammonia nitrogen concentrations shall be less
than 2.5 mg/1; however, for various waters of the state more stringent
values will be imposed. Specific examples are: Little Calumet River -
1.5 mg/1 (single daily value); Lake Michigan open water - 0.02 mg/1
(annual average) and 0.05 mg/1 (single daily value); and Lake Michigan
shore water - 0.05 mg/1 (annual average) and 0.12 mg/1 (single daily
value).
Indiana - Indiana's ammonia standards are almost identical to Illinois
as might be expected. For example, in Lake Michigan open and shore
water, and in the Little Calumet River the values are exactly the same.
An additional example is the standard for ammonia in the Indiana Harbor
Canal (the control point is the Indiana Harbor Breakwater Inner Light)
- 1.0 mg/1 for the annual average, and 1.5 mg/1 for a single daily
value.
Massachusetts - Class B - ammonia shall not exceed an average of 0.5
mg/1 as N during any monthly sampling period. Class C - not to exceed
1.0 mg/1 as N. Classes SA and SB - not to exceed 0.2 mg/1 as N.
Class SC - not to exceed 1.0 mg/1 as N.
Minnesota - Fisheries and recreation-ammonia concentrations shall be
less than the following values for these categories of use: cold water
fishery and/or whole body contact - trace; mixed fishery and/or whole
body contact - 1.0 mg/1; and resident fish and/or partial body contact -
2.0 mg/1.
Nebraska - The draft of Nebraska's water quality standards included
the following statement and the accompanying table (Table 2).
"Ammonia nitrogen concentrations shall not exceed l.U mg/1 in trout
-------
TABLE 2
MAXIMUM PERMISSIBLE LIMIT OF AMMONIA EXPRESSED AS N
AT VARIOUS pH VALUES
Maximum Limit of Ammonia
Expressed as parts per million N
pH
8.3
8.1*
8.5
8.6
8.7
8.8
8.9
9.0
Trout Waters
l.l*
1.2
1.0
0.8U
0.72
0.60
0.56
O.U8
Warm Waters
3.5
2.9
2.1+
2.1
1.8
1.1*
1.2
1.1
streams nor exceed 3.5 mg/1 in warm water streams where the pH in these
streams does not exceed a pH value of 8.3. If the pH of a stream exceeds
8.3, the undissociated ammonium hydroxide as nitrogen shall not exceed
0.1 mg/1 in trout streams nor exceed 0.25 mg/1 in warm water streams."
New York - The appropriate standard for New York is quoted as follows:
"With reference to certain toxic substances as affecting fish life,
the establishment of any single numerical standard for waters of New
York State would be too restrictive. There are many waters, which
because of poor buffering capacity and composition vill require special
study to determine safe concentrations of toxic substances. However,
based on non-trout waters of approximately median alkalinity (80 ppm)
or above for the State, in which most of the waters near industrial
areas in this State will fall, and without considering increased or
decreased toxicity from possible combinations, the following may be
considered as safe stream concentrations for certain substances to
comply with the above standard for this type of water. Waters of lower
alkalinity must be specially considered since the toxic effect of most
pollutants will be greatly increased .... Ammonia or ammonium com-
pounds — not greater than 2.0 ppm (NHo) at pH of 8.0 or above."
10
-------
North Dakota - For portions of the Red River and other designated and
related waters the ammonia concentration shall not exceed 2.0 mg/1.
Puerto Rico - For surface waters: "The concentration of ammonia shall
not "be increased "by more than 0.02 mg/1 as a monthly average. Maximum
permissible concentrations for a single sample is 3 times the established
limit, provided this limit must not be exceeded in more than 5% of such
samples."
Remedies - Ammonia Removal from Wastewater
Just as the detrimental effects of excessive ammonia concentrations
in water have been studied for many years, so have potential remedies.
(22)
Rohlich summarized some of the methods which have been investigated.
He has reported that the Guggenheim process of the 1930?s used a zeolite
for ammonia removal. Presumably this zeolite was greensand or glauconite.
It was quite effective in reducing initial concentrations of ammonia of
12 to iH mg/1 to an effluent concentration of 0.5 to 1.0 mg/1.
(22)
Rohlich also reported on the work of Nesselson who used nuclear
sulfonic cation exchangers for ammonia as well as other pollutants. De-
spite the apparent success of this method, the nonselective nature of
conventional ion exchangers makes the cost of this process prohibitive.
In addition, if conventional ion exchange were used for ammonia removal,
excessive regenerant wastes would be encountered.
(22)
Air stripping of ammonia was also discussed by Rohlich . However,
even though removals of 92 percent were reported, it was believed that
the economics of the process were unfavorable. On the other hand, more
(23)
recent experience with air stripping has resulted in more explicit data
Using a recommended value of 1 mg/1 for ammonia, research on air stripping
has shown the cost of the process to be about 2<£/1000 gal: not including
the cost of lime and solids removal from product water. The lime cost to
raise wastewater pH to 9.9 was determined to be about 3<£/1000 gal. A
cost saving would result if lime recovery was practiced. The FWPCA has
concluded that air stripping can be applied most conveniently in combina-
tion with lime treatment.
11
-------
Steam stripping has been employed widely 'in the petroleum refining
industry for ammonia removal, "but for the lower concentrations encountered
in most ammonia-laden streams it is quite costly.
Other processes have been unsuccessful in removing appreciable amounts
(2h)
of ammonia. Malhotra et al. showed that there is no ammonia removal
during alum precipitation. Likewise, electrodialysis, because of its non-
selectivity, cannot be effectively employed for ammonia removal .
Similar findings have been reported for reverse osmosis.
In summary, it would appear that a selective method for removing
the ammonium ion from wastewater would be highly effective and economical
when compared with other nonselective methods such as conventional ion
exchange, electrodialysis, and reverse osmosis. Furthermore, steam and .
air stripping are rate limited at lower ammonia concentrations and as a
result, it is very costly to achieve effluent concentrations of about
1 mg/1, a commonly recommended value. Even in distillation, ammonia appears
in the condensate. The selective ion exchange method for removing ammonia
from wastewater reported here demonstrates exceptional promise for solving
one of the nation's water pollution problems. In addition, the information
attained on selective ion exchange in this study can prove to be helpful
when applied to other problems in water pollution control.
12
-------
Conclusions and Recommendations
The removal of ammonia from secondary effluent was successfully
demonstrated on a laboratory and pilot scale using a selective ion
exchange process. The significant results of this program are:
(l) Greater than 99% ammonia removal was demonstrated in
the laboratory with two zeolite columns in series and
clarified secondary effluent containing 10 to 19 mg/1
ammonia nitrogen. From 200 to 300 column volumes of
secondary effluent can normally be processed to full
column loading.
(2) Laboratory elution. studies show that regeneration can
be accomplished with about 20 column volumes of recycled
regenerant containing lime applied at the rate of 10
column volumes per hour. More than 95% of the available
ammonia capacity was restored with this treatment. The
presence of sodium ion in the recycled regenerant mini-
mizes the volume of regenerant required for elution of
the ammonia.
(3) No liquid regenerant waste is produced since the spent
regenerant is air stripped to remove ammonia and reused
after the addition of more lime. The ammonia is exhausted
harmlessly to the atmosphere from the air stripper.
(M Residual alkalinity in a freshly regenerated clinoptilolite
bed causes some ammonia leakage during the initial part
of the loading cycle. Backwashing with ammonia free water
(i.e. product water from the clinoptilolite beds in service)
is recommended for removing the residual alkalinity.
(5) An average ammonia removal of 91% was attained for two
500 gallon clinoptilolite columns in the mobile demonstra-
tion plant with clarified secondary effluent containing
Q
16 mg/1 ammonia nitrogen. The flow rate was 6.0 gal/ft /
min for a total of 70,000 gallons per day.
13
-------
(6) An average ammonia removal of 93% was attained for two
330 gallon clinoptilolite columns with secondary effluent
containing 15 mg/1 ammonia nitrogen. Although the second-
ary effluent was simply filtered through a multimedia
filter without chemical coagulation, no zeolite bed
plugging problems were encountered with this feed. The
o
flow rate was 8.1+ gal/ft /min for a total of 100,000
gallons per day. Laboratory studies indicated fouling
of tfce zeolite with unclarified secondary effluent.
(7) Elutior: of ammonia from the mobile dejmonstration ion
exchange columns with recycled regenerant was quite
effective; however, regeneration time was longer than
anticipated due to low stripping efficiency. Optimum
air-liquid co.ntact was apparently not attained in the
air stripping \tower.
Although the number of pilot scale studie»s were limited due to
expiration of the contract, the selective ion exchange process shows
great promise as an effective, economical method for removing ammonia
from wastewater. Further pilot scale studies are recommended with the
mobile demonstration plant following minor alterations to improve the
regeneration s;ystem. Areas requiring further study include: (l)
accurate determination of clinoptilolite attrition through extended
operation of the mobile demonstration plant, (2) effect of temperature
(climate) and (3) effect of wastewater composition (i.e. high salt vs
low salt).
lU
-------
LABORATORY STUDIES
Zeolite Ion Exchange Equilibria
The ion exchange equilibria of four zeolites were investigated to
determine their selectivity for ammonium ion in the presence of other
cations normally found in relative abundance in most wastewaters. These
zeolites were chosen on the basis of their high cesium ion selectivity
The natural clinoptilolite used in this study was from the Hector,
California area. The occurrence of the Hector clinoptilolite deposit has been
(26)
described elsewhere . The AW UOO and AW 500 are zeolites produced by
(6)
the Linde Company*, and Zeoloir^ is a product of the Norton Company.**
The properties of these zeolites are listed in Table 3.
(25)
Zeolite
AW 500
AW UOO
Zeolon
Hector
Clinoptilolite
TABLE 3
ZEOLITE PROPERTIES
WT % Wt %
SiQ2/Al203 Binder H20, 25°C
Structural NH^ Capacity,
Type meq/g
l*-5
6-7
10
25
25
—
15
12
12
Chabazite
Erionite
Mordenite
1.6T
1.88
2.02
8-10
5-20
12
Clinoptilolite 1.8l
All of the capacities given in Table 3 are for the zeolite with the amounts
of binder and water indicated. Capacities measured with cations other than
ammonium can be substantially different.
Equilibrium studies were accomplished by utilizing a small column tech-
(27)
nique similar to that reported by Howery and Thomas . One-half gram of
30 to 50 mesh zeolite was placed in a straight sealing tube containing a
fritted glass disc. The sealing tubes had been cut off just below the frit
to allow placement of the tube and zeolite into a 500 ml KJeldahl flask for
ammonia determination. A solution of known chemical composition was slowly
passed through the zeolite column until equilibrium was attained. The contact
time and solution volume necessary to reach equilibrium was determined pre-
viously by passing several volumes of influent through a group of columns and
ft*
Linde Company - Tonawanda, New York
Norton Company - Worchester, Massachusetts
15
-------
ascertaining the volume and time elapsed before ammonium ion values on the
zeolite remained constant. By obtaining several equilibrium points vith con-
stant normality influents of variable cation ratio, an isotherm could be
determined for each binary system. The isotherm represents a series of
equilibrium points at the same temperature between the equivalent fraction
of one of the cations on the zeolite and the equivalent fraction of the same
cation in the equilibrium solution.
From the solution and zeolite cation compositions, a selectivity coef-
ficient, K^ , was obtained with the following relationship:
(B)?B (A )nA
Z •"
where (Ajj) , (Bjj) = normality of cations A and B in the equilibrium solution,
(A)z , (B)z = equivalent fractions of cations A and B on the zeolite, and
nA, nB = the number of moles of A and B represented in the chemical
equation for the exchange reaction of A and B.
In conjunction with ammonium ion removal, the main competing cations
in natural waters normally will be sodium, calcium, magnesium and potassium.
The present study, therefore, is concerned with zeolite ammonium selectivities
from the binary systems sodium-ammonium, calcium-ammonium, magnesium-ammonium,
and potassium-ammonium.
The use of organic cation exchange resins for selective ammonium ion
removal from systems containing divalent cations did not prove to be feasible.
A comparison of Hector Clinoptilolite and a strong acid polystyrene resin,
IR 120, is shown in Figure 2. A diagonal drawn from zero ammonium ion on
the resin and in solution to 1.0 ammonium ion on the resin and in solution
is a straight line representing no cation selectivity. Any isotherm below
the no selectivity line represents an isotherm with ammonium selectivity
values of less than 1.0, in which case the exchanger is selective for calcium.
If.the isotherm is above the diagonal line, the exchanger is selective for
the cation being measured, or ammonium ions. The IR 120 prefers calcium to
ammonium ions, as Figure 2 shows. The Hector Clinotilolite,on the other
hand, prefers ammonium to calcium ion, and hence would be of greater utility
for ammonium ion removal from calcium systems. Other organic resins were
studied and gave similar results.
16
-------
M
Hector
Clinoptilolite
0
Figure 2. The 23°C isotherms for the reaction, (Ca) + 2(NHY ) = 2(NH, )
z rj ^ ^
+ (Ca~J with Hector Clinoptilolite and IR 120. Total equilibrium
solution normality was constant at O.lNJCa) , (NH, ) = equivalent
Z *T Z
fraction of calcium or ammonium on the zeolite. (Ca,,), (NH> ) =
normality of calcium or ammonium in the equilibrium solution. The
ammonium capacity of IR 120 was U.29 meq/g of air-dried resin.
17
-------
Isotherms for the systems sodium-ammonium, calcium-ammonium, magnesium-
ammonium and potassium-ammonium with the zeolites AW 500, AW UOO, Zeolon and
Hector Clinoptilolite were determined at 23°C . Tvo general characteristics
of the isotherms were noted. The first characteristic is that all of the
above zeolites prefer potassium to ammonium ions, and the second that all
are ammonium-selective in systems containing calcium, sodium and magnesium.
Selectivity coefficients may "be determined at several points on the
isotherm, and these values plotted vs. corresponding ratios of cation normali-
ties in the equilibrium solution. The results of such a plot are presented
for magnesium-ammonium, calcium-ammonium, sodium-ammonium and potassium-
ammonium with Hector Clinoptilolite in Figures 3 and ^, respectively. It may-
be noted that Hector Clinoptilolite has a much greater preference for NH^*
over Mg and Ca+^ in strong salt solutions of these ions. This preference
or selectivity diminishes somewhat due to the mass action effect in dilute
salt solution. The above plots are useful for estimating zeolite ammonium
/po \
ion loadings for different waste compositions.
Small Column Studies
Clinoptilolite was chosen for further study in column experiments on
the basis of good .ammonium ion selectivity combined with potential low cost.
Clinoptilolite is available in several large natural deposits in the Western
United States. The small column experiments were conducted with 1.9 cm
diameter by 12.3^ cm high beds of either 20 x 50 mesh or 50 x 80 mesh Clin-
optilolite to obtain an estimation of the relationship between flow rate,
effluent ammonium content,and grain size range prior to setting up larger
columns. The results indicated that somewhere between 20 and 30 column
volumes/hr, the 20 to 50 mesh size range exhibited poor exchange kinetics
with a simulated secondary effluent. The slope of the ammonium break-
through curve became relatively shallow, and leakage of ammonium occurred
early in the loading cycle.
Improved kinetics were apparent with use of the 50 x 80 mesh Clinop-
tilolite grain size because the slope of the breakthrough curve did not change
appreciably until flow rates greater than kO column volumes/hr were attained.
A Clinoptilolite grain size range of 20 to 80 mesh would probably allow
column flow rates of 30 to 35 column volumes/hr for ammonium loading. The
18
-------
10,000
vo
1000 —
X
100
1000
10.000
Figure 3. Selectivity coefficients vs. concentration ratios of calcium
or magnesium and ammonium in the equilibrium solution with
Hector clinoptilolite at 23°C for the reaction (x) + 2(IIH} )
Z H- ft
> +
-------
ro
o
100
10
0.1 •—
0.01
J i I I I 1 I
Y=Na
Y = K
0.1
10
100
Figure k. Selectivity coefficients vs. concentration ratios of sodium or
potassium and ammonium in the equilibrium solution with Hector
clinoptilolite at 23°C for the reaction (y) + (NH, •) = (NH, )
z 4 N k z
-------
column could either be loaded downflov to avoid fluidizing the bed, or be
mechanically restrained for upflow loading. However a potential disadvantage
of small grain nizes is the high head loss often experienced in their use.
The effect of temperature on the ammonium elution rate was determined
for 20 to 50 mesh Hector clinoptilolite column loaded with an equilibrium
mixture of 1.2- NH^Cl + 0.83- KC1. ' The elution samples were collected
in an acidic solution, transferred to a Kjeldahl apparatus and made basic in
pH. The ammonia from the boiling sample was collected in boric acid and
titrated with a standard acid to the original boric acid pH. Partial elution
curves were determined with a slurry containing U. 5 g/1 of Ca(OH)p at 0°C,
10°C, ii5°C and 80°C, along with a complete elution curve at 23°C. Zero de-
grees and 80°C represented the slowest elution rates, requiring 80 to 120
column volumes to 100 percent elution. The 10°C and U5°C elution rates
were intermediate, and the 23°C curve required 3^ to hd column volumes of
a Ca(OH)2 slurry for 100 percent elution. The 23°C elution rate probably
represented an optimum between Ca(OH)2 solubility in the eluting solution
and cation exchange kinetics. A relatively high temperature resulted in
favorable exchange kinetics but low Ca(OH)2 solubility, while a low tempera-
ture yielded high Ca(OH)2 solubility but poor exchange kinetics.
Elution with a solution containing O.lJI NaCl + 5g/l of Ca(OH)2 was
tried at 23°C and U5°C. The added NaCl minimized the Ca(OH)2 temperature
effect and speeded ammonium elution, and thereby represented an improvement
over elution with Ca(OH)2 slurry alone.
At 25 column volumes, the Ca(OH)2 slurry alone had eluted 90 percent
of the ammonium on the column while the NaCl-Ca(OH)2 solution had eluted
90 percent of the ammonium, at l6 column volumes.
After some calculations on the respective pH of the equilibrium
solutions of Ca(OH)p and Mg(OH)2, it became apparent that precipitation
of the magnesium eluted from the column with a Ca(OH)2 slurry would occur.
Consequently, comparative Ca(OH.)2 slurry elution studies were made between
clinoptilolite columns loaded with the ammonium with and without magnesium.
At i+0 column volumes, the percentage of ammonium eluted was 98 for the
two columns with magnesium and 99-3 for the column without magnesium. The
presence of magnesium on the column has little or no effect on the ammonium
elution rate. Likewise, little grain size range effect was shown between 20
to 50 and 50 to 80 mesh clinoptilolite columns loaded with ammonium and
magnesium solution.
21
-------
Although a Ca(OH)2 slurry was used during the elution studies, the
flow rate was held to seven column volumes/hr. The clinoptilolite bed was
not raised at that flow rate, so the Ca(OH)p slurry only partially pene-
trated the clinoptilolite bed. In effect, therefore, the elution data were
obtained with a saturated solution of Ca(OH)2 rather than an excess Ca(OH)2
slurry. The small diameter columns were not suitable for fluidized bed
studies. It will be necessary, however, to fluidize the bed after elution
with Ca(OH) slurry to remove the Ca(OH)2-Mg(OH)2 solids. We may conclude
that saturated Ca(OH)2 solution (l.T g/1 at 23° or 0.023M) sufficed to re-
move the ammonium from the clinoptilolite.
Characterization of Secondary Effluent
Periodic grab samples of secondary effluent were taken for chemical
analysis from the No. 2 chlorinator basin at the Richland, Washington,
Wastewater Treatment Plant during the spring, summer, and fall months of
1967. This effluent stream was used for both laboratory and pilot plant
studies. The average characteristics of the secondary effluent are listed
in Table U. Since the cation concentrations including ammonium ion are of
particular importance to the ion exchange process the concentration varia-
tion over the nine month period covered is illustrated in figures 5 and
6.
The Richland Wastewater Treatment Plant is a normal high rate trick-
ling filter operation. Since the treatment plant receives essentially no
industrial wastes, the flow of approximately 3 million gallons per day is
almost entirely of domestic origin.
Table k
Average Characteristics of Richland Secondary Effluent
Ammonia as nitrogen 15 mg/1
Sodium 58 mg/1
Potassium 12 mg/1
Magnesium 8 mg/1
Calcium 3^ mg/1
Alkalinity as CaC03 180 mg/1
Ortho phosphate as phosphorus 9 mg/1
Sulfate as sulfur 20 mg/1
Nitrate as nitrogen 1 mg/1
COD 83 mg/1
pH 7.3
Specific conductance 630 mhos/cm
22
-------
60
40
en
20
0
j I
Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec.
Figure 5. Ammonia Nitrogen Concentrations in Richland Secondary Effluent
60 -
Mar. Apr. May June July Aug. Sept Oct. Nov. Dec.
Figure 6. Metal Cation Concentrations in Richland Secondary Effluent
23
-------
Two Inch Column Studies
Laboratory experiments were conducted with 2 inch diameter by 2k inch
columns of granular clinoptilolite to provide scale-up data for the mobile
pilot plant design. The flow sheet used for most of the laboratory runs is
shown in Figure 7. A second clinoptilolite column was added during the final
phase of the laboratory work to evaluate the performance of two clinoptilolite
columns in series.
The clinoptilolite beds were contained in thick wall glass pipe 2
inches inside diameter by 36 inches in length. Feed solutions were pumped
to the clinoptilolite beds with a centrifugal pump while the feed flow rate.
was regulated by means of a pneumatically operated valve. Regenerant solu-
tions or slurries were pumped through the beds with a separate centrifugal
pump.
The removal of ammonia from untreated Richland secondary effluent was
investigated since it may be desirable in some cases to remove the ammonia
without removing suspended solids. The column was operated upflow to pre-
vent plugging of the clinoptilolite beds. The removal of ammonia in this
manner is illustrated in Figure 8, along with regeneration of the bed with
a lime slurry, at a loading rate of l6.6 cv/hr. (column volumes/hour).
The 20 x 50 mesh clinoptilolite was expanded to about 125$ of the settled
volume during loading. Little fluidization occurred at this flow rate ;
however, channeling was observed throughout the bed resulting in some
loss in ammonia removal efficiency. Subsequent examination of the clinop-
tilolite indicated that organic fouling reduced the available ion exchange
capacity by about 25% during two upflow runs with untreated secondary
effluent. The capacity was regained by heating the clinoptilolite to UOO°C
for 2 hours in a muffle furnace.
The remaining 2 inch column experiments were conducted with pretreated
Richland secondary effluent. Pretreatment of the secondary effluent from
the Richland Wastewater treatment plant included: (l) lime coagulation
without activated carbon adsorption, (2) lime coagulation with carbon ad-
sorption, and (3) alum coagulation with carbon adsorption. The flow sheet
given in Figure 7 illustrates pretreatment (2). The secondary effluent was
treated with 200-300 mg of Ca(OH)p per liter to coagulate suspended matter.
After sedimentation in a 200 gallon polyethylene tank, the supernatant liquid
was decanted to a second 200 gallon polyethylene tank. The liquid was
2k
-------
Coagulation
Sedimentation
-IX-
Carbo nation
.2
-o
o>
o>
o
o
o
O
Feed Tank
D/P Cell
Mr Controlled
Valve
IX-
—tx
c
E
13
"o
O
-2
"o
ex
o
_c
O
iX-
o
Di
Figure 7. Flow Sheet for Single 2-Inch Column Runs
-------
0.6
en
c.
I °-4
3
o
o
0.2 -
0
- 12
- 8. £
o
o
- 4
0
0
40
80 120
Column Volumes
160
200
Loading Flow Rate
Column Dimensions
Clinoptilolite grain size
Temperature
Elution flow rate
Eluant
Influent NH^+
Feed
Figure 8
Loading and Elution Curves for Upflow Operation
with Untreated Richland Secondary Effluent
l6.6 cv/hr
2" diameter by 2U" depth (settled)
20 x 50 mesh
28°C
10.3 cv/hr
^.5 g/1 Ca(OH) slurry
10.9 mg/1 as KH_ - N
Untreated Richland secondary effluent
26
-------
sparged with C02 to pH 9*b to reduce the Ca+2 concentration by precipitation
of CaC03. Polyelectrolyte*(8 mg/l) was added to aid sedimentation and filtra-
tion. The mixture was then passed through an anthracite-sand filter and a
column of granular activated carbon to remove suspended matter and dissolved
organic matter. The filtrate flowed to a 500 gallon polyethylene feed tank
where the pH was adjusted to about 6.5 prior to pumping through the zeolite
column. The clarity of the filtrate after carbonation was excellent in all
runs. No clinoptilolite bed plugging problems due to suspended matter in
the feed was experienced. An example of the composition of the secondary
effluent before and after pretreatment (l) is given in Table 5.
TABLE 5
Composition of Richland Secondary Effluent
Before and After Lime Coagulation and Filtration
Before After
NH£ as N, mg/l 11.5 10.1*
Na+ , mg/l 55 51
K+ , mg/l 11 11
Mg+2 , mg/l 8.0 2.8
Ca+2 , mg/l 3lt 56
P(V3 as P, mg/l 7.5 1.6
COD 91 60
PH 7.2 6.5
Ammonia breakthrough curves for a clinoptilolite column receiving lime
clarified Richland secondary effluent are shown in Figure 9. The curves
illustrate the effect of Na on the breakthrough capacity of the clinoptilolite.
The presence of a significant amount of Na+ on the exchange sites permits
-L.O
more rapid exchange for NHj than in the case where only Ca is available for
• ,o
exchange. Self diffusion coefficients for Na and Sr (which is similar to
+2» (29)
Ca ) have been determined for Oregon clinoptilolite by Ames . The
diffusion coefficient for Na was 8.1*3 x 10 cm /sec as compared to
Q r)
1.0 x 10~ cm2/sec for Sr .
A»
Nonionic polyacrylamide
27
-------
ro
c»
c/c
0.8 -
0.6 -
0.4 -
0.2 -
0
No Regenerant Na
Regenerant Na
(B)
0
40
80
120
160
200
240
280
Column Volumes
Figure 9- Effect of Na in Regenerant on Subsequent Ammonia Breakthrough
Flow rate
Column dimensions
Clinoptilolite grain size
Temperature
Feed
Previous regeneration
Influent NH,+
20 cv/hr dovnflow.
2" x 2V.
20 x 50 mesh.
26°C.
Limed coagulated and filtered Richland secondary effluent.
(A) Iu5g/l lime at 30 cv/hr upflow, (B) ^.5g/l lime, in 0.1N_NaCT 30 cv/hr upflow.
(A) lU.O mg/1 as NH3 - N (B) 13.0 mg/1 as NH - N.
-------
The presence of Na in the lime regenerant also sharpens the ammonia
+ +2
elution curve due to the more rapid exchange rate of Na relative to Ca
Ammonia elution curves are shown in Figure 10 for 3 different regenerant solu-
tions containing lime. Curve 3 is similar to that of a saturated lime solution.
Regeneration with a mixture of CaClp and NaCl in a saturated lime solution,
as shown by curve 2 in Figure 10, represents the most economical approach.
The composition of this mixture is near that which results from continuous
recycle of a regenerant solution (0.1— NaCl + CaC^) after air stripping to
remove ammonia. Reuse of the regenerant after addition of make-up lime
eliminates the problem of disposal of a wastewater concentrate to the environ-
ment . A small amount of innocuous solid matter, largely calcium carbonate
and magnesium hydroxide, is removed from the recycled regenerant for disposal.
The use of an alkaline material such as lime for regeneration may cause
some leakage of ammonia during the initial part of the loading cycle. The
leakage is caused by residual alkalinity (i.e. lime) which increases the pH
of the column feed to the point where the ammonia is poorly ionized and there-
fore poorly adsorbed. Neutralization with C02 or HC1 was investigated^ fcut
did not appear to be satisfactory due to the agglomeration of the clinoptilo-
lite particles with CaCO^ in the case of COp or excessive consumption of acid
in the case of HC1. The latter was caused by ion exchange with the clinop-
tilolite rather than neutralization. Thorough backwashing with ammonia
free effluent from the loading cycle was found to be effective for removing
the residual alkalinity.
The performance of two clinoptilolite columns in series is illustrated
by the ammonia breakthrough curves in Figure 11. The first column had
received 193 column volumes of effluent as a polishing column in'a previous
experiment. The second column in Figure 11 was placed in series with the
first after 72 column volumes of feed had gone through the first column. The
average ammonia removal was 9&% in the product effluent from the columns.
More than 99% removal of ammonia has been demonstrated in the laboratory with
two columns in series.
The second column in Figure 11 had previously been regenerated with 0.015
M NaOH. Sodium hydroxide may be desirable as a regenerant in some cases to
avoid an increase in hardness of the effluent due to the use of lime. Regen-
eration with 0.015 M NaOH required about 35 column volumes. Higher concentrations
29
-------
32
C/G
28
24
20
16
12
8
0.1N NaCI Saturated with Ca(OH),
(1)
0.07N CaCl2 Saturated with Ca(OH), (2)
\ 0.03N NaCI t
\
\
\
\
\
\
(3)
\» 0.1N CaCL Saturated with Ca(OH),
\
\\
\
^
10
20 30
Column Volumes
40
50
Figure 10. NHvT Elution Curves for Three Regenerant Solutions
Flow rate
Temperature
10 cv/hr upflow
26°C
30
-------
C/C
0.8
0.6
0.4
0.2
0
0
First Column
Second Column
-. i „ i
40
80
120
160
200
240
Column Volumes
Figure 11. Ammonia Breakthrough for Two Clinoptilolite Columns in Series
Flow rate
Column dimensions
Clinoptilolite grain size
Temperature
Feed
Previous regeneration
20 cv/hr.
2" x 2U".
20 x 50 mesh.
26°C.
Alum coagulated and filtered Richland Effluent
KH3 - N content 18.5 mg/1.
0.015H NaOH at 10 cv/hr, Uo cv total.
-------
of NaOH may be used to reduce the volume of regenerant or the time required
for regeneration. However, precipitation of Ca(OH)? or damage to the clin-
optilolite probably will limit the strength of the caustic used. The chemical
(32)
stability of clinoptilolite has been reported by Barrer, et al, . A weight
loss of 70% was determined for clinoptilolite crystals in contact with 20%
NaOH for two days at room temperature. The relation between the amount of
attack and concentration of NaOH was almost linear.
p
A 3 inch by 10 feet air stripping column packed with 1/2 inch Intalox *
saddles was constructed in the laboratory to evaluate recycle of regenerant
solutions. The spent regenerant solutions were fed to the top of the strip-
ping column at 25 ml/min while air was introduced at the botton at 120
o
liters/min. This air/liquid ratio is equivalent to 6kO ft-3/gal. The
ammonia concentrations were reduced to less than 1 mg/1 of NH - R. The
influent air and liquid temperatures were 25° ± 2°C. The stripper perform-
ance was evaluated at 3 different air/liquid ratios with a spent regenerant
solution containing lime at pH 12 and 0.08 eq/1 CaCl2, 0.02 eq/1 Nad and
l*4l mg/1 NHo - N* The ammonia removal was 29%, 86%, and 98.8$ at 25°C for
3
air/liquid ratios of 27, 87, and 111 ft /gal., respectively. Laboratory
data on the effect of temperature at 2U ft of air per gallon of spent
regenerant showed an increase from 2h% ammonia removal at 23°C to 78$ at
50°C.
Recycled regenerant was used for 3 consecutive runs. The ammonia
elution curve for the third run which is typical of the 2 preceding runs,
is shown in Figure 12. The elution curve is relatively sharp and shows
that regeneration is essentially complete at 20 column volumes. Laboratory
studies indicate that as much as I1* mg/1 NH_ - N may be tolerated in a
recycled regenerant without impairing the ammonia removal efficiency of a
clinoptilolite column.
U. S. Stone Ware, Division of Norton Company, Akron, Ohio
32
-------
32
C/C
28 -
24 -
20 -
16
12
8
0
40
50
60
Column Volumes
Figure 12. Ammonia Elution with Recycled Regenerant
Saturated with Lime.
Flow rate 10 cv/hr upflow.
Column dimensions 2" x 2V.
Clinoptilolite 20 x 50 mesh.
Temperature 26°C.
-------
PILOT PLANT STUDIES
Demonstration Plant Description
The mobile demonstration plant houses three "basic components; a
Neptune MicroFloc, Inc.* 100,000 gpd model SWB tertiary filter unit, a
complete ion exchange and regeneration system for ammonium ion removal,
and a small wet laboratory for plant performance monitoring and control.
A photograph of the plant is shown in Figure 13.
The tertiary filter unit serves to remove suspended matter from
wastewater which might plug or foul the ion exchange beds. The unit con-
tains equipment for chemical addition, flocculation, settling and filtra-
tion. Settling is accomplished in a compact chamber which contains many
slightly inclined tube settlers. The overflow from the tube settlers is
filtered with a course-to-fine mixed media filter bed and pumped to the
backwash-filtered water storage vessel, which is a 3500 gallon portable
swimming pool.
Since the ion exchange facility is the focal point of the investiga-
tion, it was designed to be versatile and flexible. This approach has aided
greatly in achieving the experimental objectives as well as providing the
capability for investigating unusual process flow configurations. The ion
exchange facility includes three 750 gallon ion exchange vessels with re-
movable top sections. These top sections extend through the roof of the
forty foot trailer and are removed during transit. When the top sections
are not used the vessels are converted to 500 gallon total volume capacity.
A k3 inch ID stripping tower also extends through the trailer roof
during operation, and like the ion exchange vessels, its top is detachable
for trailer transport. The stripping tower is packed with about 7 feet of
1 1/2" porcelain Intalox-^ saddles. The stripping tower liquid drains into
a 200 gallon aqueous make-up tank with its own solids feeder and liquid
mixer. Air for. the stripping tower is supplied by a 10 Hp blower and is
heated by a 60 kw heater bank.
A compact piping cross header is installed underneath the trailer
frame, leaving about 18 inches clearance to the pavement of a flat road.
The piping includes 36 electrically operated ball valves which allow the
liquid flow in the ion exchange and regeneration cycles to be routed
*Neptune MicroFloc, Inc., Corvallis, Oregon
3U
-------
It
'-•>' 11 • J m
»
c
«sa SF
AMI
,
1
Figure 13. Mobile Demonstration Plant for Ammonia Removal
-------
between any conceivable origin and destination in the process. This is
achieved by providing cross connections between four general purpose cross
headers to any inlet or outlet point of the ion exchange vessels.
Six of the ball valves and one special purpose cross header serve
special functions in the regeneration cycle. One of the general purpose
headers is also diverted to serve the regeneration cycle when one of the ion
exchange columns is being renewed. Figure lU presents a schematic diagram
of the ion exchange and ion exchange regeneration system. The figure also
shows 1000 gallon regenerant storage which is provided by two 500 gallon
polyethylene tanks.
Figure 15 (photograph) shows the demonstration plant control and pro-
cess monitoring panel. Figure lU is a diagram key to the control panel
arrangement. At top left of the two-rack panel are the two temperature
controllers for the stripper,air and liquid heaters. These are fully
*
transistorized Electro-General controllers with full proportional action
with reset and reset rate adjustable provisions. These controllers each
**
modulate a Robicon SCR solid state power controller - 60 kw for the air
heater and 5 kw for the liquid heater.
Just below the temperature controllers are the switch and neon indi-
cator light arrays for the thirty-six motor-operated ball valves in the
trailer piping. Each valve may be controlled individually by one of the
switches or in a predetermined way by a master transfer switch and a
programmable 36 pole drum switch which simultaneously controls the other
35 valves as well. The individual valve switches are clustered in a 30
switch array and smaller cluster of six.
The main array is arranged to represent the valve connecting between
the indicated header (column in the array) and ion exchange vessel top or
bottom connection (row in the array). This arrangement is convenient for
the experimental purposes where the desired mode of operating the process
may be upflow or.downflow, series or parallel, or any combination of these.
The lowest panel in the left hand rack contains the start-stop controls
for the several pumps, the solids feeder, the blower, the two heaters, and
the aqueous make-up tank water feed solenoid valve. Some of the equipment
controls can be operated automatically, as well as manually, for unattended
Electro General Corporation; Hopkins, Minnesota
**Robicon Corporation; Pittsburg, Pennsylvania
36
-------
to
Ion
Exchange
Vessels
Header
"Inlet"
Header
"1-2"
Header
"2-3"
Header
"Producf
Special
Header
"R-Out"
-®-
-®-
A r:
/ ^ i ^*
Flow Controller
U
i n
/ vi ^ v
AIR-NH3
Exhaust
-®-
-®~
Product
-®-
Solids
Feeder
Aqueous
Makeup*1
-®-
Ion Exchange
Pump
1000 Gallon
Storage
Filtern
Regeneration [J
Pump
Flow Controller
Figure lU
Ion Exchange Process Schematic
Regenerant
Return Pump
Stripping
Tower
Blower
-------
1
ttaaa
struts
a.aaa
aaaa
aaisss
IS IS
« * *
9
T
ft
fe-t-v«-**iw.^^M^^^MM. — „— ^^^ — — ^— _^__J___J^J^ ____________
Figure 15. Demonstration Plant Control Panel
38
-------
operation. The ion exchange feed pump, for example, can be turned on and
off automatically "by the tertiary filter unit automatic controls, depend-
ing on the water level in the backwash storage pool. The solids feeder
can be operated automatically when the regenerant pH drops below a preset
value as measured by the regenerant pH monitor on the right hand panel.
Wastewater Pretreatment
This section will deal with the operation and performance of the Neptune
MicroFloc wastewater pretreatment unit which was described in the previous
section. Depending on the influent wastewater composition, pretreatment
requirements may range from none or simple filtration up to complete treat-
ment with the addition of powdered activated carbon, a chemical coagulant,
and a polyelectrolyte followed by flocculation, settling and filtration.
Prior to operation of the filter unit, laboratory jar tests were con-
ducted to establish required chemical concentrations for treatment of Richland
secondary effluent. Based on the results of the jar tests and trial runs
with the filter unit the chemical concentrations employed for the first two
runs were as follows: l) powdered activated carbon, TO mg/1; 2) lime, 56
mg/1; 3) alum, 2^0 mg/1; and h) polyelectrolyte, 2.5 mg/1. The lime was
added to maintain the pH near 7. Table 6 contains analytical data for the
pretreatment step for the first two runs.
COD reduction through the pretreatment system ranged from 3^ to 6l%.
This relatively low COD removal is primarily a result of a very short con-
tact time between the waste stream and the carbon prior to addition of the
coagulating chemicals. However, the effluent from the system was sparkling
clear from visual observation which is supported by the turbidity data re-
ported. The effluent was also odorless. As would be expected with the alum
concentration employed, the phosphate removal was highly efficient.
The first two runs were made at 70% of the design flow rate because of
short filter run lengths, averaging about 2 hours. The short filter run time
was caused by a'poor hydraulic flow pattern to the tube settlers which was
later corrected by a simple alteration in flow pattern from the flocculator.
Trailer test runs numbers 3 and k were conducted with only filtration
of the secondary effluent as pretreatment. Turbidity of the secondary efflu-
ent was reduced from 10 JTU to approximately 6.3 JTU. Suspended solids were
reduced from 28 to 23 mg/1.
39
-------
TABLE 6 Analytical Data for Wastewater Pretreatment
Secondary Effluent
Filter Effluent
Sample
1
2
3
1+
5
COD
mg/1
106
115
98
106
Turbidity
JTU
11.2
10.1
12.8
9.8
PV3
mg/1 P
9.3
8.5
9.3
8.7
COD
mg/1
TO
^5
U9
66
Turbidity
JTU
0.63
0.66
0.56
0.27
PO ~3
mg/1 P
<0.l6
O.16
0.16
0.16
0.16
-------
Ion Exchange Processing
Studies were conducted with the mobile demonstration plant at the
Richland wastewater treatment plant to evaluate the effectiveness of the se-
lective ion exchange process for removing ammonia nitrogen from secondary
effluent on a pilot scale. The mobile pilot plant was operated with 500
gallon clinoptilolite beds at a 70,000 gpd rate with clarified secondary
effluent and at a 100,000 gpd rate with 330 gallon clinoptilolite beds
and unclarified secondary effluent. The influent stream was pumped down-
flow through either a single clinoptilolite bed or two beds in series.
Series operation permits greater utilization of the available ion exchange
capacity since the service cycle of a single ion exchange bed is terminated
at relatively low breakthrough whereas the first bed of a series can
operate to a relatively high breakthrough. Freshly regenerated beds are
placed at the end of the series to replace the loaded beds upon their
removal from service.
The series mode of operation employing three ion exchange beds is
described by the steps below starting with two beds in service while a third
bed is being regenerated.
(l) When regeneration of the third bed is complete the regenerant
is drained from the bed to the regenerant storage tanks.
(2) The freshly regenerated bed is rinsed with a small amount
of water to recover any regenerant remaining in the bed and
replace evaporated water.
(3) The first bed in series is removed from service when it
approaches maximum loading.
(M The freshly regenerated bed is placed in series upflow after
the No. 2 column for a short period of time to remove residual
alkalinity and small particulate matter (e.g., CaCO_). Ammonia
concentration in the effluent from the No. 2 column should be
relatively low during the backwash period.
(5) When the backwash of the freshly regenerated column is complete
it is placed in downflow service.
(6) The loaded column is drained in the meantime and a solution of
NaCl and CaCl2 (total 0.1 eq/l) containing lime is pumped
through the bed upflow for regeneration.
-------
(7) Spent regenerant is air stripped to remove ammonia, make-up lime
is added, and the regenerant is recycled back through the column.
(8) Precipitated solids such as magnesium hydroxide and calcium
carbonate settle out in the regenerant storage tanks and are
periodically removed.
The first two ion exchange runs were conducted with clarified secondary
effluent. Only a single bed was used in the first run since high head loss
developed when an attempt was made to place a second bed in series with the
first. The differential pressure across both beds reached the ion exchange
vessel pressure test limit of 28 psig immediately upon placing the second
bed in service. It was possible to place two in series during the second
run, however, since the differential pressure remained just under 28 psig.
No significant pressure increase was observed during either run.
All of the clinoptilolite beds underwent short regeneration cycles
following the initial flushes made with clarified secondary effluent. The
beds were backwashed with ammonia free effluent following regeneration.
However, the backwash flow may have been insufficient to remove some of
the small particulate matter. Addition of a storage tank for backwash
water may be helpful, but it is doubtful whether the design flow rate of
100,000 gpd can be attained with two 500 gallon beds of 20 x 50 mesh clino-
ptilolite in series due to high pressure.
Breakthrough curves for the two clinoptilolite beds used in Run 2 are
shown in Figure l6. The first bed was operated alone until ammonia break-
through was detected at a throughput of 75,000 gallons. At this point a
second bed was placed in series with the first. Average ammonia removal
was 91% for the 15^,000 gallons processed during the second run and the
average effluent ammonia nitrogen concentration was 0.53 mg/1.
Lime solution containing 0.1 eq/1 of Nad + CaCl2 was employed for
regenerating the exhausted zeolite beds from the first two demonstration
runs. A curve showing percent ammonia removal during regeneration and air
stripping versus time is presented in Figure 17 for the column exhausted
in Run 2. Loading data indicates that 17.5 Ibs. of ammonia was adsorbed
whereas the curve in Figure 17 indicates that regeneration was near
completion at 15.2 Ibs. of ammonia eluted. The difference is believed to
be largely the result of fluctuation in the influent ammonia concentration
which occurred outside of the scheduled sampling period.
U2
-------
10
CT
8
.o
I 6
CD
O
1 4
on H
o>
0
•O-
•cro-
0
5X104
IxlO5
1.5X105
Gallons Throughput
Figure 16. Ammonia breakthrough curves for 'C1 and 'A' columns, Run No. 2.
Flov rate
Bed volume
Clinoptilolite grain size
Feed
Influent NH+
50 gpm
500 gallons
20 x 50 mesh
Alum coagulated Richland secondary effluent
16.3 mg/1 NH.. - N average
-------
25
Time, hr
Figure IT. Run No. 2: Regeneration and Stripping Operation
-------
The ion exchange beds were reduced in size from 500 to 330 gallons
of clinoptilolite to evaluate their performance at the design flow rate
of TO gpm (100,000 gpd). The flow could not be achieved with two 500
gallon beds in series due to pressure difficulties. A bed volume of
330 gallons was selected since it was the maximum volume of zeolite
which could be placed in the ion exchange vessels without the top section.
One third of the available space in the vessels was thus alloted for bed
expansion during backwashing. Chemical treatment of the secondary effluent
was omitted in order to study the feasibility of this mode of operation.
The multimedia filter was used essentially as a strainer to remove large
particulate matter which might plug the ion exchange beds.
Ammonia breakthrough data with the filtered but unclarified secondary
effluent showed an average ammonia nitrogen concentration of 1.1 mg/1 for
138,250 gallons of product water. The ammonia removal averaged 93 percent.
No increase in head loss occurred but the ammonia removal efficiency for
the first column loaded was poor. The loss in efficiency is believed to
have been caused by channeling in the zeolite bed.
Ammonia Stripping Operation
Ammonia is removed from the ion exchange regenerant at pH 10 -12
by countercurrent air stripping. The stripping tower used is 3 1/2 feet
inside diameter by 11 feet 9 inches overall height. 6.6 feet of 1 1/2
inch Intalox saddle packing provides a six theoretical transfer unit
performance at the range of gas and liquid rates used in the stripping
operation. Table 7 summarizes the specifications of the stripping tower
and its packing. Figure 18, a cross section drawing of the tower, also
shows the six inch thick de-entrainment mesh used at the tower top and
the liquid separator at the bottom.
9f
Air is supplied to the system by a 10 horsepower Sheldons blower
rated to deliver 1800 scfm at 20 inches water. The air passes from the
blower through a 60 kva heater bank where its temperature may be increased
to the 20-35°C range required at the tower air inlet to provide acceptable
stripping performance.
The regenerant liquid enters the top of the stripping tower and sprays
down onto the packing from a liquid distributor ring. The regenerant is
scrubbed by the rising air as it passes through the packing, and then it
is collected at the bottom of the tower and routed back to the aqueous
*Sheldons Manufacturing Corporation, Bensenville, Illinois
-------
11'-9"
1?' Flange and
Blind Cover
for Packing
Removal
Air Outlet
Boss Flange
j r
I L
6" York Mesh
Demister
;/£:p IWPl^Wpl
'i'K ife^-j^f?'-• :--i *' ft^T-h'tf^f. • *.-••• A&f&t
^mim
.? Grate, 1" Mesh
'-' • -y^'-"'
UDUUU
Liquid
Outlets
L_t
J I
Boss Flange
Figure 18. Cross Section Drawing of Stripping Tower
-------
TABLE 7
Stripping Tover Specifications
p
Tower Cross Section 9.62 ft
Packing Height 6.6 ft.
Packing Material l 1/2 inch Intalox saddles, Ceramic
Air Rate HlOO scfm
(i860 Ibs/ft -Hr.)
Liquid Rate
at 18 gpm 936 Ibs/ft -Hr
Air/Liquid Ratio 228 ft /gallon
-------
make-up tank for chemical fortification and recycle to the ion exchange
vessel being regenerated.
The performance of the stripping tower in Run 2 is presented in
Figure 19. The stripping efficiency actually achieved is considerably
(33)
lower than that reported elsewhere , observed in the laboratory studies,
or predicted for the stripping tower operations. The most probable reason
for the poor performance observed is liquid channeling preferentially down
the tower side walls without being sufficiently contacted by the upflow-
ing air stream. It is believed that the tower performance can be upgraded
to its predicted efficiency by an improved liquid inlet distributor which
should allow a complete regeneration-stripping time cycle of 18 hours or
less.
It should be emphasized that the stripping tower was designed to fit
the space limitations of the trailer van and as such the tower is relatively
small. In order to provide adequate air-liquid contact an excessive
amount of energy is required to drive the required volume of air through
this small tower. In an actual plant process a relatively large stripping
/ ~_ \
tower, such as a cooling tower would be used to avoid a high pressure
differential across the tower.
U8
-------
.:
100
so
1 60
a
o
cc
^ 40
2Q
0
25
2o
24
22
20
IS
lo
14
12
PH
f*
I :
-I /
I
% Removal
10 20
Elapsed Time, hours
Temperature
12
11.6
11.2
10.8
10.4
10
9.6
9.2
Figure 19. Run Ho. 2: Stripper Operation
-------
REFERENCES
1. Babbitt, H.E. and E.R. Baumann, Sewerage and Sewage Treatment ,
(John Wiley and Sons: New York), 1958.
2. Klein, Louis, River Polll^i"**:, Volume I. Chemical Analysis, (Butter-
worths: London), 1959-
3. McKee, J.E. and H.W. Wolf, '(Water Quality Criteria," Second Edition,
Publication No. 3-A, State Water Quality Control Board, State of
California, The Resources Agency, (1963).
h. Pomeroy, R.D. and G.T. Orlob , '^Problems of Setting Standards and of
Surveillance for Water Quality Control" , Publication No. 36, State
Water Quality Control Board, State of California, The Resources Agency
(1967).
5. Jones, J.R.E., "Fish and River Pollution, (Butterworths : London) 196U.
6. Petz Handbook of Industrial Water Conditioning, (Betz Laboratories:
Philadelphia), 1962.
7. LWater Quality Criteria, Report of the National Technical Advisory
Committee to the Secretary of the Interior, Federal Water Pollution
Control Administration, April 1, 1968.
8. Sawyer, C.N., "Fertilization of Lakes by Agricultural and Urban
Drainage: Jour. New England Water Works Assn. , 6l, 2, 109-127.
9. Hoagland, D.R. , Lectures on the Inorganic Nutrition of Plants, (Chronica
Botanica Co.: Waltham, Mass.),
10. Dugdale, V.A. and R.C. Dugdale, "Nitrogen Metabolism in Lakes. III.
Tracer Studies of the Assimilation of Inorganic Nitrogen Sources,"
6 iLimno. and Ocean 10 1, 53-57-
11. Weinberger, L.W. , D.G. Stephan, and' P.M. Middleton, "Solving Our
Water Problems - Water Renovation and Reuse," New York Academy of
Sciences Meeting, Division of Engineering, December 8, 1965.
12. Behrman, A. S., /.Water is Everybody's Business, The Chemistry of Water
Purification, (Anchor Books: Garden City, N.Y. ) 1968.
i
13. Camp, T.B. , tWater and Its Impurities (Reinhold Publishing Corp.:
New York) 1963.
lU. ,The Cost of Clean Water, Vol. II, Federal Water Pollution Control
Administration, U.S. Dept. of the Interior, 1968.
15. Sylvester, R.O., "Nutrient Content of Drainage Water from Forested,
Urban and Agricultural Areas," Algae and Metropolitan Wastes , Public
Health Service, Dept. of Health, Education, and Welfare, April I960.
p. 80.
50
-------
16. The Cost of Clean Water, Vol. Ill, I.W.P. No. 1, "Blast Furnace and Steel
Mills, FWPCA, USDI, Sept. 1967.
17. Nemerow, N.C., Theories and Practices of Industrial Waste Treatment,
(Addison-Wesley Pub. Co.:Reading, Mass.), 1963.
18. The Cost of Clean Water, Vol. Ill, I.W.P. No. 5, "Petroleum Refining "
FWPCA, USDI, Nov. 1967.
19. Beychok, M.R., Aqueous Wastes from Petroleum and Petrochemical Plants,
(John Wiley and Sons: New York), 1967.
20. Gurnham, C.F., (ed.), Industrial Wastewater Control, (Academic Press:
New York), 1965. p. 226.
21. The Cost of Clean Water, Vol. Ill, I.W.P. No. 3, "Paper Mills," FWPCA,
USDI, 1967.
22. Rohlich, G.A., "Chemical Methods for the Removal of Nitrogen and
Phosphorus from Sewage Plant Effluents", Algae and Metropolitan Wastes,
Public Health Service, Dept. of Health, Education and Welfare, April
I960. pg. 130.
23. Summary Report; Advanced Waste Treatment Research Program, July 196U -
July 1967, Publication WP-20-AWTR-19, Federal Water Pollution Control
Administration, U.S. Dept. of the Interior, (1968).
2k. Malhotra, S.K., G.F. Lee, and G.A. Rohlich, "Nutrient Removal from
Secondary Effluent "by Alum Flocculation and Lime Precipitation,"
Inter. Jour Air Water Pollution, J3, 8 and 9, ^87-500.
25. Ames, L.L. Jr. "Cation Sieve Properties of the Open Zeolites Chabazite,
Mordonite, Erionite and Clinoptilolite," Amer. Min. H6, 1170-1131,
1961.
26. Ames, L.L. Jr., Sand, L.B. and Goldich, S.S. "A Contribution on the
Hector, California Bentonite Deposit," Econ. Geol., j>3s 22-37, 1958.
27. Howery, D.G. and Thomas, H.C. "Ion Exchange on the Mineral Clinop-
tilolite," Jour. Phys. Chem., 69, 531-537, 1965.
28. Ames, L.L. "Zeolitic Removal of Ammonium Ions from Agricultural Waste-
waters", 13th Pacific Northwest Industrial Waste Conference Proceed-
ings, 135-152, Washington State University, April 1967.
29. Ames, L.L. Jr. "Self-diffusion of Some Cations in Open Zeolites",
Am. Mineral, j?£, 1*65-^75, 1965-
30. Dean, R.B. ''ultimate Disposal of Waste Water Concentrates to the
Environment", Environmental Science and Technology, j?, 1079-1086, 1968.
51
-------
31. Mercer, B.W., Ames, L.L. Jr. and Touhill, C.J. "Ammonia Removal
from Wastewater by Selective Ion Exchange", Presented to 155th
American Chemical Society Meeting, April 1967.
32. Barren, R.M., R. Papadopoulos and L.V.C. Rees. "Exchange of Sodium
in Clinoptilolite by Organic Cations", J. Inorg. Nucl. Chem. 29,
201*7-2063, 1967.
33. Slechta, A.F., and Gulp, G.L., "Water Reclamation Studies at the
South Tahoe Public Utility District", J. Water Pollution Control
Federation. 39, 787-813, 1967.
------- |