WATER POLLUTION CONTROL RESEARCH SERIES
17080 DAR 09/71
OPTIMATION OF AMMONIA REMOVAL
BY ION EXCHANGE USING CLINOPTILOLITE
U.S. ENVIRONMENTAL PROTECTION AGENCY
-------
WATER POLLUTION CONTROL RESEARCH SERIES
The Water Pollution Control Research Series describes the results and
progress in the control and abatement of pollution in our Nation's
waters. They provide a central source of information on the research,
development, and demonstration activities in the Environmental
Protection Agency, through inhouse research and grants and contracts
with Federal, State, and local agencies, research institutions, and
industrial organizations.
Inquiries pertaining to Water Pollution Control Research Reports should
be directed to the Chief, Publications Branch, Research Information
Division, Research and Monitoring, Environmental Protection Agency,
Washington, D. C. 20460.
-------
OPTIMIZATION OF AMMONIA REMOVAL BY
ION EXCHANGE USING CLINOPTILOLITE
by
Sanitary Engineering Research Laboratory
College of Engineering
and
School of Public Health
University of California
Berkeley, California 94720
for the
ENVIRONMENTAL PROTECTION AGENCY
Project #17080 DAR
September 1971
For sale by the Superintendent of Documents, U.S. Government Printing Office
Washington, D.C., 20402 - Price $1.50
-------
EPA Review Notice
This report has been reviewed by the Environmental Protection Agency
and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the
Environmental Protection Agency, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
-------
ABSTRACT
The zeolite ion exchanger clinoptilolite was investigated with the
objective of optimizing its application to ammonia removal from waste-
waters. The study included multiple cycle pilot plant operations at
three municipal sewage treatment plants. Particular attention was
given to cation interference with exhaustion performance and with
minimum cost regeneration.
The ammonia capacity of clinoptilolite was found to be nearly constant
over the pH range of 4 to 8, but diminished rapidly outside this range.
In regeneration the pH was critical in determining the NaCl require-
ments, a higher pH favoring lesser amounts of salt. However, at a pH
over 12.5 zeolite attrition became excessive and exchanger makeup
contributed significantly to operating costs.
An average ammonia removal of 95.7% was obtained in demonstration
studies on three municipal wastes having an NH3-N content of about
20 mg/£. The cost of ammonia removal using clinoptilolite for a
10-mgd plant operating under these conditions was estimated to be
$0.082/1000 gal. Ammonia removal down to less than 0.5 mg/a NH3-N
is technically feasible, but only with shorter exhaustion runs and
greater regenerant requirements.
This report was submitted in fulfillment of Grant No. 17080 DAR by
the University of California for the Environmental Protection Agency.
Partial support was provided by the University.
n
-------
CONTENTS
Chapter Page
I. CONCLUSIONS 1
II. RECOMMENDATIONS 7
III. INTRODUCTION 9
IV. ION EXCHANGE CHARACTERISTICS OF THE ZEOLITES 15
V. NATURE AND SCOPE OF THE INVESTIGATION 39
VI. EXPERIMENTAL EQUIPMENT AND PROCEDURES . . . 47
VII. SATURATION PERFORMANCE OF CLINOPTILOLITE 61
VIII. REGENERATION STUDIES 79
IX. PROCESS PERFORMANCE 105
X. DESIGN CONSIDERATIONS AND COST ANALYSIS 123
XI. ACKNOWLEDGEMENTS 137
XII. REFERENCES 139
XIII. GLOSSARY 147
XIV. APPENDICES 151
-------
FIGURES
Figure Page
1. Generalized Ion Exchange Isotherms -]7
2. Idealized Zeolite Particles 28
3. Isotherms for Exchange of NH4 for K, Na, Ca , and
Mg++ on Clinoptilelite 34
4. Column Unit Located at the SERL Treatment Facility 43
5. Column Unit Used in Studies at EBMUD and CCCSD 48
6. Schematic Illustration of Column Unit 49
7. Grain Size Distribution of Clinoptilolite 54
8. Treatment Systems Used in Wastewater Studies at SERL .... 57
9. Treatment System Used in Studies at EBMUD 58
10. Treatment System Used in Studies at CCCSD 58
11. Exhaustion of Clinoptilolite in Na Form 65
12. Variation of Ammonia Exchange Capacity with Competing
Cation Concentration 67
13. Exhaustion of Clinoptilolite in Ca Form 70
14. Ammonia Concentration Histories for Ca- and Na-
Clinoptilolite 71
15. Batch Results - Effect of pH on Ammonia Exchange 74
16. Relative Effect of pH on Ammonia Exchange Capacity 74
17. Effect of pH on Ammonia Breakthrough 76
18. Effect of Flow Rate on Regeneration 83
19. Variation of Ammonia Elution with Flow Rate 84
20. Ammonia Elution - No Nad 86
21. Ammonia Elution - 0.049 Ib NaCl/gal 86
VI
-------
FIGURES (Continued)
Figure Pa9e
22. Ammonia Elution - 0.10 Ib NaCl/gal 87
23. Ammonia Elution - 0.17 Ib NaCl/gal 87
24. Ammonia Elution - 0.24 Ib NaCl/gal 88
25. Ammonia Elution - 0.73 Ib NaCl/gal 88
26. Variation of Ammonia Elution with Regenerant Strength ... 89
27. Volume of Regenerant Required for 95 Percent Ammonia 92
Elution
28. Effect of Regenerant Composition on Efficiency 95
29. Laboratory Columns Used in Attrition Study 98
30. Comparison of Column Rinse Requirements . 102
31. Envelope of Ammonia Leakage in Tests at SERL Pilot Plant . .110
32. Variation of Ammonia Leakage with Amount of Salt Used
in Previous Regeneration 113
33. Variation of Ammonia Leakage with pH of Previous
Regeneration 113
34. Cost of Regeneration - No Regenerant Reuse 131
35. Cost of Regeneration Reuse of Regenerant 131
-------
TABLES
Table
1. Ion Exchange Capacities of Clinoptilolite 32
2. Separation Factors for Ammonia Exchange on Clinoptilolite
at XNH3-N = °'5 33
3. Size of Selected Cations 36
4. Average Composition of SERL Tap Water 55
5. Clinoptilolite Exchange Capacity 62
6. Influent Chemical Composition for Saturation Runs 63
7. Equilibrium Values for Saturation Runs 64
8. Influent Chemical Composition for Exhaustion of Na- and
Ca-Clinoptilolite 68
9. Chemical Composition for Column pH Runs 75
10. Column Ammonia Capacity for pH Runs 77
11. Composition of Column Influent for Regeneration Studies . . 80
12. Column Performance During Regeneration Studies 81
13. Amount of Chemicals Required for Regeneration 93
14. Weight Loss of Clinoptilolite Exposed to 2% NaOH
(pH 13.3) 97
15. Weight Loss of Clinoptilolite in Column Attrition Studies .
16. Comparison of Column Rinsing Procedures
17- Cation Composition of Primary Effluent During SERL
Pilot Plant Runs 106
18. Cation Composition of Column Influent and Effluent
During SERL Pilot Plant Runs 107
19. Series I Operational Data for Studies at SERL
Pilot Plant 109
vm
-------
TABLES (Continued)
Table EM6.
20. Effect of Column Flow Rate on Ammonia Leakage Ill
21. Series II Operational Data for Studies at SERL
Pilot Plant 114
22. Cation Composition of EBMUD Wastewater 116
23. Operational Data for Studies at EBMUD Pilot Plant 118,
24. Cation Composition of CCCSD Wastewater 119
25. Column Performance for Studies at CCCSD 120
26. Properties and Operating Conditions for Clinoptilolite . . 124
27. Operating Characteristics for a 10-mgd Clinoptilolite
Ion Exchange Facility 126
28. Cost of Chemicals and Clinoptilolite 129
29. Cost Summary 10-mgd Clinoptilolite Ion Exchange Plant . . 134
IX
-------
I. CONCLUSIONS
The principal objectives of this investigation were: 1) to determine
the effect of water composition, ionic form of the zeolite, and pH on
the exhaustion performance of clinoptilolite; 2) to optimize the regene-
ration cycle of clinoptilolite; 3) to conduct studies to determine the
performance of clinoptilolite in removing ammonia from chemically
different sewages; and 4) to develop design criteria and estimate
treatment costs associated with the removal of ammonia using clinoptilolite
Experimental studies were conducted using clinoptilolite beds 4 in. in
diameter by 3 ft in length. Tests in chemically defined systems using
fortified tap water were made to determine ion exchange properties of
clinoptilolite in column operation. Subsequent tests with sewage were
made to optimize column operation and to evaluate the performance of
clinoptilolite in wastewater systems. A cost analysis of the process
was made based on the experimental results of the study. Conclusions
are presented with respect to the various parts of the study with specific
implications regarding the operation and design of clinoptilolite ion
exchange units listed where appropriate.
EXHAUSTION PERFORMANCE OF CLINOPTILOLITE
The exhaustion performance of clinoptilolite was examined using tap
water containing cation concentrations typical of domestic sewages.
Clinoptilolite was exhausted in the Na and Ca forms to compare their
operational characteristics. Results indicated that column kinetics
were considerably more favorable for exhaustion and regeneration using
the Na form.
The effect of water composition on the ammonia exchange capacity was
analyzed by exhausting clinoptilolite with waters having different
chemical compositions. For a relatively constant influent ammonia
concentration, the ammonia exchange capacity decreased sharply with
increasing competing cation concentrations representative of "average"
sewages. At higher concentrations the ammonia exchange capacity decreased
to a much lesser degree as the cation concentration of the feed was
increased. The results of these tests may be used to estimate the
ammonia exchange capacity of clinoptilolite for wastewaters containing
cation concentrations which are not unusually different from concen-
trations used in tests made during this study.
A model was.formulated to describe the effect of pH on ammonia exchange
in a ternary NH^-Na-H system. This model described the solid phase
ammonium ion concentration as a function of pH for specific values of
-------
the NH^-H and NH4-Na selectivity coefficients and the armionia dissocia-
tion constant. Experimental tests were made using both 2-hr batch
equilibrations and column runs to demonstrate the validity of this model.
The results of these tests make it possible to understand the effect of
pH on column performance during exhaustion and to combine the ammonia
exchange process with other unit processes required to meet specific
treatment objectives.
The following specific conclusions were made regarding the exhaustion of
clinoptilolite in columnar systems:
1. The ammonia exchange capacity of domestic wastewaters can be esti-
mated from the cationic strength (similar to total ionic strength)
of the column influent. The ammonia exchange capacity was observed
to decrease sharply with increasing competing cation concentrations
to a cationic strength of about 0.01 mole/£. Increases of cationic
strength above this value continued to decrease the exchange capacity,
but to a much lesser degree.
2. While the total ammonia exchange capacity was identical for Na- and
Ca-based clinoptilolite, the breakthrough exchange capacity was more
than twice as great for Na-clinoptilolite. This observation was
accounted for on the basis of the greater mobility of the smaller
sodium ion in the zeolite pore spaces.
3. While this phase of the study was principally concerned with the
exhaustion performance of clinoptilolite, it became apparent that
regeneration using calcium salts was much less efficient than using
sodium salts. Approximately three times as much regenerant,
expressed on an eauivalent basis, was required for complete ammonia
elution using CaCl2 and Ca(OH)2 compared to NaCl and NaOH. Operation
of clinoptilolite using sodium salts for regeneration resulted in a
greater throughput per cycle to ammonia breakthrough and led to
economies of regeneration.
4. Optimum conditions for ammonia exchange exist between pH 4.0 and pH
8.0 with little variation in ammonia exchange capacities between
these values. However, the ammonia exchange capacity decreased
rapidly outside this range. Results of these tests were predicted
by a model of the exchange reaction and independently determined
selectivity coefficients.
5. The ineffectiveness of ammonia sorption at high pH corroborated
observations that regeneration of an ammonium-based clinoptilolite
is best accomplished at high pH. The exchange model predicted an
ammonia exchange caoacity of 0.53 meq/g at pH 10, 0.08 meq/g at pH
11, and 0.008 meq/g at pH 12 for clinoptilolite in equilibrium with
a solution containing 60 mg/£ Na and 20 mg/i NH3-N. While regene-
ration is not an equilibrium process, these values indicate that
elution of ammonia becomes much more favorable as the pH of the
regenerant is increased.
-------
REGENERATION
Regeneration of clinoptilelite was examined to determine the effects of
flow rate, regenerant pH, and regenerant NaCl concentration on the
regeneration efficiency and on the volume of regenerant required. (Here
regeneration efficiency was defined as the equivalents of NH3-N removed
during exhaustion divided by the equivalents of Na used for regeneration
expressed as a percentage.) Results indicated that the regenerant pH
was the controlling factor in determining the amount of regenerant
required to remove ammonia from clinoptilolite. It was hypothesized
that the unionized ammonia formed at the high pH was able to diffuse
through the zeolite pores more readily than the ammonium ion. This
coupled with the maintenance of a maximum differential between solid
and solution phase ammonium ion concentrations at high pH resulted in
an increased rate of ammonia elution.
Attrition studies were made to determine the stability of clinoptilolite
in the presence of caustic solutions. In small column tests designed
to simulate exhaustion and regeneration cycles, attrition rates of 0.25,
0.35, and 0.55%/cycle were measured for exposure to pH 11.5, 12.0, and
12.5 solutions, respectively. Although the attrition rate appeared to
decrease after exposure to approximately 100 simulated regeneration
cycles, it could not be concluded from these tests that this would lead
to a decrease in the required replacement rate of clinoptilolite.
The following specific conclusions were made with respect to the
regeneration of clinoptilolite:
6. Regeneration of clinoptilolite with caustic solutions is necessary
if the process is to be feasible for ammonia removal. However, the
strength of caustic used for regeneration will be limited by the
attrition of clinoptilolite in caustic solutions.
7. The pH at which clinoptilolite was regenerated affected both the
volume of regenerant required and Degeneration efficiency. Regene-
ration at pH 12.5 was more effective and more efficient than
regeneration at pH 11.5 or 12.0.
8. Increasing the regenerant NaCl concentration beyond a certain value
at a particular pH had no effect on the volume of regenerant
required. For regeneration at pH 12.0 and 12.5, no benefit was
realized by using a salt concentration greater than 0.10 Ib NaCl/
gal. Likewise, increasing the salt concentration beyond 0.17 Ib
NaCl/gal at pH 11.5 produced no improvement in regeneration
performance.
9. Flow rate had no effect on regeneration efficiency over the range
of 4 to 20 BV/hr. Flow rates of 25 BV/hr produced minor impairment
of regeneration performance, but regeneration at a flow of 30 BV/hr
resulted in unacceptable performance judged by the amount of ammonia
eluted per unit volume of regenerant.
-------
in. The rinse requirement varied with the buffering capacity of the
rinse water. Eleven bed volumes of product water was sufficient
to reduce the pH to 10.0, while 35 BV of tap water having a lower
buffer capacity was required for equivalent rinsing. The appli-
cation of small amounts of acid to the bed prior to rinsing reduced
the rinse requirement. However, this reduction was insufficient
to justify the added expense of purchasing and storing acid.
PROCESS PERFORMANCE
Process demonstration studies were conducted at three locations to pro-
vide measures of column performance. Observations made during these
tests may be summarized as follows:
11. The performance of clinoptilolite was related to various operating
variables. Flow rates in the range of 7.5 to 15 BV/hr had no effect
on ammonia effluent values. Ammonia leakage was observed to decrease
with increasing levels of regeneration.
12. An overall ammonia removal of 95.7% was achieved in these tests with
an average effluent ammonia concentration of 0.75 mg/a NH3-N.
Removals at the SERL treatment facility averaged 97.8% with an
average effluent concentration of 0.39 mg/a NH3-N for one test
series and 91.5% with an effluent concentration of 1.7 mg/a NH3-N
for a second series. However, in the second series of runs an
effluent average of 0.94 mg/a NH3-N could have been achieved if
runs had been stopped after a throughput of 135 BV instead of the
180 BV run length which was used. Average effluent concentrations
of 0.20-0.30 mg/£ NH3-N could have been achieved in the first series
of runs for run lengths of 90 BV. Removals in tests at the East
Bay Municipal Utility District pilot plant averaged 94.0% with an
effluent value of 0.71 mg/«, NH3-N. In tests at the Central Contra
Costa Sanitary District an average removal of 97.5% with an average
effluent concentration of 0.50 mg/a NH3-N was achieved.
COST OF AMMONIA REMOVAL
Results obtained in the experimental phases of the study were utilized
in the conceptual design and cost analysis of a 10-mgd clinoptilolite
ion exchange facility. The design included the use of gravity flow
reinforced concrete vessels as ion exchange reactors. Estimates were
based on costs of rapid sand filters. Treatment costs were calculated
for a system where regenerant would be wasted after one use and for a
system in which regenerant would be stripped of ammonia and reused.
The following conclusions were reached:
-------
13. The cost of regeneration where regenerant would be used only once
was strongly dependent on the concentration of salt used in the
regenerant. Costs were least for regeneration at pH 12.5 because
of the higher regeneration efficiency at this pH. The minimum cost
of regeneration was $0.092/1000 gal for a waste having 20 mg/£ NH3-N
using a regenerant containing 0.049 Ib NaCl/gal at pH 12.5. This
corresponded to a regeneration level of 5.2 Ib NaCl/cu ft. Costs
using a regenerant containing only NaOH with the pH raised to 12.5
were slightly higher. Because chemical prices and transportation
costs will vary according to the treatment plant location, the
choice of regenerant composition in this case will depend on the
specific location being considered.
14. When reuse of regenerant was considered, regeneration costs were
relatively insensitive to the NaCl concentration of the regenerant
but were significantly influenced by zeolite replacement costs.
The minimum cost of regeneration was $0.042/1000 gal for a waste
containing 20 mg/£ NH3-N and was achieved using a regenerant composed
of either 0.10 or 0.17 Ib NaCl/1000 gal with sufficient lime to
raise the pH to 11.5. This corresponded to regeneration levels of
18.5 and 24.6 Ib NaCl/cu ft, respectively; however, the amount of
salt required for regenerant makeup corresponded to 1.2 and 1.5 Ib
NaCl/cu ft, respectively.
15. Because the cost of salt is heavily dependent on transportation
charges, the cost of regeneration will vary for areas where the
transport distance is significantly different from the 300 miles
for NaCl and NaOH and 200 miles for CaO assumed in this analysis.
16. The cost of replacement-clinoptilolite was estimated including
400 miles transportation of clinoptilolite. If the zeolite must
be shipped farther, clinoptilolite replacement costs will increase
and regeneration at a lower pH might result in lower treatment costs.
17. The total cost of ammonia removal using clinoptilolite was estimated
to be $0.134/1000 gal where regenerant is used only once and $0.082/
1000 gal where regenerant is reused. Costs for disposal of spent
regenerant solutions were not included in the cost estimate for the
case where regenerant would be wasted after one use. These costs
indicate that the reuse of regenerant by air stripping ammonia from
the spent regenerant solution will be the most feasible method of
operation. However, in areas where transportation charges for
chemicals are minimal and where regenerant solutions can easily be
disposed of, wasting of regenerant solutions after one use might
be competitive with treatment costs where regenerant is reused.
-------
II. RECOMMENDATIONS
Results of this study showed that the ammonia exchange capacity of
clinoptilolite was significantly influenced by the concentration of
competing cations in the wastewater and by the ammonium concentration.
However, it was not possible to develop a multicomponent model suitable
for the prediction of both effective capacity and leakage. For example,
it is expected that increased cation concentrations will not only result
in a decreased ammonia exchange capacity, but that higher ammonia leakage
prior to breakthrough will also result. In addition, the effect of
varying ammonia concentrations on column performance have not been
identified. Ammonia concentrations ranging from less than 10 to more
than 25 mg/n NH3-N might be expected in municipal wastewaters, depending
on the nature of the wastewater and on the types of treatment processes
which precede ammonia removal. Diurnal concentration fluctuations and
the time of their arrival with respect to the ion exchange exhaustion
cycle will also affect column performance. It is recommended that a
systematic study of the relationship between cation composition and
performance be carried out.
Additional factors relating to the effective ammonia exchange capacity
of clinoptilolite concern the influence of process design on the exchange
capacity. The effective ammonia exchange capacity per unit weight of
zeolite will increase with the depth of the bed; however, the extent to
which an increase in bed depth will increase the useful run length is
not precisely known. In addition, little information exists concerning
the change of effective capacity in increasing the diameter/height ratio
of the exchanger bed from experimental to prototype units. Both of
these considerations will be important in the design of clinoptilolite
ion exchange units and are worthy of further investigation.
In portions of this investigation chemically precipitated wastewaters
were applied to clinoptilolite columns without prior filtration. In
these systems the clinoptilolite served as both exchanger and filter.
No impairment of the ion exchange function of the zeolite was observed
due to plugging of the bed or fouling of exchange sites. While some
removal of suspended solids and turbidity were achieved, it is believed
that the addition of another filter medium, perhaps plastic chips, above
the clinoptilolite and the injection of a small amount of alum or
polyelectrolyte ahead of the column would greatly improve the filtration
performance of this unit. By combining two unit processes, a considerable
savings in overall treatment costs would be realized. This matter also
merits further investigation.
While data obtained from laboratory studies in this investigation
demonstrated that a significant loss of clinoptilolite will occur from
exposure to caustic regenerant solutions, longer-term attrition data
-------
from operating systems are needed. This information should be obtained
from cyclically operated units using actual wastewaters as the column
influent.
Estimated costs for ammonia removal using clinoptilolite made in this
study showed that the reuse of regenerant will result in significant
reductions of process operating costs. While previous work has indicated
that regenerant reuse is possible, many factors relating to this aspect
of the process remain to be demonstrated. The exact amounts of caustic
and NaCl which must be added to the regenerant after each cycle of use,
the effect of the accumulation of potassium and calcium in regenerant
solutions on regeneration performance, and the loss of regenerant volume
during each cycle are factors which must be investigated before a more
thorough evaluation of regenerant reuse can be made. In addition more
information is needed concerning the design and operation of towers to
be used for the stripping of solutions containing high concentrations
of ammonia. It is recommended that a study of problems relating to the
reuse of regenerant solutions be made to determine the most effective
methods of minimizing regeneration costs.
-------
III. INTRODUCTION
THE CONCERN FOR AMMONIA IN WASTEWATERS
Traditionally, emphasis in wastewater treatment has been placed on the
removal of biologically degradable organic material, suspended solids,
and floating substances. The water quality objectives requiring such
treatment were to produce a clear effluent, which, when mixed in the
receiving water, would produce no more than acceptable oxygen depletions
and no gross signs of pollution or objectionable odors. As a means of
meeting these objectives, physical and biological treatment developed
in the 1930's generally proved to be a satisfactory solution to pollu-
tion problems. Properly operated, these processes removed sufficient
organic matter to prevent,oxygen depletion in the receiving water,
provided adequate removal of. suspended matter, and generally made it
possible to achieve some degree of control of oil, grease, and other
floating materials. Any removal of nitrogen achieved in these pro-
cesses was incidental to the fulfillment of the treatment objectives.
More recently, increased use of water brought about by a rise in the
standard of living and the pressure of population growth has had the
effect of requiring a reevaluation of waste treatment objectives.
Recent interest in the removal of nitrogen from wastewaters is consis-
tent with current waste treatment objectives of producing a water
which is not detrimental to the environment and of treating to a level
of quality suitable for direct reuse.
Need for Nitrogen Removal
Investigation into the causes of eutrophication have led many workers
to conclude that a nutrient such as nitrogen or phosphorus is likely
to be the growth-limiting factor in many cases. While phosphorus has
most often been implicated as being the major sewage constituent
contributing to eutrophication, the factor which limits algal growth
in a particular water can be correctly defined only in the context of
the local aquatic environment. In studies of eutrophication in Lake
Tahoe [1], data from chemostat assays indicated that nitrogen was the
limiting nutrient. -However, it was pointed out that nitrogen concen-
trations might only be indicative of other substances which do control
algal growth rates.
In developing water quality standards for the Potomac River, the
Department of the Interior has recommended that nitrogen removal be
required at the 240-mgd Blue Plains plant and at the 30-mgd Piscataway
treatment plant in Washington, D. C. [2]. In studies of water quality
in the San Francisco Bay and Delta, questions have been raised
-------
concerning the role of nitrogen in maintaining suitable water quality.
In one study it was concluded from batch assays that nitrogen was the
limiting algal nutrient [3]. However, in other work [4] a nitrogen
mass balance indicated that nitrogen concentrations did not limit algal
growth. Disputes concerning the exact causes of eutrophication and
the role of sewage effluents in stimulating algal growths probaby will
not be resolved for some time. However, it seems unlikely that all
eutrophication problems will be explained by attributing stimulation
to a single factor either nitrogen or phosphorus.
In some locations demands on available water resources have increased
to the extent that water reuse for both industrial and domestic pur-
poses has become economically attractive. If water is to be reused for
municipal supplies, ammonia will adversely affect the efficiency of
chlorination of the water as well as increase chlorine consumption.
While ammonia in concentrations found in sewage is apparently not toxic
to man, the PHS Drinking Water Standards [5] limit nitrate nitrogen to
10 mg/ii because of its role in infant methemoglobinemia. Thus, nitrogen
removal will be necessary wherever wastewaters are to be reused for
drinking purposes. For cases where water is reused for fishing and
recreation, ammonia may be toxic to game fish. An ammonia concentration
of 0.7 mg/a has been reported as toxic to trout by McKee and Wolf_[6].
However, ammonia toxicity is a function of pH, fish species and size,
and other factors in addition to the ammonia concentration. Ammonia
can also cause oxygen depletion in the receiving water as it is
oxidized to nitrate.
Methods of Nitrogen Removal
Methods of removing nitrogen from wastewaters can be grouped as either
biological or chemical-physical processes. General reviews of these
methods have been written by Stern [7], Rohlich [8], McCarty [9], and
Samples [10]. Nitrogen removal achieved in conventional waste treat-
ment processes is primarily due to microbial denitrification. Removals
reported tend to be erratic and are usually not sufficient to meet
water quality criteria where nitrogen removal is required.
Bioloqical processes specifically designed for nitrogen removal include
microbial nitrification-denitrification and algal harvesting. Nitri-
fication-denitrification using the standard activated sludge process
requires a sufficient mean cell residence time to allow nitrifying
bacteria to become established in the system. This usually requires
long aeration periods, thus negating the economic advantages of high-
rate biological systems. Even in a compartmentalized system such as
the one proposed by Barth eit. jil_. [11], treatment periods are long and
problems develop in maintaining individual systems having uniquely
different biological functions. Algal stripping of nitrogen requires
considerable land area compared to other processes. Removals are
dependent on weather conditions, and cell separation is still a costly
and uncertain process [9].
10
-------
Chemical and physical processes for nitrogen removal have the advantages
of being more amenable to control than biological processes and more
adaptable to the fluctuating flows and concentrations inherent to muni-
cipal wastewater systems. While air stripping has the potential of
being the least expensive method of removing nitrogen, the process
becomes less efficient in cold weather and is impractical to use at all
when the temperature drops below freezing. This precludes serious
consideration of air stripping in most northern parts of the country
until improved designs are available. An additional problem with the
stripping process is that scale deposition occurs in the tower as a
result of the high pH required for ammonia removal.
Nitrogen removal using conventional ion exchange resins is unattractive
because of the preference of these exchangers for ions other than the
ammonium and nitrate ions. In addition, disposal of the regenerant
brine poses a problem in some locations. However, work by Ames and
Mercer [12-14] demonstrated that the ammonium selective zeolite,
clinoptilolite, is effective in removing ammonia from wastewaters. By
regenerating the zeolite with a high pH brine solution, it is possible
to strip the ammonia from the regenerant and reuse the regenerant
solution, thus reducing regenerant costs and minimizing the problem
of brine disposal. The work by Mercer showed that ammonia removals
of greater than 95% can be achieved using clinoptilolite.
As a unit process, ion exchange is easily controlled to achieve almost
any desired product quality. The efficiency of the process is 'not
significantly impaired at temperatures usually encountered in the
United States. Ion exchange equipment can be automatically controlled
and requires only occasional inspection and maintenance.
While a number of processes have been proposed to accomplish nitrogen
removal, full-scale treatment units are still in the planning and testing
stages. Perhaps the most obvious reasons for this is that the presence
of nitrogen in wastewaters has become objectionable only in the last
few years. At the present time plans for nitrogen removal facilities i
are underway in several locations where existing or anticipated uses of
a waterway have dictated the need for this type of treatment. However,
an equally compelling reason for the absence of nitrogen removal in
wastewater treatment has been a combination of high costs and a lack
of adequate technology applicable to process design and operation.
Studies of nitrogen removal methods have generally been more concerned
with proving that a particular process works than with defining opera-
tional problems and treatment costs. Although very recent work has
begun to answer some of these questions, a definite need exists for
more information applicable to process operation and design.
This investigation will concentrate on defining various operational
parameters necessary to efficiently remove ammonia by ion exchange
using clinoptilolite. By studying different aspects of column operation,
11
-------
it will be possible to evaluate more clearly the applicability of this
process to operating conditions and to more accurately predict column
performance and costs.
OBJECTIVES
This study was concerned with the optimization of ammonia removal from
wastewaters by ion exchange using clinoptilolite. The general objec-
tive of the investigation was to establish the conditions under which
clinoptilolite can be used most effectively and economically for ammonia
removal- The fulfillment of this objective involved the experi-
mental determination of certain ion exchange properties of clinopti-
lolite which are important in its use for ammonia removal. These
results were used to define optimum operating conditions for clinopti-
lolite exchangers and to estimate the cost of ammonia removal using
clinoptilolite. The necessary assumptions made in arriving at valid
cost estimates are listed where appropriate. However, the reader is
reminded that the cost of a single process is not independent of the
system in which it is placed.
The study had the following specific objectives:
1. Determination of factors influencing the exhaustion performance of
clinoptilolite including the effect of pH on ammonium exchange.
2. Optimization of the regeneration cycle of clinoptilolite to estab-
lish the least-cost method of operation.
3. Determination of the performance of clinoptilolite under normal
conditions by operating test facilities at several wastewater
treatment plants.
4. Development of a basis for the design of clinoptilolite exchange
systems.
5. Establishment of the costs associated with the construction and
operation of processes for the removal of ammonia using clinopti-
lolite.
A brief review of the contents of this report will serve to acquaint
the reader with its organization. In Chapter IV the general definitions
and fundamentals of ion exchange theory, past work concerned with the
removal of ammonia by ion exchange, and the properties of clinoptilolite
are reviewed. While this material provides a general background for
the rest of the report, it is not essential for an understanding of
the conclusions of the study. However, the material reviewed in this
chapter was chosen to complement observations made later in the report.
12
-------
Chapter V considers the potential for the use of clinoptilolite for
ammonia removal and sets forth the rationale for the specific phases
of the experimental work. In Chapter VI the equipment used in the
study is described and the analytical methods and operating procedures
are outlined. The findings of experimental investigations are presented
and discussed in Chapters VII, VIII, and IX. Work presented in Chapter
VII concerns the effect of pH, ionic form of the zeolite, and water
composition on the exhaustion performance of clinoptilolite. Regen-
eration of clinoptilolite, considered in Chapter VIII, includes \
individual regeneration runs, studies of clinoptilolite attrition in
the presence of caustic solutions, and column rinsing tests. In
Chapter IX the performance of clinoptilolite using three different
wastewaters is presented. Results of the study are used in discussing
the design of ion exchange units and the cost of ammonia removal by
this process in Chapter X.
13
-------
IV. ION EXCHANGE CHARACTERISTICS OF THE ZEOLITES
ION EXCHANGE THEORY
Equilibrium Relationships
Ion exchange may be defined as the stoichiometric, reversible exchange
of ions between a liquid and solid which produces no significant
changes in the structure of the solid. The mass action equilibria
expression provides a useful model for ion exchange behavior. In a
binary system, the reaction,
bA+a + aBZ, £ bAZa + aB+b (1)
D d
expresses the reversible equilibrium where a and b are the valences of
ions A and B, respectively, and Z is the exchange site in the solid.
Helfferich [15] has referred to ions having charge signs opposite that
of the exchange sites as "counter ions" and ions having the same charge
sign as the exchange site "co-ions." This reaction may be expressed as
the equilibrium constant,
00
k
(O
in which (aU, (a)fl^ , etc. are the activities of the various species.
Because of tne difficulty in measuring activities, especially in the
solid phase, it is convenient to use concentrations uncorrected for
activity. In doing so, the equilibrium constant in Equation 2 varies
with concentration and has been termed the "selectivity coefficient"
by Helfferich [15],
where q is the solid phase ionic concentration in meq/g and c is the
solution phase concentration in meq/&. Alternatively, the selectivity
coefficient can be expressed in terms of dimensionless concentrations,
15
-------
a-b
These variables are expressed in terms of the total solution concen-
tration, C0, in meq/£ and the total exchange capacity, Q, in meq/g.
Thus, x = c/C0 and y = q/Q.
In performing calculations with equilibrium data it is frequently
desirable to make corrections for solution and solid phase activities.
In dilute solutions, when ions A and B have the same valence, the
solution phase activity coefficients may be taken as unity in most
cases. However, for ions of unequal valence, the distribution of these
ions depends on the relative dilution of the two species. Activity
corrections for the solution phase in dilute solutions is usually made
using the Debye-Huckel equation [16]. However, activity corrections
in the solid phase are more difficult. If the exchanger could be
considered as an independent solid phase, activity in the particle
could be taken as constant and equal to unity. However, it is more
correct to view the components of the solid phase as forming miscible
solid solutions with one another [17]. In addition, heterogeneity of
the particle matrix makes a single description of exchange sites
tenuous. Corrections which have been proposed include considering
solid phase activity as a function of the number of ions per unit
weight of exchanger or as a function of the mole fraction of an ion
in the solid phase [18]. Baetsle [19] reviewed several more complex
methods of correcting for solid phase activity which consider the
exchanger to be an orderly system of identical sites. While the problem
of solid phase activities might be overcome in binary systems by
determining values of the selectivity coefficient over a range of solid
phase compositions, the problem is more complex in multicomponent
systems in which the solid composition is difficult to define. Because
of the complex manner in which activities vary, no method of making
solid phase corrections has gained wide acceptance.
The preference of an ion exchanger for one ion relative to another in
binary systems is often expressed as the "separation factor,"
A _ /3\ /c\ /y\ /x\ ,_,
aB - lei (aI = xA (vL (5)
Because the numerical value of the separation factor is not affected by
the choice of concentration units, equilibrium data are often expressed
in this way. If the equivalent fraction of ion A in the solid phase,
y/\, is plotted against the equivalent fraction of A in the solution, XA:
16
-------
three cases can be identified corresponding to a < 1, a = 1, and a > 1
as shown in Figure 1. Isotherms which are concave upward, a < 1, are
designated as being "unfavorable" to the uptake of ion A; those which
fall along the rising diagonal, a = 1, are termed "linear" and exhibit
no preference for ion A or B; and curves which are concave downward,
a > 1, are referred to as "favorable" isotherms since the solid prefers
ion A to ion B [20]. Ion exchange operations almost always are concerned
3
o-
UJ
'8.0 0.10.2 0.3 0.4 0.50.6 0.7 0.80.9 1.6
Equivalent Ion Fraction in Solution Phase, XA
FIGURE 1. GENERALIZED ION EXCHANGE ISOTHERMS
with systems in which the ion of concern has a separation factor greater
than unity during exhaustion.
COLUMN THEORY AND RATE PHENOMENA
Most ion exchange operations are carried out in columns as a much
larger percentage of the capacity of the ion exchanger can be utilized
17
-------
than with batch methods. The performance of column processes will
depend on the properties of the particular ion exchanger, the composi-
tion of the column influent, and the operating conditions and process
arrangement. Over the past 20 years, a considerable body of theory
has been developed to describe column performance. While these
theories are frequently not directly applicable to ion exchange
applications, they can be used to provide a better understanding of
column behavior and to predict the response of ion exchange processes
to changes in design and operating procedures.
In the previous section it was mentioned that in most applications of
ion exchange, equilibrium is necessarily favorable with respect to the
ion being removed. In this case the exchange zone will establish itself
and move at a constant speed through the column provided that flow
through the column and the composition of the feed remain constant.
This type of behavior is referred to as "constant pattern" because
the sharpening effect of equilibrium and the dispersive effects of
the rate of exchange tend to balance each other. This results in the
exchange front maintaining a constant shape as it moves through the
column. The boundary is said to be "self-sharpening" because it
becomes steeper relative to the length of the column as the exchange
front passes through a progressively deeper section of the bed. For
this type of column behavior, increasing the length of the column will
have the effect of increasing the breakthrough capacity of the column.
If the separation factor is less than one, a "proportionate pattern" is
attained because the exchanging ion is attracted to the exchange sites
in proportion to the solution phase concentration of that ion. It can
be imagined that the exchanging ions in this case have to "push" them-
selves onto the exchanger, and the less the solution phase ion concen-
tration, the less they are forced onto the exchanger. This type of
boundary is "nonsharpening" as the exchange front becomes more diffuse
as the column length is increased [21]. For the case where a = 1, the
boundary becomes neither steeper nor more diffuse relative to column
length as the front passes through the column.
Rate Processes
Models describing the variation of effluent concentration as a function
of throughput are concerned with three factors: equilibrium, stoichio-
metry, and rate. The rate of exchange is governed by one or more
diffusional steps which have been listed as: 1) external fluid phase
or film diffusion, 2) fluid phase pore diffusion, 3) reaction at the
liquid-solid interface, 4) solid phase internal diffusion, and 5) mixing
by molecular diffusion or axial dispersion [20,22]. These mechanisms
will affect column behavior, either individually or In combination with
one another. The reaction rate at the phase interface is usually very
rapid and does not limit the rate of exchange [23]. While molecular
diffusion probably influences column performance, the effects are
usually small except at very low flows. Axial dispersion increases
with the linear velocity through the column.
18
-------
Film Diffusion. The transport rate of ions between the bulk fluid
phase and the exchanger particles can be estimated from the relation-
ship,
dy_
dt
2.62 (D- F/S)
0.5
J.5
QPI
(x-x*)
(6)
where Df is the fluid phase diffusivity in sq cm/sec, F is the volu-
metric flow rate in nu/sec, S is the column cross-sectional area in
sq cm, dp is the particle diameter in cm, pt, is the bulk density in
g/cu cm, t is time in sec, and x* is the equilibrium value of x [20],
All abbreviations and nomenclature used in this report including
appropriate units are summarized in the Glossary, Chapter XIII. From
this relationship it is apparent that film diffusion will tend to be
limiting when exchanger particles are small, when the solution phase
concentration is low, and when the superficial velocity, F/S, through
the column is small.
Pore Diffusion. Pore diffusion in the fluid phase has been represented
by the following relationship [20]:
dy _ ypore pore
dt ,2
y*-y
Qpb [1 + (a-l)y]
1/2
(7)
Here, ^nore 1>s a correction factor, Dpore is the pore diffusion
coefficient in sq cm/sec, e is the voids ratio, y* is the equilibrium
value of y, and other variables are as previosuly defined. Although
pore diffusion occurs in the fluid phase, this equation indicates that
its rate is primarily a function of solid phase concentrations. As
the equation indicates, pore diffusion is likely to control mass
transfer rates where particles are large and solid phase concentrations
are near the equilibrium values.
Solid Diffusion. Solid phase diffusion includes diffusion through a
homogeneous, permeable solid which results from a concentration gradient
within the particle. Several models have been proposed for solid
diffusion, but the closest approximation is given by the quadratic
driving force model of Vermeulen [24]:
dy . \ 60
dt .2
y*2-y2
2y
(8)
19
-------
where Vn "is a correction factor, Dp is the solid phase diffusivity in
sq cm/sec, and the other variables are as defined previously. Solid
diffusion controlled reactions are the same function of particle
diameter as those for pore diffusion, but the rate of solid diffusion
is seen to be a different function of the solid phase concentrations.
However, as in the case of pore diffusion, solid diffusion tends_to
be limiting when particle phase concentrations are near equilibrium
values. The solid phase diffusivity is affected by the nature of the
exchanging ions as well as by the varying ionic composition of the
particles [15,22].
Rate Controlling Mechanisms. The rate which limits ion exchange in
column operations may be a result of more than one of the above men-
tioned diffusional steps and, in addition, the rate controlling mechanism
may change as the column becomes more saturated. For example, near the
rising limb of a breakthrough curve the solid phase concentration
gradient will be high while the solution phase gradient will be low
near the bottom of a column. Thus, film diffusion tends to control the
initial portions of the breakthrough. On the other hand, near the
upper end of the breakthrough curve, the solid phase gradient will be
low and the solution phase gradient high; thus, the tendency is toward
a shift in the rate-controlling mechanism to either pore or solid
diffusion. It should be noted that these tendencies are relative to
the particular exchanger being used and that it is entirely possible
that pore or solid diffusion could predominate throughout the break-
through. In this case film diffusion would contribute more to the
diffusion resistance at the beginning of the breakthrough than at the
end, but would, at neither time, control the rate of exchange.
Because the diffusional mechanisms which affect column behavior act
together, either in series or parallel, methods have been developed to
account for this in predicting column behavior [25]. For combined
fluid phase or particle phase resistances, the effective rates are
obtained by adding individual resistances if they act in parallel or
by adding the reciprocal resistances to obtain the reciprocal combined
resistance for series combination. In the case of inorganic zeolites
having a bimodal internal structure consisting of individual crystal-
lites bound together into larger particles, the exchange rate will be
governed by a series combination of pore diffusion and particle
diffusion, with the crystallite diameter replacing the particle
diameter, dp, in Equation 8 [20].
While the rate mechanisms discussed above are valid for ion exchange
reactions, the equations which relate these rates to column and particle
variables are limited to cases which can be described mathematically
without undue complexity. These equations apply to: 1) the average
concentration in the neighborhood of a single particle, 2) two-
component exchange, and 3) spherical particles of uniform size.
20
-------
Column Theory
Reaction Kinetic Model. Numerous theories have been proposed to
describe column performance under various conditions. These methods
make it possible to predict the column effluent concentration profile
from properties of the exchanger which can be measured by laboratory
batch-type tests. One approach is that taken by Glueckauf [21] in
which column performance was described in terms of either fluid or
particle diffusion. Thomas [26] developed a solution for systems in
which the rate is described by a second-order rate equation. These
equations have been used infrequently because, as discussed in a
previous section, exchange rate is seldom controlled by the reaction
rate at phase boundaries and because Thomas' solution involved the use
of functions which are difficult to evaluate. Hiester and Vermeulen
[27] utilized this method to obtain solutions in the form of families
of concentration-history curves. Solutions were obtained in terms of
two dimensionless parameters: N, the number of transfer units analogous
to the number of theoretical plates in distillation column, and T, the
throughput parameter. N can be defined generally as,
. P(F/S)C /hS
where c and d are functions of the rate mechanism involved as defined
in Equations 6, 7, and 8 and h is the column height. T is the ratio
of meq of ions passed through the column to the capacity of the column.
In the ideal case, complete saturation of the column corresponds to
T = 1. This solution was made more workable by allowing rates of fluid
or particle phase diffusion to be expressed as an effective reaction
rate.
Assumptions made in the Hi ester-Vermeulen solution include: 1) a
symmetric system (although a correction procedure was outlined for
ions of unequal valence), 2) binary exchange, 3) homogenous particle
structure including uniform distribution and nature of exchange sites,
4) uniform feed concentration, 5) uniform flow rate, 6) uniform initial
sorbent concentration within the particles, 7) constant temperature,
and 8) constant separation factor.
From this discussion several qualitative observations may be made
concerning column behavior. In the Hiester-Vermeulen model, the
effluent concentration is a function of three parameters: the separa-
tion factor, a; the number of transfer units, N; and the throughput
parameter, T. A sharp breakthrough curve, and therefore a greater
utilization of column exchange capacity, is favored by large values of
a and N. The separation factor will be determined by the exchanger
being used. In this respect, the preference of the exchanger for the
21
-------
counter ion in the feed should be large, but not so large as to make
regeneration difficult. From Equation 9 it is apparent that a large
N is favored by a small velocity of fluid through the bed, a large
bed depth, and a small particle diameter. The diffusivity, D, is
fixed, and little can be done to vary it once the exchange material is
chosen. In considering the rate mechanisms given in Equations 6, 7,
and 8, it is seen that increasing the linear velocity through the bed
increases the exchange rate only in the case of film diffusion. In
this case, increased flow rates act to reduce the thickness of the
film surrounding the particles. Increasing the particle size is seen
to adversely affect the rate of pore and solid diffusion more than film
diffusion. The curves shown by Hiester and Vermeulen reveal that sharp
breakthrough cannot be achieved even for very large values of a if N
is less than about 5.
Pore- and Solid-Diffusion Limited Systems. For particle phase diffu-
sivities, the reaction-kinetic solution of Hiester and Vermeulen assumes
a rate governed by a linear driving force. The quadratic driving force
model of Vermeulen given in Equation 8 provides a better fit to observed
diffusion rates. A solution utilizing this model, as well as a solu-
tion for pore diffusion limited systems, was developed by Hall et al.
[28]. It applies specifically to constant pattern conditions and to
cases where the equilibrium conforms to a Langmuir isotherm. The
solution lacks flexibility in that cases in which the rate of exchange
is governed by combined mechanisms cannot be treated by this model.
However, it is thought that fluid phase resistance is of little signi-
ficance in ion exchangers having uniform internal pore structures such
as the zeolites.
Multicomponent Systems
The foregoing discussion of ion exchange theory can be applied quanti-
tatively only to binary systems. Unfortunately, any application of ion
exchange in wastewater treatment will involve multicomponent equilibria.
The complexities introduced in dealing with multicomponent systems
include having to cope with a different separation factor for each pair
of exchanging species and, in the case of solid diffusion, a different
diffusion rate for each pair of ions.
For equilibrium relationships, it is possible to solve for the multi-
component equilibria knowing the selectivity coefficients for each ion
pair. However, as pointed out above, in systems involving species of
unequal valence, activity corrections would be necessary. Dranoff and
Lapidus [29] have found that ternary equilibria can be predicted from
binary equilibria when the data are treated as adsorption isotherms and
when multivalent cations are treated as an equivalent concentration of
univalent ions. However, for mass-action equilibria, the lack of
suitable solid-phase activity correction methods prohibits accurate
prediction of multicomponent equilibria from binary data.
22
-------
For the general case of multicomponent systems having nonlinear equili-
bria, no method has yet been developed for determining the rate depen-
dent, breakthrough behavior. Until recently very little theoretical
consideration .of multicomponent behavior in fixed beds had appeared in
the literature. Klein e£ aj_. [30,31] discussed the considerations
involved in equilibrium calculations and presented a method of deter-
mining concentration profiles in constant selectivity coefficient
systems. This method is based on equilibrium operation and thus predicts
ideal process performance.
Despite the fact that established theory cannot yet deal explicitly with
the complex problems of ion exchange encountered in wastewater treat-.
ment, it is useful as a qualitative guide for understanding and inter-
preting column phenomena. The very fact that exchange column behavior
is so complex is a compelling reason for having a basic theoretical
background to aid in the interpretation of ion exchange performance
and to guide the design and operation of ion exchange systems.
AMMONIA REMOVAL BY ION EXCHANGE
Use of Conventional Resins
Perhaps the first use of ion exchange to remove ammonia from wastewater
was the zeolite filter in the Guggenheim process [32,33]. Using a
"preferred type of zeolite," between 66 and 93% removal of ammonia was
achieved at a loading rate of about 20 BV/hr (BV/hr x 0.125 = gal/cu
ft-min). The amount of regenerant used corresponded to approximately
64 eq NaCl/eq NH3-N removed. Ammonia from the spent regenerant was
removed by air stripping in a packed tower and the regenerant reused.
The zeolite also acted as a filter for the chemically precipitated
waste, but the resulting headless occasionally made it necessary to
backwash the bed before ammonia breakthrough was reached. No mention
was made of a loss of capacity of the zeolite due to clogging or
organic fouling.
Nesselson [34] investigated the use of Amber!ite IR-120 and Nalcite
HCR strong acid exchange resins for ammonia removal from an activated
sludge effluent. Nalcite HCR, which showed the better performance,
required 8.8 Ib NaCl/lb cations removed for regeneration. Regeneration
was accomplished using .25 to 35 Ib NaCl/cu ft resin with the regenerant
volume amounting to a minimum of 5.8% of the product water volume.
Flow rates in exhaustion ranged from 48 to 67 BV/hr. Because of
unfavorable selectivity with respect to the ammonium ion, the ammonia
exchange capacity of the resin was only about 0.11 meq/nu (corresponding
to approximately 0.27 meq/g.or 2.4 kgr CaC03/cu ft) out of a total
operating exchange capacity of 0.78 meq/nu (1.95 meq/g or 17.1 kgr/cu
ft as CaC03) with a water having a hardness of 380 mq/n as CaC03. The
regenerant requirement was approximately 20 eq NaCl/eq NH3-N removed.
23
-------
With a hardness of 142 rag/;, as CaC03, a 32% increase in ammonium
exchange capacity was realized.
Tests made to determine the effect of prefiltration of the activated
sludge effluent on column performance showed that unfiltered effluent
could be applied to the columns without adversely affecting ammonia
removal efficiency. While no problems of excessive headless through
the bed were reported, the throughput to a breakthrough of 1 mg/x, NH3-N
averaged only 100 BV/run. For the design of an ammonia removal unit,
Nesselson recommended removal of temporary hardness using lime and
regeneration using NaCl with recovery of ammonia by air stripping at
a high oH and subsequent reuse of the regenerant. It was estimated
that salt costs alone would be $0.26/1000 gal.
The removal of ammonia using Duolite C-25 resin was investigated at
the South Tahoe Public Utility District [35,36]. Because the hardness
of Lake Tahoe sewage was much less than in the sewage used by Nesselson
- 55 as opposed to 380 mg/i as CaC03 - ion exchange seemed more feasible
at Lake Tahoe. The resin column was exhausted downflow at 24 BV/hr
until an ammonia breakthrough of 1 mg/£, NH3-N was reached. The column
was regenerated with 10 Ib NaCl/cu ft in a 10% solution at a rate of
1.6 BV/hr. This corresponded to the use of 5 eq NaCl/eq NH3-N removed.
Exhaustion studies showed that an average of 365 BV could be treated
prior to breakthrough which represented an increase of 350% over
previous investigations. It was estimated that salt costs would be
$0.04/1000 gal based on regeneration with 10 Ib NaCl/cu ft. For this
reason and because of anticipated problems of brine disposal, further
investigation of the ion exchange process was not attempted.
In these studies it is significant to note the strong dependence of the
ammonia exchange capacity of the resin on water hardness. In Nesselson's
experiments only about 15% of the exchange sites were used for ammonia
exchange. Nesselson observed a significant increase in ammonia exchange
capacity as the NH3-N/total cation ratio increased. In the studies at
Lake Tahoe using a very low hardness sewage, the throughput to a break-
through of 1 mg/£ NH3-N was more than three times as great as that
achieved in Nesselson's experiments.
Although specific data regarding regeneration efficiency, i.e., equi-
valents of NH3-N removed during exhaustion divided by the equivalents
of NaCl used for regeneration, were not given in these reports, values
calculated from other data indicate that regeneration efficiency was
only 1.5 to 5'".- for the Guggenheim process and in Nesselson's studies,
while data reported in the Tahoe report indicated a value of about 20%.
Because of the small fraction of exchange sites occupied by ammonium
ions, lower regeneration efficiencies would be expected for ammonia
removal than for softening. Reuse of regenerant was investigated only
by Gleason and Loonam. Regenerant reuse was not considered in cost
estimates made by Nesselson or in the Tahoe report.
24
-------
Use of Clinpptilolite
The first suggested use of clinoptilolite for ammonia removal was
contained in a paper by Ames [12] who determined the selectivity series
for various zeolites and found that clinoptilolite and a synthetic
zeolite, AW 400, produced by the Linde Company were the most promising
for ammonia removal from wastewaters. Subsequent reports [13,37] done
under the direction of Mercer concentration on column studies using
clinoptilolite for ammonia removal because of its favorable ammonium
ion selectivity and potential low cost. Tests were made using 20 x 50
mesh material. In order to evaluate column performance using secondary
effluent without pretreatment, tests were made in 2-in. columns with
upflow loadings of 16.6 BV/hr. High leakage of approximately 1 mg/£
NH3-N and shallow breakthroughs observed in these runs were attributed
to channeling of the flow through the bed. Organic fouling resulted in
a 25% loss of capacity during the two upflow runs.
In preliminary studies, the columns were regenerated upflow using a
lime slurry; however, saturated lime solutions containing added sodium
and calcium salts as counter ions were used in later studies. The
saturated solution performed as well as the slurry and eliminated
deposition of lime particles in the bed. During most of these studies,
the regenerant flow rate was 10 BV/hr upflow, although flows of 30 BV/hr
were used in two instances. The required regenerant volume was reduced
significantly by adding NaCl to the saturated lime solution. In com-
paring elution curves using various regenerants, approximately 50 BV
regenerant was required for complete ammonia elution using a 0.05 M
CaCl2 solution (0.046 Ib CaCl2/gal) saturated with lime while only
27 BV and 20 BV were required for saturated lime solutions containing
0.035 M CaCl2 plus 0.03 M NaCl and 0.1 M NaCl (0.049 Ib NaCl/gal),
respectively. Exhaustion concentration histories were compared for
columns previously regenerated with 4.5 g/a lime slurries (0.038 Ib
lime/gal) containing no NaCl and 0.1 M NaCl (0.049 Ib NaCl/gal).
Leakage was approximately the same for both cases, but the breakthrough
capacity to 1 mg/n NH3-N leakage was 20% greater for the column
regenerated with added NaCl. Ammonia exchange capacities were 0.20
meq/g (0.15 meq/ma or 3.3 kgr/cu ft as CaC03) and 0.23 meq/g (0.17 meq/
m£ or 3.7 kgr/cu ft as CaC03) for the columns regenerated with the lime
slurry and with the lime-salt combination, respectively. No attempt was
made to determine whether the increased capacity was due to more complete
regeneration using the lime-salt solution or to more favorable NH4-Na
exchange kinetics during exhaustion. Regeneration using 0.015 M NaOH
was demonstrated to require 35 BV for complete removal of ammonia.
Studies were conducted to elucidate the effectiveness of recycled
regenerant. Ammonia was removed from the regenerant solution by air
stripping in a laboratory column. Regenerant used for three cycles was
as effective as fresh regenerant.
Additional tests were performed in three locations using a mobile
demonstration plant [14]. The breakthrough capacity to 1 mg/£ NH3-N
25
-------
for flow rates between 6.5 and 9.7 BV/hr with influent ammonia concen-
trations of 15 to 17 mg/i NH3-N ranged from 100 to 120 BV. The ammonia
leakage prior to breakthrough averaged 0.7 mg/«. NH3-N. Operation of
two columns in series resulted in a 60% increase in the utilized ammonia
exchange capacity per column. However, no data were presented to
demonstrate the relative regenerant requirements for single columns and
those exhausted in series operation.
A batch regeneration technique was studied in which 2 to 4 BV_of
regenerant were recycled through the column until saturated with ammonia.
Subsequently, the spent regenerant was stripped of ammonia and reused.
The results of these studies indicate that clinoptilolite has much
potential for removing ammonia from wastewater. Because of its greater
ammonia selectivity, clinoptilolite has an advantage over conventional
resins for ammonia removal. Tests in the Battelle demonstration unit
showed that ammonia removals greater than 95% can be achieved using
clinoptilolite. Regeneration was accomplished using lime-salt solutions
which were effectively reused after ammonia was removed from the solution
by air stripping.
ION EXCHANGE PROPERTIES OF THE ZEOLITES
In order to be used in large-scale process applications, an ion exchange
material must meet certain basic requirements related to both its
physical and chemical properties. Specifically, the material must be
available in abundant supply and at a reasonable price; it must possess
good physical stability toward abrasion and be chemically stable toward
the fluids with which it will come in contact; it must have a high
exchange capacity and exhibit favorable selectivity for the intended
use; and the exchange kinetics should allow a large part of the equili-
brium capacity to be utilized in column operations. In this section the
properties of zeolites will be discussed and compared to conventional
ion exchange resins.
Structural Properties
Stoichiometrically the zeolites are derived from the formula (Si02)n
by the periodic substitution of Al for Si atoms with sufficient alkali
metal and alkaline earth cations to maintain electroneutrality. The
zeolite framework consists of tetrahedral SiO^ and Al-O^5 combined into
crystal structures. Oxygen atoms are shared among groups in such a way
that cavities are formed within the zeolite structure which accommodate
cations needed to maintain electroneutrality within the crystal lattice.
Passageways and restrictions between exchange sites in zeolites are
referred to as channels, while the enlargements containing the exchange
sites are called ion cages or cavities. Openings between zeolite
26
-------
crystals but within the same aggregate particle are referred to as
pores. These terms are illustrated in Figure 2. Although zeolites
may exist as fibrous, lamellar, or rigid three-dimensional structures,
the members of the latter group have been studied most extensively for
their ion exchange properties [38].
The exchange capacity of zeolites is a function both of the Al/Si
ratio of the zeolite framework and the access of the interstitial
fluid to the ion cages. Exchange capacities of selected zeolites
calculated from hydrated formula weights reported by Amphlett [38]
ranged from 2.30 to 5.30 meq/g. Capacities reported by Sherry [39]
for similar zeolites were based on the anhydrous weight and were
correspondingly higher. Ames measured the exchange capacity of chaba-
zite and obtained a capacity of 2.59 meq/g for sodium and potassium
ions, while the cesium capacity was 2.20 meq/g. This compared to
calculated capacities of 4.95 and 4.00 meq/g given by Sherry and
Amphlett, respectively.
Ion Sieve Properties
The extent to which the cations within the zeolite may be exchanged
depends on the nature of the zeolite cages specifically, the size of
the opening and the degree to which the channels are interconnected.
Because the rigid, three-dimensional crystal lattice contains definite
sized openings into the ion cages, zeolites exhibit ion sieve proper-
ties to a much greater extent than conventional ion exchange resins.
The extent of ion sieving exhibited by a zeolite depends primarily on
the size of the channel opening. In the more dense zeolites some ions
are completely excluded from the channels. However, in the "open"
zeolites, all alkali metal and alkaline earth cations have access to
the passageways, although partial sieving action is observed due to the
stripping of hydrated water from the ion as it enters the zeolite
cavity [39]. In this case the preference of the zeolite for an ion
is a function of the energy with which the water of hydration is bound
to both the cation and to the zeolite, the size of the ion, and its
valence. If lattice forces are strong enough to overcome the energy
of solvation, the ion may have access to the zeolite channels on the
basis of the crystal radius of the ion [15].
In addition, steric considerations of internal pore geometry affect the
capacity as well as the selectivity of the zeolite. The extent of
exchange between ion pairs may depend on whether an ion may enter one
end of a channel while the exchanged ion leaves by the other end, or
whether the two ions must pass at one point within the channel. These
questions were considered in Hey's theory of ion exchange mechanism
discussed by Marshall [40]. Channel constrictions between zeolite
cavities also influence the preference of a zeolite for univalent ions.
Bivalent ions cannot easily occupy positions at channel restrictions
for steric reasons; however, failure to completely satisfy the
27
-------
Interparticle
pore
Aggregate
zeolite particle
20 x 50 mesh
(0.30-0.84 mrr)
Zeolite channel or
irvtracrystal line oore
Ion cage or cavity
Zeolite
crystallite;
. < 5 u dia.
Zeolite-
inter crystal line
pores
FIGURE 2. IDEALIZED ZEOLITE PARTICLES
28
-------
electroneutrality requirements of exchange sites adjacent to the
restriction leads to instability resulting in the preference of the
exchanger for univalent ions [41]. Similarly the volume of the ion
cage may be insufficient to accommodate the required number of ions
needed to satisfy the charges within the cage. In this case only
partial exchange of the particular ion would be possible [38]. Both
the degree of channel restriction between cavities and the anion site
separation will influence ionic equilibria in these cases. In channels
restricted only slightly, it may be possible for multivalent ions to
situate themselves in such a way to satisfy exchange sites on either
side of the restriction. The anion site separation is partially a
function of the exchange capacity of the zeolite and is determined by
the number of AlO^ groups per unit volume of exchanger.
Another factor which complicates the explanation of zeolite selectivity
is the existence of more than one type of cavity in a zeolite. When
this occurs, each size cavity may exhibit different ion sieving
properties resulting in only partial exchange of some ions. While all
functional groups in zeolites originate from the substitution of AlO^
groups within the crystal structure, the functionality of these groups
varies according to the structure of the particular zeolite and, more
specifically, according to the particular cage composition [39].
The structural properties of zeolites which affect selectivity also
influence the rate of diffusion of ions through the zeolite. Exchanging
ions proceed to exchange sites by diffusion through the liquid in the
pores and through the channels of the zeolite framework. The rate of
exchange in zeolites is usually controlled by diffusion within the
particle which can be l.O6 times slower in zeolites than in organic
resins [38]. However, Ames states that diffusion through zeolites
composed of crystallites cemented together into aggregate particles
is much faster than diffusion through the coarsely crystalline zeolites
[42]. These crystallites are usually less than 5 microns in diameter.
In such zeolites the rate of exchange would tend to be limited more by
diffusion through the pores surrounding the crystallites than by
diffusion through the zeolite channels. That particle phase diffusion
would be favored in zeolites can be seen from a comparison of average
pore diameters for zeolites and organic resins. The minimum free
diameter of zeolite channels reported by Barrer [43] ranged from 2.2
to approximately 9 A. The average pore diameter of strong acid resins
reported by Kunin [44] averaged 12 A for 8% crosslinking. Thus,
zeolite channels are not only smaller than the pores in strong acid
exchangers, they are also not able to swell to accommodate larger
ions. From the structural properties of zeolites considered above, the
relationship between zeolite selectivity and the rate of exchange
becomes more evident. The definite sized channels and cavities in the
zeolite framework which are largely responsible for the unusual
selectivity of zeolites result in poor kinetics of exchange compared to
conventional resins. In applications where the selectivity of a
particular zeolite is desirable, a sacrifice must be^madeJjL tfirms_jifL-
exchange kinetics.
29
-------
In summary, the exchange capacity of a zeolite is a function of the
Al/Si ratio of the zeolite framework and the access of ions to the
exchange sites. The unusual selectivity exhibited by some zeolites is
primarily a result of specific channel geometry as related to cation
size and charge, and the binding energy of the structural water in the
zeolite lattice. The rate of exchange in zeolites is usually controlled
by particle phase diffusivities and is less than the rate in synthetic
resins. However, when selectivity for a particular ion is desired, the
rate of exchange may be large enough to make feasible the use of a
particular zeolite.
Properties of Clinoptilolite
Classification. Mineralogically, the zeolites are classified as a
family in the silicate group. Specifically, they are defined as hydrated
alumino-silicates of uni- and bivalent bases which can be reversibly
dehydrated to varying degrees without undergoing a change in crystal
structure and which are capable of undergoing cation exchange [38]. The
general composition of zeolites is given by the formula (M,N2)0-A1203-
nSi02-mH20 where M and N are, respectively, the alkali metal and
alkaline earth counter ions present in the zeolite cavities [38].
Clinoptilolite was first identified by Schaller [45] and has been the
subject of several mineralogical inquiries since that time. Kostov
[46] recognized Clinoptilolite as a heulandite rich in silica. However,
Berman [47] and Pabst [48] both classed Clinoptilolite as a distinct
mineral in the heulandite group. More recently, Mumpton [49] found
that while the properties of Clinoptilolite approximated those of
heulandite, Clinoptilolite exhibited increased thermal stability due
to its higher content of silica. On this basis, it was recommended
that Clinoptilolite be retained as a valid mineral species. Mumpton
defined Clinoptilolite as a "mineral of the zeolite group having a
molecular composition close to
(Na20)0-70(CaO)0>10(K20)0-15(MgO)0. 05-Al203-8.5-10.5 Si02-6-7 H20
and a structure similar to heulandite, but which can clearly be dis-
tinguished from heulandite by optical, x-ray, thermal, and chemical
means." This composition is essentially the same as ones given by
Berman [47], Pabst [48], and Barrer et al_. [50].
The largest known deposite of clinootilolite in the United States is
found in southern California within a deposit of bentonite called
hectorite because of its proximity to Hector, California [51]. Although
Ames stated that Clinoptilolite is a common material found in bentonite
deposits in the western United States, no information was given concern-
ing the quantities of Clinoptilolite present in any deposit.
30
-------
Clinoptilelite from the Hector deposit has been found to contain from
5.to 15% impurities consisting of quartz, feldspar, and unaltered glass
with traces of calcite and montmorillonitic clays [42,50]. Ames [52]
also used clinoptilolite from the John'Day formation in Oregon which
he reported to have a purity of 95% or greater. Sheppard [53] reported
the occurrence of clinoptilolite in sedimentary deposits in 66 locations
in the United States. However, no information was given concerning the
purity or abundance of clinoptilolite in these deposits. The presence
of clinoptilolite in Jaoan has been reported by lijima [54] and by
Minato and Utada [55]. Minato and Utada reached many of the same con-
clusions concerning the similarities between clinoptilolite and heulan-
dite as did Mumpton [49]. The primary difference between these deposits
of clinoptilolite and those found in the United States was that the
clinoptilolite of Japan was reported to be found in potassium, calcium,
and sodium forms, whereas that found in the United States has been
reported to be predominantly in the sodium form.
Capacity. Although the total ion exchange capacity of a material is by
no means a complete description of its ion exchange properties, it is an
indication of the applicability of the substance for process use. For
example, the mineral glauconite (New Jersey greensand), which was widely
used in water softening before the development of organic exchangers,
has a total exchange capacity of 0.17 meq/g [56]. In comparison, the
exchange capacities of strong acid cation exchangers are usually 4 to
5 meq/g. In such, a case it is doubtful that any differences in selec-
tivity between these exchangers could be great enough to warrant use
of the greensand.
Values of the exchange capacity of clinoptilolite as determined by
various investigators are listed in Table 1. Because the exchange
capacity is dependent on the method used as well as the ions involved,
the conditions under which these capacities were determined are also
given. The total exchange capacity of clinoptilolite measured by these
investigators ranges from 1.6 to 2.0 meq/g and is slightly lower than
the average for zeolites. Several observations may be made concerning
these values. For the ions used, there is no apparent difference in
the accessibility of the ions to the exchange sites. In comparing the
experimentally determined exchange capacity to the composition of
clinoptilolite, Barrer et al. [50] found that their exchange capacity
corresponded to 98% of th~e possible capacity. Thus, it appears that
practically all of the exchange sites in clinoptilolite are accessible
to alkali metal ions. It also appears that the acid wash employed by
Ames does not affect the exchange capacity. In fact, the acid wash
would be expected to increase the exchange capacity both by opening
pores blinded by acid soluble impurities and by removing impurities
which contribute to the weight of the exchanger but not to the exchange
capacity. However, the values reported by Ames are generally lower
than other reported values.
31
-------
TABLE 1
ION EXCHANGE CAPACITIES OF CLINOPTILOLITE
Reference
Ames [52]
Ames [57]
Ames [12]
Howery and
Thomas [58]
Barrer,
Papadopoulos,
and Rees [50]
Frysinger [59]
Capacity
meq/g
1.7
2.0a
1.7
1.7
1.81
2.05
2.04
1.97
1.83
1.61
Ion Used
Na+ replaced by
Cs+; Cs+ re-
placed by Na+
Li+ replaced by
Cs+
Ba++ replaced
by Cs+
NHj
Na+
Cs+
NHj
Na+ replaced by
NH+
Cs+
Size Material
Not specified
0.25-0.50 mm
(35x60 mesh)
30x50 mesh
30x80 mesh
18x30 mesh
20x40 mesh
Preparation of
Material
Washed in 10% HN03;
put in proper form by
contact with satu-
rated solution of the
appropriate chloride
salt for two days
Washed in 10% HC1 ;
put in proper form by
contact with satu-
rated solution of the
appropriate chloride
salt for 21 hours;
washed in distilled
water
Not specified
Contacted with NaCl ;
rinsed with distilled
water
Contacted with NaCl;
rinsed with distilled
water
Not specified
Method
Double tracing tech-
nique; substantiated
by titration of H+
based sample
Shallow bed method
using radioactive
tracers
Small column method;
direct determination
of NH£ by distillation
of clinoptilolite in
kjeldahl flask
Determination of Na+
and Cs+ radiochemical ly;
determination of NHj by
difference from an Na+
based clinoptilolite
Na+ measured radio-
chemically; NHt deter-
mined by analyses of
nitrogen in the NH1;
exchanged sample
Column method using
radioactive Cs+ as
tracer
aClinoDtilolite from John Day Formation, Oregon. Other samples from Hector, Calif, except for that
used by Frysinger who did not specify the source.
u>
N)
-------
Selectivity. In order to identify the preference of clinoptilelite
for various ions, Ames [60] determined the cesium capacity of clinopti-
lolite using 0.01 N cesium solutions in the presence of 1 N concentra-
tions of the competing cations. Capacities were calculated from the
point (C/CQ)CS =0.5 using breakthrough curves obtained from laboratory
columns. The results showed that cesium was most preferred by the
zeolite and that the more closely a competing cation approached cesium
in size, the more selective clinoptilolite became for that ion. Three
replacement series were identified for univalent, bivalent, and tri-
valent metal ions. The cesium capacity in the presence of the ammonium
ion was intermediate between the series for univalent and bivalent ions,
indicating the influence of the electronic structure of the ion in
determining selectivity. The order of preference for the various ions
decreased in the order,
Cs+ > Rb+ > K+ > NHt > Ba++ > Sr++ > Na+ > Ca++ > Fe+3 > A1+3 > Mg > Li.
In later work Ames.[12] determined the equilibrium isotherms for ammonia
and other cations which are present as macrocomponents in wastewaters.
These isotherms, which are reproduced in Figure 3, illustrate that
clinoptilolite is selective for ammonia relative to all of these ions
except potassium. Throughout this report solid phase concentrations,
q, selectivity coefficients, K, etc. for ammonia are designated NH^-N
because the ammonium ion is the species exchanged. However, solutions
phase ammonia concentrations are more correctly designated as C|\|H3_N
as both unionized ammonia and the ammonium ion are measured in ammonia
determinations. At low pH there is no significant difference in total
NH3-N and the ammonium ion concentration. Additional nomenclature for
high pH solutions is given in Chapter Vi In all cases these species
are designated as .nitrogen in accordance with conventional water and
wastewater treatment practice. The separation factors, defined by
Equation (5), were calculated from these isotherms and are shown in
Table 2. Because they were not constant over the entire concentration
range, the values shown in Table 2 were calculated for the point
xNHs-N =0.5 for comparative purposes.
TABLE 2
SEPARATION FACTORS FOR AMMONIA EXCHANGE ON
CLINOPTILOLITE AT XNH _N = 0.5a
System
NH4-Mg
NH4-Ca
NH^-Na
NH^-K
NHi -N
Separation Factor, a., z
at xNHs_N =0.5
17.2
6.7
4.6
0.39
*Data from Ames [12]
33
-------
0.0
Total Solution
Concentration 0.1 N;
Temperature 23° C
I
I
I
I
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Equivalent fraction of NH3-N in Solution Phase, XNH N
FIGURE 3- ISOTHERMS FOR EXCHANGE OF NH^ FOR K+, Na+,
Ca++, AND Mg++ ON CLINOPTILOLITE [12]
34
-------
In strong acid organic resins, the effects of electroselectivity [15]
and swelling pressure are the primary factors influencing resin selec-
tivity. In dilute solutions under normal operating conditions this
results in ions of higher valence and smaller hydrated radius being
preferentially exchanged into the resin [44]. The selectivity sequence
for organic resins generally follows the lyotropic or Hofmeister series
and for sulfonated divinyl benzene resins has the order,
Ba++ > Sr++ > Ca++ > Cs+ > Rb+ > Mg++ - K+ > NH£ > Na+ > H+ > Li+ [61].
Although selectivity for ions of equal charge follows the lyotropic
series in both clinoptilolite and strong acid exchangers, the selec-
tivity among ions of different valence is quite different.
On the basis of the cation replacement series established for clinopti-
lolite, Ames [60] concluded that the preference for a. given cation by
clinoptilolite was dependent on the relationships between cation size,
cation charge, and electronic structure of the cation. Temperature in
the range 20°-60°C affected selectivity significantly only for Na-Cs
exchange. Because clinoptilolite is one of the open zeolites, the
effect of cation size is that of partial ion sieving and not complete
ion exclusion. In order ,to determine the effect of hydrated water on
clinoptilolite selectivity, Ames [62] conducted exchange studies with
clinoptilolite and other open' zeolites in molten salt solutions of
lithium and potassium nitrates. These studies showed that the struc-
tural water of clinoptilolite was not as firmly bound to the zeolite
framework as it is to other, less selective, open zeolites. Thus, the
structural water .of clinoptilolite is free to exert a more intense
sieve effect on cations entering the zeolite cavities. In the case of
a high field strength cation such as lithium, structural water is
attracted to the cation, preventing it from approaching the exchange
site as closely as in open zeolites having more tightly bound water.
Low field strength cations suclras ammonium do not attract as much
water of hydration and, therefore, are freer to migrate through the
lattice and closely approach exchange sites.
Steric and ion,sieve effects in clinoptilolite have been discussed by
Barrer £t a^. 1.50]. The degree of exchange of various alkylammonium
ions in clinoptilolite was explained in terms of the free dimensions of
channel entrances. Because dimensions of the clinoptilolite framework
have not been determined, Barrer assumed that the crystal structure of
clinoptilolite was identical to that of heulandite. While the similar-
ities in composition between heulandite and clinoptilolite support
this assumption, differences in exchange properties of the two zeolites
raise doubts as to the completely iso-structural nature of the two
materials [60], Nevertheless, the correlation between the degree of
exchange of organic ions and the size of the heulandite windows seems
to indicate a reasonable similarity between the two structures. Barrer
used crystallographic data of Merkle and Slaughter [63] showing ellip-
tical windows of two different sizes. Free dimensions of the larger
window were approximately 3.5 by 7.9 A and of the smaller window, 3.0
by 4.4 A. The existence of windows of two sizes in the clinoptilolite
35
-------
framework could explain the sigmoid isotherms for Na-Sr, Na-Ca, and
Ca-Sr exchange reported by Ames.[64] and the variation of separation
factors calculated from isotherms shown in Figure 3.
The partial sieve action in clinoptilolite may be illustrated by
comparing ionic diameters in Table 3 to the free diameters of the
crystal windows. It can be seen that there is ample room for all
of
TABLE 3
SIZE OF SELECTED CATIONS
Ion
Li
Na
K
NHh
Rb
Cs
Mg
Ca
Sr
Ba
Crystal
Diameter9
A
1.20
1.90
2.66
2.80
2.96
3.38
1.30
1.98
2.26
2.70
Hydrated
Diameter^
A
20.0
15.8
10.6
10.7
10.2
10.1
21.6
19.2
19.2
17.6
Crystal diameters from Pauling [65].
Hydrated diameters from Jenny [66]
as found in Grim [67].
these ions to enter the cavities based on the crystal diameter of the
ions, but that none of the ions can enter without some of the water of
hydration being stripped. Based on crystal diameters, sufficient room
exists for any two ions to pass in the larger channel. However, several
ion pairs could not pass in the smaller window.
In an effort to explain the general preference of clinoptilolite for
univalent ions, Ames [64,68] examined the selectivity of zeolites having
different Al/Si ratios. Using the A!/Si ratio as a general predictor
of the distance between exchange sites in the zeolites, Ames found that
bivalent ion selectivity decreased with decreasing Al/Si ratios. This
is explained by the fact that bivalent ions cannot remain stable and
occupy two exchange sites which are relatively far apart. However, the
inadequacy of this phenomena to completely explain selectivity between
ions of different valence was shown by the fact that clinoptilolite is
somewhat more strontium selective than other zeolites having higher
Al/Si ratios [68]. Howery and Thomas [58] attempted to account for
36
-------
nonideality of exchange on clinoptilolite by a thermodynamic treatment
of the exchange sites. Two suppositions were made concerning the nature
of the exchange sites: 1) that the sites were identical and differences
in binding energies were influenced only by the population of neighboring
sites, or 2) that binding energies varied from site to site but that
sites were so widely separated that neighboring ions exerted no influence
on each other. However, from results obtained it was impossible to
attribute exchange behavior to one of the models.
Diffusivity. Ames [42] has reported that diffusion in clinoptilolite,
which is composed of crystallites imbedded in a binder, is much more
rapid than in the coarsely crystalline zeolites. Using an expression
proposed by Boyd, Ames [62] found the particle phase diffusion coeffi-
cient to be 1.66 x 10~7 cm2/sec for Na-Cs exchange at 25°C in 0.134-mm
diameter particles. In another study the particle phase diffusivity
was found to be 3.88 x 10~7 cm2/sec at 29°C for 0.25 to 0.50-mm dia-
meter particles [69]. These values are an order of magnitude smaller
than those measured by Boyd et^ ail_. {23] of 2.0 x 10"6 and 2.9 x 10~6
cm2/sec for Cs-K and Na-NH^ exchange, respectively, in 0.23-mm particles
of Amber!ite IR-1. In investigating film diffusion, it was not possible
to compute the film diffusion coefficient. However, values obtained by
Ames [69] for a parameter containing the film diffusity were very close
to values obtained by Boyd for Amber!ite IR-1 when corrected to similar
conditions. Boyd estimated film diffusion coefficients to be about
1.8 x 10~5 cm2/sec. Thus, while film diffusion coefficients are
approximately the same for clinoptilolite and organic resins, particle
phase diffus.ivities are an order of magnitude less for clinoptilolite.
Barrer [50] measured diffusion coefficients for various alky!ammonium
ions in clinoptilolite and concluded that intraparticle diffusion in
pore spaces between crystallites was the rate controlling step. Values
of the diffusion coefficient ranged from 10~8 to 10~9 cm2/sec. While
some pore diffusion is not usually affected by the nature of the
exchanging ions, differences in hydration characteristics of organic
and inorganic cations might be expected to account for differences in
these values and those measured by Ames.
In relating diffusion in clinoptilolite to selectivity, Ames [62]
concluded that the different selectivities of erionite and clinopti-
lolite were not caused by differences in diffusion rates, but rather
that cation selectivity differences influence exchange kinetics. From
studies of the exchange rate of ions under conditions where film
diffusion limited the rate of exchange, Ames found that exchange rates
did not correlate to liquid diffusion coefficients, but that interaction
of the ions with the exchange sites also had to be taken into account.
In another study Ames [70] showed that self-diffusion of ions, i.e.,
diffusion of one ion in an exchanger in the absence of a concentration
gradient, decreased as the ion valence increased.
Stability. The instability of natural clays and zeolites toward acids
and alkalis has been known since these materials were widely used in
37
-------
water softening [71]. Studies of the effect of pH on the clinoptilolite
structure have been limited, but Ames [60] found that the material was
considerably more acid resistant than other zeolites. However, Barrer
and Makki [72] found that treatment of clinoptilolite with acid in 1 N
and higher concentrations resulted in progressive displacement of
aluminum from the crystal framework leaving only the hydroxylated silica
after treatment with 5 N acid.
The stability of clinoptilolite in the presence of alkali has been
investigated by Barrer et_ aj_. [50]. Samples of clinoptilolite were
shaken with varying concentrations of NaOH for two days at room
temperature. Weight loss during this period was taken as a measure of
the degree of chemical attack by the alkali. The relation between
weight loss and concentration of NaOH was reported to be nearly linear
with a 70% loss of weight occurring from exposure to 20% NaOH. Thes-e
results indicate that regeneration of clinoptiloli te with caustic
solutions might result in loss of the zeolite from chemical attack.
38
-------
V. NATURE AND SCOPE OF THE INVESTIGATION
Based on the previous work concerned with ammonia removal and the ion
exchange properties of clinoptilelite reviewed in Chapter IV, the specific
nature and scope of this study may be outlined. It is the purpose of
this chapter to assess the potential utility of clinoptilolite for
ammonia removal, to outline the rationale of the study, and to define
the specific scope of the investigation
POTENTIAL ADVANTAGES OF CLINOPTILOLITE
As an ion exchange process, the* ase of clinoptilolite has several advan-
tages over other methods of ammonia removal. The widespread application
of ion exchange processes has led to the development of equipment which
functions almost automatically ahd requires relatively little maintenance.
Ion exchange processes are by nature stable and predictable with respect
to product quality. While the total exchange capacity of clinoptilolite
is somewhat less than that of synthetic organic resins, its selectivity
for the ammonium ion compensates for this fact, giving it an advantage
over conventional ion exchange resins. Exchange kinetics in clinopti-
lolite, though poorer than in organic resins, are rapid enough to permit
favorable column utilization during exhaustion. In areas where disposal
of regenerant brine is a problem, the high pH solution used to regenerate
the zeolite can be stripped of ammonia and reused.
In the final analysis the applicability of clinoptilolite for ammonia
removal is heavily dependent on process costs. While cost estimates
prepared for nitrogen removal processes are based on preliminary operating
and design information, they do provide a guide for assessing the poten-
tial for the use of clinoptilolite for ammonia removal. The biological
nitrification-denitrification process was discussed by McCarty [9] who
estimated that costs would be approximately $0.08/1000 gal. Denitrifi-
cation in this process was accomplished using anaerobic filters with
methanol added as a carbon source. Costs for the three-stage biological
process investigated by Barth et.ll. [11] were not estimated, but it
seems likely that nitrogen removal costs for this process would not
differ significantly from the cost of the system proposed by McCarty.
'* ^'
The cost of ammonia removal by chlorination is highly dependent on
transportation costs' unless on-site production vf, chlorine is feasible.
Given a chlorine cost of $0.04/lb and the assumption that 10 parts of
chlorine will completely oxidize 1 part of ammonia, chemical costs for
ammonia removal by chlorination will be $0.075/1000 gal for an initial
ammonia concentration of 20 mg/a NH3-N. This cost considers only the
39
-------
cost of chlorine and does not include capital costs, maintenance costs,
or the cost of dechlorination if required. Ammonia removal by super-
chlorination might be undesirable for water reuse applications where
toxicity to aquatic life and tastes and odors are objectionable. Gulp
[73] has estimated that ammonia removal by air stripping using a low
headless, cross-flow tower costs $0.017/1000 gal. Costs estimated by
Smith and McMichael [74] for the same type of tower range from $0.036/
1000 gal for a 10-mgd plant to $0.026/1000 gal for a 300-mgd plant.
Parrel 1 [75] has quoted a cost of $0.029/1000 gal based on experience
at the South Tahoe Public Utility District. These estimates do not
include the cost of raising the pH of the waste prior to ammonia strip-
ping. The estimated cost of ammonia removal using clinoptilolite was
$0.16/1000 gal for a proposed installation at the South Tahoe Public
Utility District [76]. However, Dean [77] reported a cost of between
$0.03/1000 gal and $0.06/1000 gal based on operating experience of the
Battelle demonstration plant and suggested that a cost of about $0.10/
1000 gal sounded most reasonable. Cost estimates for the construction
of advanced waste treatment facilities at the Blue Plains wastewater
treatment plant in Washington, D. C. prepared by the Bechtel Corp. [78]
revealed that ammonia removal by ion exchange would cost $0.103/1000 gal
for a 240-mgd plant and $0.097/1000 gal if the plant were expanded to
300 mgd'. Estimates for nitrification-denitrification ranged from
$0.118/1000 gal to $0.124/1000 gal for a 240-mgd plant and $0.112/1000
gal to $0.117/1000 gal for expansion to 300 mgd, the range of costs for
each size design depending on whether or not alum was added to the
nitrification stage.
This comparison shows that air stripping is the least costly method with
present technology. However, cost estimates for air stripping might not
accurately reflect cost of tower replacement due to delignification and
maintenance costs required to remove scale from the tower packing. If
these problems do not prove to be serious air stripping will probably
be the preferred method in areas where low temperatures are not a
problem and where ammonia discharge to the atmosphere is not prohibited.
Although costs of other methods listed range from $0.075 to $0.10/1000
gal, the preliminary nature of these estimates suggests that the cost
of any of these processes is approximately $0.10/1000 gal. On this
basis the use of clinoptilolite for ammonia removal is, with the excep-
tion of air stripping, competitive with other methods.
SPECIFIC AREAS OF STUDY
Effect of Water Composition on Ammonia Exchange Capacity
The ammonia exchange capacity will vary in proportion to the NH3-N/total
cation ratio of the column influent. Although the equilibrium column
loading may be ideally predicted knowing the influent composition and
selectivity coefficients as described by Vermeulen, Klein, and Hiester
40
-------
[20] and as shown by Mercer [14] for small columns, solid phase activity
corrections must be made for accurate application of this method to
waters of widely varied composition. However, the equilibrium proper-
ties of clinbptiTolite indicate that the ammonia exchange capacity is
a function of cationic strength. The term cationic strength refers to
the ionic strength of the water considering only the cation species
present. Cationic strength will be designated I+ and is calculated
using the expression for ionic strength [16]:
1+ = }£>! z?) (10)
where m-j is the cation concentration of the ith species and Zj is the
valence of the particular cation. Ammonia selectivity coefficients
determined by Ames [12] show that zeolite selectivity for ammonia
increases as the competing cation/ammonia ratio increases. In addition,
Equation 4 shows that the selectivity of the exchanger for multivalent
cations decreases as the total solution concentration increases. While
the ammonia exchange capacity will decrease with increasing cationic
strength, these considerations indicate that the decrease will become
less with increasing cationic strengths. To determine the nature of
the variation of ammonia exchange capacity with water composition, runs
were made saturating clinoptilelite with waters having cation composi-
tions representative of chemically different sewages.
Effect of pH on Ammonia Exchange
In removing ammonia by ion exchange, the pH of the waste is of primary
importance. The pH of the waste prior to ion exchange may range from 5
to 11, or higher, depending on the chemical used in precipitation
processes preceding the ion exchange unit. The pH will also play an
important role in regenerating and rinsing of ion exchange beds. The
effect on the exchange of ammonia will be twofold. At high pH, ammonia
will be predominantly in the unionized form, unavailable for exchange.
At lower pH levels, hydrogen ions in the water compete with ammonium
ions for exchange sites.
The influence of pH may be examined using equilibrium ion exchange
relationships. Although this is done most correctly using a binary
NH^-H system, the result does not correctly predict behavior in a multi-
component system. In the binary case, concentrations of both ammonium
and hydrogen ions decrease as the pH increases. However, the equili-
brium relationships are such that the clinoptilolite remains predomi-
nantly in the NH4 form even at high pH values. For the case of the
multicomponent wastewater system, other ions, unaffected by changes in
pH, will replace ammonia in the exchanger at high pH. To represent the
multicomponent wastewater system, a ternary NH^-Na-H system will be
considered. Although this can correctly be described only by a ternary
diagram or by appropriate activity corrections, the approximate solution
41
-------
is sufficient to describe a system having a relatively constant equili-
brium sodium concentration.
Three equilibria are involved:
NhJ + NaZ ^ NH^Z + Na+
NHj + HZ + NHUZ + H+
NH3 + H+ $ NHt
These equilibria may be described by the NH^-Na and NH4-H selectivity
coefficients and the ammonia equilibrium constant:
q'Na
N
Because it is necessary to distinguish between ionized and unionized
ammonia, CN^.N designates the ammonium ion concentration as nitrogen,
while CNH^-N refers only to the unionized ammonia concentration, again
as nitrogen. The total ammonia nitrogen concentration, analogous to the
concentration measured in a kjeldahl distillation, is designated c^ _^
such that, 3
In order for equilibrium concentrations to be compatible with the units
of Keq, they are expressed in moles/£ as indicated by the brackets. For
the symmetric univalent system, selectivity coefficients, normally
expressed using concentrations in milliequi valents per liter, may be
used without correction for the difference in units. Solid phase concen-
trations may still be expressed as meq/g.
In the solid chase the sum of individual concentrations will equal the
total exchange capacity, Q.
42
-------
Equations 11 and 12 may be solved for
r~ j_ "i r~
Equation 15.
_N and substituted into
li
[cNa] qNHlt-N [CHJ
I
[_CNIVNJ
The concentration of CM is a function of pH, and may be expressed
[Ch]
of CMH
in terms of Equations 13 and 14,
KIH,-N =T
L k J [CH] + Keq
Substituting Equation 17 into Equation 16, the result is
(17)
,
[c
LC
NH-N
1 Tc 1
J LCHJ
Na
Tc 1
LCJ
..
which expresses the solid phase concentration of ammonia as a function
of the total equilibrium ammonia concentration in the solution, the
equilibrium pH, and the equilibrium sodium concentration.
The constants necessary for the use of this relationship are Keq, K|\ja4~ ,
KNHt-N5 and Q. The ammonia equilibrium constant is 5.55 x 10~^0 at 25°C
[T6]. The value of K^~H is 4.6 from Table 2. The value of KNHl*-N
has not been determined directly, but may be calculated from otner data.
From the Cs-NH4 isotherm given by Mercer and Ames [79], K^|4_|^ = 5.
From data given by Ames [42], Ktjs= 10. Because of the variation of
equilibrium data for clinoptilolite with concentration, average values
were chosen for these coefficients. The ammonium-hydrogen selectivity
coefficient may be estimated from these data:
10
Equation 18 then becomes,
43
-------
qNH1+-N = [CN ] [cj + 5.55 x lO'10 LcHJ + 5.55 x KT10 '
1 + _!ia - D - + d - (18)
4.6 [c
NH,-NJ [
-------
be expected in cyclic operation, the dependence of column performance
on flow rate and the nature of the previous regeneration, and the capa-
city :of clinoptilolite using different sewages'for the column influent.
Questions related to fouling and the sustained capacity were also of
interest.
Process Costs
To establish the economic feasibility of the clinoptilolite ion exchange
process, a detailed cost estimate was prepared for the optimum operating
conditions determined during the experimental phase of the study.
45
-------
VI. EXPERIMENTAL EQUIPMENT AND PROCEDURES
EXPERIMENTAL COLUMN UNITS
The experimental phases of the study involved runs using both synthetic
waters and treated sewage from the SERL pilot plant, the East Bay
Municipal Utility District (EBMUD), and the Central Contra Costa Sanitary
District (CCCSD). Two exchange column units were used in these tests.
The SERL unit was designed primarily to be used for runs in which only
single column operation was required. The column unit used in studies
at EBMUD and CCCSD was designed for easy transport and for flexibility
of operation when several columns were employed in series. Both units
were constructed so that modifications could easily be made for upflow
or downflow operation, backwashing, and regeneration.
SERL Column Unit
Columns. The column unit used in runs made at SERL is pictured in
Figure 4. Basically the unit consists of two 4-in. ID plexiglass
columns 6 ft in length. Construction details of the columns are shown
in Figure 1, Appendix A. The exchange medium was supported on a plexi-
glass plate containing 5/16-in. holes for flow distribution and a 210 y
mesh nylon screen. Piping to the column consisted of 1/2-in. PVC pipe
and fittings and was constructed so that each column could be removed
from the unit by uncoupling unions located near the ends of the columns.
Valves which came into contact with regenerant solutions were PVC ball
type while other valves were gate type of brass. A schematic diagram
of the column system is shown in Figure 6. Columns were supported on
a welded frame made from 1-1/2 in. x 1-1/2 in. x 1/4 in. steel angle.
Influent to the columns was pumped to a constant head tank located on
top of the frame. Flow was regulated by a rate controller consisting
of float valve which discharged into a small reservoir. Changes in flow
rate were made by adjusting a 1/2-in. gate valve located on the float
chamber. This simple and inexpensive arrangement provided excellent
control over flow rates and made it possible to maintain flows with
less than five percent deviation from the desired value. Flow could be
checked by rotometers located in the piping between the constant head
tank and the column inlet.
Regeneration was accomplished by pumping regenerant through a Vanton
Flexi-Liner pump (Vanton Pump Co., Hillside, N.J.) to a constant head
tank from which it flowed to the columns. Regenerant was prepared in
a 200-gal Nalgene tank. Following the addition of caustic and mixing,
47
-------
oo
FIGURE 4. COLUMN UNIT LOCATED AT THE
SERL TREATMENT FACILITY
FIGURE 5. COLUMN UNIT USED IN STUDIES
AT EBMUD AND CCCSD
-------
Regenerant
Constant Head Tank
Influent
FTexi-Liner
Pump \
i
Influent Constant
Head Tank
Variable
'Speed Pump
Regenerant
Storage
Drain
Unions
PVG Ball
Valves
PVC Ball
Valves
^E
t
t
fl
3
1
V"
^' X
S~. J
/
!
iL
4" ID x 6'
s Plexiglass
' Column
Check Valve
\
>
£^ s
j, | Float Va
Drain
Influent
Rotometer
Backwash
Rotometer
Hose from
Water Tap
Drain
FIGURE 6. SCHEMATIC ILLUSTRATION OF COLUMN UNIT
49
-------
the solution was allowed to settle overnight. Subsequently, supernatant
was siphoned from the tank to a 200-£ plastic container using a floating
intake.
Headless measurements were made by means of piezometer taps located at
6-in. intervals in the side of the columns. Plastic tubing from the
taps ran to a board attached to the side of the column frame which was
marked at 0.02 ft intervals.
Sampling. Composite effluent samples were collected over 2- to 6-hr
periods, depending on the nature of the test being run using a Misco
fraction collector, Model 6510, (Microchemical Specialties Company,
Berkeley, California). The sample was pumped through plastic tubing
using a Sigmamotor Model T-6 pump (Sigmamotor, Inc., Middleport, N. Y.)
to a glass tube fastened to the edge of the rotating collector. Funnels
placed on the table directly beneath the collector caught the sample
and transmitted it to bottles placed beneath the table. Composite
influent samples were taken from the head tank over longer time inter-
vals using the same Sigma pump.
Column Units Used in EBMUD and CCCSD Studies
The columns used in studies at EBMUD and CCCSD are shown in Figure 5.
This unit included four 4-in. ID by 6-ft columns used for ion exchange
and two 6-in. ID by 6-ft columns used as filters. Details of the 4-in.
columns, which were essentially the same as the 6-in. columns, are
shown in Figure 2, Appendix A. A bed of river gravel was included at
the bottom of these columns to obtain better flow distribution during
backwashing. However, some channeling was still evident when columns
were backwashed.
Influent was pumped throuah the columns using a three-stage Moyno pump
(Robbins and Myers, Inc., Springfield, Ohio) run by a 1/2-hp DC motor.
A pressure relief valve on the outlet side of the pump protected the
columns from excessive pressure. Connections between columns were made
using neoprene hose and tubing with PVC unions to facilitate changes
from one operating mode to another. A rotometer mounted on the column
frame was used to measure both influent and backwash flow rates.
To achieve more effective backwashing, a multiple jet surface wash
device was incorporated into the column design. This feature, illus-
trated in Figure 3 in Appendix A, proved to be very effective in
breaking up the cake which frequently formed at the surface of the
first column.
The column framework was constructed from Super Strut channel (Imperial
Strut and Hanger Co., Oakland, California) and was designed for
disassembly and transport. A walkway located behind the columns
provided easy access to the top of the columns as well as stability for
50
-------
the column supports. Regeneration was accomplished in a manner similar
to that for the SERL unit. Regeneration flow rates were regulated using
a float valve controller. Sampling procedures were identical to those
used with the SERL unit.
The filter used ahead of the ion exchange unit consisted of 2 ft of
sand having an effective size of 0.27 mm and a uniformity coefficient
of 1.63 and 1 ft of anthracite reported by the distributor to have a
size range of 0.85-0.95 mm. While this filter performed well during
the study, more mixing between layers occurred than was desirable.
Better headless distribution would have resulted had a larger sand
been used.
ANALYTICAL PROCEDURES
Sampling Procedure
Wastewater samples taken to monitor performance of treatment processes
upstream from column units were collected continuously over 24-hr
periods and were refrigerated prior to analysis. While it was not
possible to refrigerate column effluent samples as they were taken,
samples were refrigerated after they had been collected prior to analysis.
Composite samples were collected in the morning and were usually analyzed
the same day except on weekends. Priority was given to ammonia, COD,
and BOD determinations.
Analytical Methods
Chemical and physical analyses were generally performed in accordance
with Standard Methods [80] or the FWPCA Methods for Chemical Analysis
[81]. However, in several instances the detailed procedure followed
was that given in SERL Analytical Methods [82].
Ammonia Nitrogen. Samples were made alkaline to the phenolphthalein
endpoint and distilled in a kjeldahl apparatus. Samples containing
more than 3 mg/£ NH3-N were titrated with 0.01 N h^SO^ using an alpha-
zurine-methyl red indicator [82]; those containing less than 3 mg/a
NH3-N were nesslerized using a Bausch and Lomb Spectronic 20.
Total Hardness, Calcium, Magnesium. The EDTA titrimetric method was
employed using Calmagite indicator for total hardness determinations
and hydroxy napthol blue indicator (both products of Mallinckrodt
Chemical Works) for calcium determination [82]. Magnesium was deter-
mined by difference from total hardness and calcium values.
Sodium, Potassium. Sodium was determined by flame spectrophotometry
using a Beckman DU spectrophotometer [82]. Potassium was originally
determined by flame spectrophotometry. However, most determinations
51
-------
were made by atomic absorption soectrophotometry using a Perkins-
Elmer model 290 B atomic absorption unit because of the increased
stability of this method.
Alkalinity, pH. The pH was read using a Beckman model 76 pH meter
equipped with a Corning Triple-Purpose electrode designed to measure
pH in the range 0 to 14. Alkalinity was determined by potentiometric
titration to pH 4.3.
Organic Nitrogen. Organic nitrogen was determined according to FWPCA
Methods.
Chemical Oxygen Demand. A modification of the normal COD procedure was
used which permitted more accurate determination of samples having low
COD values. The only difference in this procedure and the normal method
was that a smaller refluxing unit and less concentrated potassium
dichromate and ferrous ammonium sulfate were used [82].
Biochemical Oxygen Demand. BOD values were determined according to
Standard Methods except that sample dilutions were made directly in
the BOD bottle.
Suspended Solids. Suspended solids were determined by filtration of
samples through Whatman GF/C glass fiber filters. Filters were pre-
washed and baked in a 560°C furnace prior to use. Samples were dried
at 105°C to constant weight, then stored in individual desiccator jars
prior to weighing.
Total Solids. Total solids were determined according to FWPCA Methods.
Turbidity. A Rossum model 600 turbidimeter (Rossum Instrument Co.,
Cupertino, California), previously calibrated with a standard silica
solution, was used for measurement of turbidity.
Total and Total Soluble Phosphorus. Phosphorus determinations were
made by digesting samples to convert all forms of phosphorus to ortho-
phosphate followed by measurement of orthophosphate using the ANS
reduction method [80,82],
PREPARATION OF THE CLINOPTILOLITE
Clinoptilolite was obtained from the Baroid Division, National Lead
Company, Houston, Texas, which mines clinoptilolite from the Hector,
California deposit. Originally, only mine run material consisting of
1- to 2-in. chunks was available. A later shipment consisted of minus
4 mesh material. The only guide in crushing the zeolite came from
crushing tests reported by Berry [83] who used a Hazemag impact crusher.
A recovery of 52% was reported for 16 x 70 mesh material (U. S. Standard
Sieve).
52
-------
Because of limited facilities, crushing was accomplished in batches.
Clinoptilolite received Tn chunk form was run through a jaw crusher set
to crush the material into 1/2- to 1-in. pieces. Subsequently, all
material was fed through a cone crusher set to deliver the desired
20 x 50 mesh size. Following a rough separation, +20 mesh material
was again passed through the cone crusher set with a slightly smaller
opening. Finally, all material was screened through 20 x 50 mesh
screens to obtain the proper gradation. A recovery of approximately
50% of the original shipment was realized. As the material was used,
it was put into columns, soaked to remove air bubbles from the pores,
and backwashed to remove dust from the bed.
A comparison between the grain size distributions for this material and
clinoptilolite obtained from the Battelle Northwest demonstration plant
[14] is shown in Figure 7. Although no effort was made to mix different
size fractions to obtain a specific product, the size distributions are
very similar. The sample from Battelle had an effective size of 0.38 mm
and a uniformity coefficient of 1.63, while the sample crushed in this
study had an effective size of 0.34 mm and a uniformity coefficient of
1.59.
CONDUCT OF THE STUDY
Selection of Column Operating Conditions
Wherever possible, data reported by other investigators were used in
determining column operating conditions to be used in this study. Ames
[60] reported the effect of different particle sizes on cesium exchange.
The breakthrough volume to c/C0 = 0.05 for 18 x 60 mesh particles was
70% of that for 60 x 100 mesh particles, while leakage occurred
immediately using 10 x 18 mesh particles. As a balance between exchange
kinetics and headless, 20 x 50 mesh particles were used in this investi-
gation. This was also the size material selected by Mercer [13,14,37]
in the Battelle demonstration plant.
The optimum flow rate for ion exchange processing is usually the greatest
flow which does not result in a sacrifice of kinetics. In studies to
determine the effect of flow rate on exchange kinetics of 20 x 50 mesh
clinoptilolite, Mercer [13,14,37] found that the breakthrough curve
became much shallower when the flow was increased from 20 to 30 BV/hr.
In this investigation a flow of 15 BV/hr was most often used both
because it resulted in convenient run times to the ammonia breakthrough
and because headless was not excessive. The bed depth used in all
studies was 3 ft.
Columns were regenerated upflow. This was desirable both to remove
suspended solids from the bed collected during exhaustion and to
minimize the entrapment of precipitated Mg(OH)2 and Ca(OH)2 suspended
53
-------
1.0
0.9
0.8
0.7
0.6
0.5
0.4
I 0.3
O)
0.2
0.1
I I
Prepared for this study
I
_L
I
I
1
l
I
0.5 1
10 20
FIGURE 7.
30 40 50 60 70 80 90
Cumulative Weight Passing, %
95
98 99
GRAIN SIZE DISTRIBUTION OF CLINOPTILOLITE
99.9 99.99
-------
in the regenerant solution. In addition, countercurrent regeneration
should be more efficient than cocurrent regeneration because of the
position of ammonia in the upper part of the bed. In cocurrent
regeneration ammonia must be pushed through the entire column length
and this requires a larger regenerant volume.
The bulk density and specific gravity of clinoptilolite have been
measured and reported by Barrer e_t al. [50]. The specific gravity for
18 x 30 mesh material was 2.19 and We specific gravity of wet particles
was 1.61. Ames [60] measured a bulk density of 0.79 g/cu cm. The
clinoptilolite used in this study was found to have a specific gravity
of 2.38, a wet particle specific gravity of 1.59, and a bulk density
of 0.74 g/cu cm.
Runs Using Synthetic Systems
During the course of the study several runs were made using chemically
modified tap water for the column influent. The average composition of
tap water used for makeup and for column rinsing is shown in Table 4.
TABLE 4
AVERAGE COMPOSITION OF
SERL TAP WATER
Ion
Na
K
Ca
Mg
pH
Concentration
10
0.
26
4
8.
mg/n
9 mg/K,
mg/Ji
mg/£
2
Technical grade chemicals were used to prepare solutions of the desired
concentration. Chloride salts were used except for a few runs in which
MgS04 was used. Tests were run to determine the concentration of Na, K,
Ca, Mg, and NH3-N present as impurities in the chemicals. The maximum
concentration of a single impurity amounted to 0.2% by weight.
Initially feed water was prepared in 500-gal batches, then pumped to an
influent storage tank. In later runs feed water was prepared continu-
ously in a mixing tank from which the solution was pumped directly to
the column head tank. Tap water was fed to the mixing tank from a
constant head tank in order to eliminate flow fluctuations caused by
55
-------
water pressure. A stock chemical feed solution was prepared in a
separate container and pumped to the influent mixing tank through a
Sigmamotor pump. Overflow from the influent head tank was returned
to the mixing tank while overflow from the mixing tank was wasted.
For runs requiring adjustment of the pH of the column influent, HC1
or NaOH was added to maintain a constant pH using the pH analyzer
described in the next section. For influent pH less than 9.5, the
pH probe was immersed in the mixing tank. When a pH equal to or
greater than 9.5 was used, the column influent was mixed in the floccu-
lator described in the next section, then passed through a clarifier
to allow precipitated solids to settle from the solution.
SERL Operations
Wastewater experiments were conducted using both chemical and chemical-
biological process trains as shown in Figure 8. Wastewater from primary
or secondary sedimentation basins of the SERL waste treatment plant was
pumped to the chemical precipitation unit at a flow of 1 gpm by a
variable-speed pump powered by a 1/3-hp Bodine DC motor and Minarik
speed controller (Minarik Electric Co., Los Angeles, Calif.). Lime
addition to pH 9.5 or 11.0 was controlled by a pH analyzer preset to
the desired pH value. This control unit consisted of a Beckman model
900 C pH analyzer (Beckman Instruments, Inc., Fullerton, Calif.), a
Foxboro model 62H controller (Foxboro Instruments, Foxboro, Mass.), and
a Minarik model W-53AM speed controller coupled to a 1/8-hp Bodine DC
motor (Minarik Electric Co., Los Angeles, Calif.). The motor drove a
Sigmamotor model T-6 pump (Sigmamotor, Inc., Middleport, N. Y.) which
pumped the lime slurry to the influent waste stream. The coagulated
wastewater was flocculated in a 3-compartment reactor having an average
detention time (defined as reactor volume divided by flow rate) of
15 min/compartment and settled in a 30-in. diameter conical-bottom
clarifier at a loading of 300 gal/sq ft-day. Because of nitrification
during activated sludge treatment, NH4C1 was added to the secondary
effluent when the treatment scheme in Figure 8b was being used to
maintain an ammonia concentration of 15 mg/a NH3-N.
Recarbonation to a pH between 7 and 8 was accomplished in a 6-in. dia-
meter, 4-ft deep plexiglass cylinder through which C02 passed counter-
current to the liquid. Following recarbonation the wastewater was
stirred in a single reactor having a detention period of ten minutes,
then settled in a clarifier identical to the one described above. From
a 15-gal holding tank the waste was pumped to the column head tank.
EBMUD Operations
Studies at the East Bay Municipal Utility District were conducted using
the chemical treatment system illustrated in Figure 9. Primary effluent
was pumped to the treatment unit from a channel of the main treatment
56
-------
Primary
Sedimentation
Activated
Sludge
T
Lime
Flocculation Sedimen-
tation
Floccu-
lation
Recar-
bonation
Sedimen-
tation
Ammonia
Exchange
a. Primary Sedimentation Followed by Chemical Coagulation
Primary Activated
Sedimentation Sludge
Lime
Flocculation Sedimen- Recarbon- Sedimen-
tation ation Floccu- tation
lation
TT
Ammonia
Exchange
b. Activated Sludge Followed by Chemical Precipitation
FIGURE 8. TREATMENT SYSTEMS USED IN WASTEWATER STUDIES AT SERL
57
-------
Line
Primary
Effluent 1
C
D
Flocculation
Sedimentation Recarbon-
ation
T
Filtration Ammonia
Exchange
FIGURE 9. TREATMENT SYSTEM USED IN STUDIES AT EBMUD
1
Lime to
^
1
pH 10
1 r*
5-10.8
:\j
Primary
Sedimentation
Floccu-
lation
Sedimen-
tation
Recarbon-
ation
T
JL
T
I
Ammonia
Exchange
Filtration
FIGURE 10. TREATMENT SYSTEM USED IN STUDIES AT CCCSD
58
-------
plant at a flow of 20 gpm. Lime was added in a concentration determined
periodically from jar tests. Monitoring of pH during part of the test
period showed that the pH of the flocculated waste ranged from 9.5 to
12. The flocculated waste was settled in a rectangular clarifier having
a detention time of 80 min and a surface loading of 800 gal/sq ft-day.
The chemical effluent was recarbonated by bubbling C02 through a 5-in.
diameter, 4-ft plexiglass cylinder immersed in the effluent trough of
the sedimentation tank. Following recarbonation to a pH between 7 and
8 the waste was pumped to the column unit. During part of the testing
period the precipitated effluent was treated by sorption of organics
through 4.5 ft of activated carbon (Filtrasorb 300, Calgon Corp.,
Pittsburgh, Pa.) or 3.3 ft of a macroporous resin (Duolite ES-33,
Diamond Shamrock Chemical Co., Redwood City, Calif.) prior to ammonia
removal. This modification had no effect on clinoptilolite performance,
but the runs made using this treatment scheme are identified in Appendix
E, Table 3.
CCCSD Operations
The treatment system illustrated in Figure 10 was used in studies at
the Central Contra Costa Sanitary District. Partially settled sewage
was pumped to the treatment unit at a flow of 4 gpm from a point near
the influent end of the primary clarifier. The influent sewage
received approximately 15 min of settling prior to being pumped to the
chemical treatment unit. Lime addition to a pH of 10.510.8 was
controlled by a pH analyzer. The coagulated waste was flocculated in
a single-compartment flocculator having a detention time of 25 min.
The waste was settled in a clarifier having an overflow rate of 240
gal/sq ft-day and a detention time of 4 hr. The effluent was recar-
bonated directly in the effluent wet well as described in the previous
section and pumped to the column unit.
59
-------
VII. SATURATION PERFORMANCE OF CLINOPTILOLITE
A necessary precondition for the integration of ammonia exchange on
clinoptilolite into advanced wastewater treatment systems is the exis-
tence of data defining the optimum exhaustion and regeneration cycle.
The purpose of this chapter is to consider the effect of water composi-
tion on the effective ammonia exchange capacity, to describe the effect
of pH on the performance of clinoptilolite, and to establish general
guidelines for the use of clinoptilolite under various operating condi-
tions. Because the characteristics of ion exchange processes are most
readily observed in systems in which no concentration fluctuations occur,
the work presented in this chapter consists of tests using synthetic
systems composed of tap water supplemented with inorganic salts. The
composition of these systems was representative of inorganic ions found
in domestic wastewaters. In examining column performance, the funda-
mentals of ion exchange theory and the properties of zeolites discussed
in Chapter IV must be used to interpret data in a way which will be
meaningful to the more general application of this process. In these
respects this work is of a more fundamental nature than other sections
of this report. However, the primary objective of this work was to
provide answers to specific questions regarding the application of
clinoptilolite in real systems. The engineering implications of the
work are specifically discussed in a concluding section.
ION EXCHANGE CAPACITY
The ion exchange capacity of clinoptilolite was measured using a small-
column method. Fresh clinoptilolite was prepared by washing several
times in an erlenmeyer flask to remove fines, equilibrating with 1 M
NaCl for two days, and washing with distilled water until the water
showed no chlorides when tested with AgN03. Subsequently, the material
was air dried, then dried overnight in a 105°C oven. Clinoptilolite
samples weighing 10 or 20 g were placed in 50-nu burettes and saturated
with 1 M NH^Cl at a flow of approximately 50 nu/hr for two days. Excess
ammonia was washed from the column by rinsing with distilled water for
two days.
Three methods were used to determine the ammonia exchanged into the
zeolite:
1. Direct distillation of ammonia by placing the NH4-form clinoptilolite
in a kjeldahl flask.
2. Elution of ammonia from the clinoptilolite by passing 0.5 M NaCl
through the column at 50 nu/hr for two days and determining the
ammonia content of the eluant.
61
-------
3. Elution of ammonia using 0.25 M CaCl2 at a flow of 50 nu/hr for
two days and determining the ammonia content of the eluant.
Ammonia was most easily distilled from zeolite particles passing a 100
mesh sieve. Thus, all samples were crushed and sieved prior to ammonia
analyses. Ammonia from -100 mesh particles could be removed in a
distillation volume of about 200 ma. The contents of each column were
thoroughly mixed before taking a sample for crushing.
Experimentally determined exchange capacities are shown in Table 5.
These values lie in the same range as values reported by other investi-
gators (cf. Table 1). The exchange capacity of 1.95 meq/g determined
from the direct disination method is identical, within the limits of
experimental error, to the value of 1.97 meq/g determined by Howery and
Thomas [58]. The value of 1.88 meq/g determined by elution of ammonia
with NaCl is very close to the value of 1.85 meq/g reported by Barrer
[50] who used the same method.
TABLE 5
CLINOPTILOLITE EXCHANGE CAPACITY
Method Used for
Ammonia Recovery
Exchange Capacity9
meq/g
Distillation of clinoptilolite particles
Elution using NaCl
Elution using CaCl2
1.95 (1.93-1.98)
1.88 (1.86-1.91)
1.77 (1.75-1.79)
Range of experimental values indicated in parenthesis.
The results show that the exchange capacity is affected by the method
used to recover ammonia from the zeolite. Values measured by the direct
distillation method and by elution with NaCl are nearly identical. A
small loss of ammonia from the NaCl eluant could account for the lower
value obtained by NaCl elution. It is also possible that a small amount
of ammonia remained in the zeolite and could have been eluted had more
NaCl been run through the column. Both of these values correspond
closely to the exchange capacity of 1.87 meq/g calculated by Barrer [50]
from the composition of clinoptilolite. It is reasonable to assume that
ammonium and sodium ions have access to practically all exchange sites
within the clinoptilolite framework. Elution with calcium resulted in
a smaller exchange capacity than did elution with sodium. From these
results it is not possible to determine whether the lower capacity was
due to a slower rate of exchange of calcium for ammonium ions or to the
inaccessibility of calcium to some of the exchange sites.
62
-------
EFFECT OF WATER COMPOSITION ON AMMONIA EXCHANGE CAPACITY
The objectives of these exhaustion studies were 1) to examine the
characteristics of breakthrough curves for clinoptilolite columns
exhausted with waters containing different concentrations of competing
cations, 2) to determine the equilibrium characteristics of the various
ions in the zeolite, and 3) to determine the effect of water composition
on the ammonia capacity of clinoptilolite.
In order to compute the total ammonia capacity of the zeolite and to
determine the selectivity coefficients, runs were continued until the
ammonia concentration in the column effluent equaled the concentration
in the influent. The studies were conducted using 3 ft of clinoptilolite
exhausted at a rate of approximately 15 BV/hr. The actual flow rates
and weights of material used for each test are indicated on the concen-
tration history curves. The compositions of the column influents for
these tests are shown in Table 6. Columns for all runs, except run 1,
were prepared by complete regeneration with 30 BV of regenerant contain-
ing 0.35 M NaCl (0.17 Ib NaCl/gal) and a pH of 12.5 adjusted using NaOH.
Clinoptilolite for run 1 was regenerated with 40 BV of 0.35 M NaCl
(0.17 Ib NaCl/gal) adjusted with NaOH to a pH of 12. Columns were
exhausted using water having cation compositions typical of domestic
wastewater.
TABLE 6
INFLUENT CHEMICAL COMPOSITION FOR SATURATION RUNS
Run
1
2
3
4
5
6
Constituent
NH3-N
mg/£
18.8
19.0
19.0
19.4
19.0
16.4
Na
mg/£
13.3
62.1
60.5
55.0
84.1
175
K
mg/A
1.0
11.7
12.5
14.0
14.8
35.4
Ca
mg/A
26
63
66
67
138
140
Mg
mg/£
5
8
16
3
53
48
pH
7.9
7.9
7.9
6.0
8.0
7.9
Equilibrium Characteristics
Selectivity coefficients were determined from concentration histories
measured for each ion during these runs using Equations 3 and 4. Solid
phase concentrations were determined by graphical integration of the
area between the breakthrough- curve and the influent ion concentration.
63
-------
The concentration history for run 1 is shown in Figure 11 as an example
of the results of these tests. Results of runs 2 through 6 are shown in
Appendix B. Calcium and magnesium data appear somewhat scattered in
these figures. In most instances'magnesium concentrations are lower and
calcium values are higher than expected. This was probably due to
measurement of some magnesium as calcium because of a buffered pH in
the calcium determination too low to precipitate all magnesium in the
sample solution. In these cases the actual data points are plotted in
the figures, but smooth curves have been drawn to match the results
expected from adjustment of these points. Adjustments to curves were
made only where an increase in the ""magnesium and an equivalent reduction
of calcium, or vice versa, resulted in more regular curves.
Examples of selectivity coefficients, are shown in Table 7. Values for
these runs illustrate differences in selectivities for a water typical
TABLE 7
EQUILIBRIUM VALUES FOR SATURATION RUNS
Item
KNa^
KNH4-N
KNC^-N, A/g
C~N> */g
Run 2
Experimental
Values'
1.12
0.60
1.32
1.84
Values from
Binary System
Ames [12,]
7.4
0.39
1.24
2.07
Run 6
Experimental
Values
1.44
.0.45
1.12
a
Values from
Binary System
Ames [12]
11.6
0.41
1.80
6.85
Value of q.. too small to be determined from experimental data.
of an "average" sewage (run 2) and for a highly mineralized water
characteristic of a sewage having a high hardness and subject to infil-
tration of saline water (run 6). Because the zeo-14te was initially in
the Na form, it was not nossible to directly measure q^a or the total
exchange capacity, Q. Solid phase sodium concentrations were obtained
by subtracting the sum of other solid phase concentrations from the
total exchange capacity, 1.9 meq/g. This method has the disadvantage
of accumulating errors in other solid phase concentrations into the
value of q^j,. Failure to reach equilibrium or the inability of ions to
be exchanged into all sites measured in laboratory exchange capacity
measurements are reflected in artificially high values of
64
-------
4.0
in
3.0 -
o~
QJ
o
E 2.0
to
s-
O)
o
.O.
o
1.0 -
0.0
Pbv = 5150 g
F = 16.6 BV/hr
(C0)NH _ = 1-34 meq/£
(C0)Ca =1.30 meq/a
(C0)N = 0.58 meq/£
100
200
300
700
800
400 500 600
Throughput, BV
FIGURE 11. EXHAUSTION OF CLINOPTILOLITE IN Na FORM
900
1000
-------
Selectivity coefficients determined by Ames [12] in batch binary systems
are also shown in Table 7 and may be compared to values obtained from
multicomponent systems in this study. In both runs the values of K^
measured from column data were less than corresponding values obtained
by Ames in batch systems. Values for the other selectivity coefficients
calculated from column data varied somewhat from values determined by
Ames. Three explanations of these differences may be offered: 1) that
equilibrium was not achieved throughout the column during the length of
the test, 2) that the column was not completely regenerated prior to the
beginning of the run, or 3) that binary equilibrium data do not adequately
describe multicomponent relationships. The first postulate seems unlikely
unless equilibrium is approached very slowly near the upper end of the
breakthrough curve. Data obtained in these tests indicated that equili-
brium was attained in each case. However, it is possible that diffusion
in the particle phase becomes extremely slow as equilibrium is approached
(cf. Equations 7 and 8). While care was taken to completely regenerate
columns to the Na form prior to each run, it is possible that some ions
located near the center of zeolite particles were incompletely eluted.
Effective Ammonia Exchange Capacity
The correlation of the ammonia exchange capacity of clinoptilolite to
the characteristics of the waste to be treated will aid in determining
the applicability of zeolitic ion exchange for ammonia removal. Condi-
tions for which the breakthrough ammonia exchange capacity is very small
will result in short run times and higher capital and operating expenses.
Ammonia exchange capacities both to 1 mg/a NH3-N breakthrough and to
complete column saturation for the runs shown in Figure 11 and in
Appendix B were calculated by integrating the area above the breakthrough
curves. The results, illustrated in Figure 12, show the variation of
ammonia exchange capacity with the cationic strength of the column
influent. While this relationship is definitely empirical and repre-
sents a simplification of the complex effect of competing cation concen-
tration on ammonia exchange capacity, it does illustrate the effect of
increased mineral strength on the exchange capacity of a relatively fixed
ammonia concentration. The reader is referred to Chapter V for a
discussion of the rationale of this approach. The influent ammonia
concentration for these tests varied between 16.4 and 19.0 mg/«, NH3-N
and was representative of concentrations found in domestic wastewater.
The influent compositions of runs 2, 3, and 4 were most indicative of
a chemically treated domestic sewage. Exchange capacities in this
range of cationic strength illustrate the sensitivity of qNHi+-N to changes
in water composition. The cationic strength for run 6 was higher than
that usually encountered in domestic wastewaters. However, values of
QNHv-N f°r runs 5 and 6 indicated that q^n^-N decreases slowly at a
cationic strength greater than about 0.01. While these relationships
will not be valid for any conceivable combination of cations, they do
provide an estimate of ammonia exchange capacities which can be expected
when clinoptilolite is used with many domestic wastewaters.
66
-------
0.7
0.6
CD
CT
0.5
cr
o
O)
O
fO
-------
Breakthrough ammonia capacities calculated to 1 mg/K, NH3-N leakage, also
shown in Figure 12, bear a similar relationship to cationic strength.
(Capacities were calculated to a breakthrough of 1 mg/«, NH3-N because
water quality requirements for ammonia removal are likely to stipulate
a product water containing no more than 1 mg/£ NH3-N.) Ammonia break-
through capacities averaged about 60% of the total ammonia capacity in
these runs.
COMPARISON OF PERFORMANCE OF Na- AND Ca-CLINOPTILOLITE
To determine the effect of the counter ion initially in the zeolite on
exhaustion performance, clinoptilolite columns in Na and Ca forms were
exhausted using tap water containing 20 mg/a NH3-N and the chemical
composition shown in Table 8. Clinoptilolite used for these tests had
previously been exhausted to saturation using waters having similar
cation concentrations such that the solid phase concentration of ions
contained in each sample of clinoptilolite was approximately the same.
In this way it was possible to make a comparison between the effective-
ness of sodium and calcium as regenerants. Clinoptilolite was placed
TABLE 8
INFLUENT CHEMICAL COMPOSITION FOR EXHAUSTION
OF Na- AND Ca-CLINOPTILOLITE
Ion
NH3-N, mg/£
Na, mg/s>
K, mg/£
Ca, mg/£
Mg, mg/£
PH
Concentration
Na Form
18.8
13.3
0.9
26
5
7.9
Ca Form
18.5
9.2
0.8
27
4
8.1
in the Na form by regeneration using 40 BV of a solution composed of
0.35 M NaCl (0.17 Ib NaCl/gal) with the pH raised to 12 using NaOH.
This corresoonded to the use of 45 Ib NaCl/cu ft and was sufficient to
reduce the effluent ammonia concentration in the regenerant to less than
3 mg/e NH3-N. Detailed regeneration studies described in Chapter VIII
indicate that a comparable degree of regeneration could be achieved
using a somewhat smaller amount of regenerant if regenerant optimization
is of major concern. Clinoptilolite in the Ca form was prepared by
68
-------
regeneration, with 0.17 M CaCl2 (0.16 Ib CaCl2/gal) in a saturated lime
solution having a pH of 12.2. However, the conversion of clinoptilolite
to the Ca form proved difficult. Only about 60% of the ammonia in the
zeolite was removed in 30 BV of regenerant. An additional 20 BV was
needed to complete regeneration of the column, bringing the total amount
of CaCl2 required to 147 lb/cu ft.
Concentration histories for these two runs are shown in Figures 11 and
13. Effluent concentrations of sodium, potassium, and magnesium in
Figure 13 indicate that these ions were not eluted from the column by
calcium regenerant. On the contrary, sodium and, to some extent,
potassium were displaced from the column by ammonia during exhaustion.
This occurred because these ions occupied larger fractions of the
zeolite exchange sites during a previous run and, therefore, were
present in greater than equilibrium quantities at the beginning of this
particular run. Effluent concentrations of calcium and magnesium for
Na-clinoptilolite in Figure 11 show a definite breakthrough indicating
that these ions were removed from the zeolite by the sodium regenerant.
While the elution of these ions does not directly affect the ammonia
capacity of the zeolite during exhaustion, this does illustrate the
restricted mobility of calcium in the. zeolite. These results also
indicate the importance of pH in the elution of ammonia from clinopti-
lolite. Because ammonia was the only ion significantly removed from the
zeolite by the calcium regenerant, this implies that the conversion of
the ammonium ion to unionized ammonia by the high pH regenerant leads
to more rapid elution .of ammonia from the zeolite pores.
The relatively greater difficulty of regenerating clinoptilolite with
calcium than sodium ions can be explained from the clinoptilolite selec-
tivity series. Because the zeolite is more selective for sodium than
calcium, it can be argued'that'regeneration should be more easily
accomplished using sodium salts. However, it follows that during
exhaustion the kinetics of NH^-Na exchange would be less favorable than
for NH^-Ca exchange. Slopes of the normalized ammonia breakthrough
curves for these two cases shown in Figure 14 reveal that the kinetics
of NH^-Na exchange was superior to NH4-Ca exchange. This demonstrates
the influence of particle phase diffusion on exchange kinetics. Although
clinoptilolite selectivity results in the preference Na > Ca, steric
factors which affect both selectivity and diffusion of ions through the
zeolite restrict the mobility of calcium ions more than sodium ions.
This results in superior exchange kinetics for sodium relative to
calcium regardless of whether these ions are entering the zeolite or
are being displaced from it.
These observations are explainable in terms of the size of the hydrated
cations (cf. Table 3). Due to its large size (19.2 A diameter), calcium
ions cannot approach exchange sites as closely as other ions and,
therefore, are less preferred by the zeolite. However, once in the
zeolite, the mobility of the ions into the fluid is restricted because
of.their affinity, for the structural water contained in the zeolite
69
-------
-j
o
2.0
1.8
1.6
4^1.4
cr
Si.2
^
| 1.0
IT3
I 0.8
O)
o
3 0.6
0.4
0.2
0.0
1 1
Pbv = 4700 g
F = 18.2 BV/hr
(C0)Ca = 1.36 meq/jl
(C0)M = 0.3 meq/£
100
200
300
700
400 500 600
Throughout, BV
FIGURE 13. EXHAUSTION OF CLINOPTILOLITE IN Ca FORM
800
900
1000
-------
1.0
0.8
o
o
0.6
0.4
0.2
0.0
200
« Ca Form
(CO)|\JH _N = 18>5 m9/£
Pbv = 4700 g
F = 18.2 BV/hr
Na Form
Pbv = 5150 g
F = 16.6 BV/hr
400 600
Throughput, BV
800
1000
1200
FIGURE 14. AMMONIA CONCENTRATION HISTORIES FOR Ca- AND
Na-CLINOPTILOLITE
-------
framework. On the other hand, the sodium ion, having a hydrated diameter
of only 15.8 A, is freer to migrate through the zeolite channels and to
approach more closely the exchange sites.
The total exchange capacity of the Na-clinoptilolite calculated by
integration of the area above the breakthrough curve was 0.68 meq/g,
and that of the Ca-clinoptilolite was 0.69 meq/g. However, the capacity
to 1 mg/£ NH3-N breakthrough was reached in 220 BV for the Na form, but
in only 100 BV for the Ca form. Thus, while the total exchange capa-
cities for the two counter ions were approximately the same, superior
column kinetics for Na-NH^ exchange resulted in a breakthrough capacity
for Na-clinoptilolite which was more than twice as great as that for
the Ca form.
Breakthrough Curve Characteristics
Several observations concerning the kinetic performance of clinoptilolite
columns may be made from the concentration histories presented in this
section and in other parts of the chapter. From Figure 11, and Figures
1 through 5 in Appendix B, the selectivity series of clinoptilolite may
be determined by the order in which the ions appear in the effluent. As
sodium was the ion initially in the zeolite, the sodium curves must be
looked upon as "reverse" breakthrough curves. The order in which ions
appear in the effluent identifies the order of selectivity as K > NH4 >
Na > Ca > Mg which agrees with the results of Ames.
Although breakthrough curves for ammonia on Na-clinoptilolite are shown
in Figure 11 and in Appendix B, the shape of the curve is best illustrated
in Figure 14. All curves demonstrate a relatively sharp breakthrough at
the lower end of the curve followed by a decrease in slope as the
breakthrough progresses. The upper end of the curves tapers off and
approaches equilibrium relatively slowly. This type of curve is
characteristic of exchangers in which either solid or pore diffusion
limits the rate of exchange. While the sharp breakthrough in the lower
portion of the curve tends to increase the breakthrough capacity of the
zeolite, the shallowness of the rest of the curve shows that a signifi-
cant portion of the zeolite remains unsaturated when the breakthrough
occurs.
EFFECT OF pH ON AMMONIA EXCHANGE
As discussed in Chapter V, the uptake of ammonium ions in clinoptilolite
will be significantly affected by pH. An optimum range for ammonium
exchange will exist which will be determined by competition of exchange
sites for ammonium and hydrogen ions at low pH and by NH3-NH+ equilibria
at high pH. Be determining the optimum pH range it will be possible in
column operation to adjust the influent pH to achieve the maximum removal
72
-------
of ammonia. In Chapter V a model of this exchange reaction was developed
which is verified with experimental data in this section. Preliminary
batch tests were made in a ternary NH^-Na-H system using ion concen-
trations similar to those found in domestic wastewaters. Subsequent
column tests were made using an influent solution containing all cations
present as macrocomponents in domestic wastewaters.
Batch Tests
Preliminary batch tests were performed to determine the influence of pH
on the uptake of ammonia from a sample NH4-Na-H solution not unlike that
encountered in wastewaters. Equilibration solutions consisted of 20 mg/a
NH4-N and 60 mg/a Na in the form of chloride salts which were adjusted
to pH values ranging from 2 to 12 using HC1 and NaOH. One hundred
milliliters of this solution were contacted with 0.5 g of clinoptilolite
for 2 hr. Clinoptilolite was prepared by contact with 1 M NaCl, washing
to remove excess sodium, and drying at 105°C as described above in the
determination of clinoptilolite ion exchange capacity. Control samples
containing no clinoptilolite were run to determine the amount of ammonia
lost to the atmosphere. Samples were shaken in 125-m£ erlenmeyer flasks
using a Burrell Wrist-Action Shaker (Burrell Corp., Pittsburgh, Pa.).
Flasks were capped with rubber stoppers to limit escape of ammonia, to
the atmosphere. For these tests ammonia was determined by direct
nesslerization [80] since no interferences were present.
Results of the 2-hr equilibration are shown in Figure 15. The optimum
nH range for ammonia exchange was between 6 and 8. At pH values above
8 the solid phase ammonia concentration decreased rapidly to a pH of
10.7 where no detectable ammonia exchange took place in the 2-hr period.
The data taken at final pH values below 6.0 were influenced by acid
soluble impurities in the clinoptilolite which acted as a buffer,
increasing the final pH of the solution and distorting the observed
effect of pH on ammonia exchange below pH 6.
Sodium added to adjust the pH to higher levels depressed the solid phase
ammonia concentration, q|\jHi+-N> resulting in less exchange than would be
expected by hydroxyl ion effects alone. However, the amount of sodium
added in raising the pH to 10 was only about 2 mg/£, while the amount
added in raising the pH to 11 was about 20 mg/£. Thus significant
sodium interference would not be expected below a pH of about 11. This
may have affected samples having final pH values of 10.7 and 11.8 in
Figure 15, but little ammonia exchange would be expected at this pH on
the basis of ammonia equilibria. The effect of pH on the ammonia capa-
city may also be expressed as the ammonia capacity at a given pH relative
to the capacity at a pH unaffected by NH^-H exchange or by the presence
of unionized ammonia. Ammonia exchange is least affected by these
factors between pH 6 and 8. For the batch results in Figure 15 the
ammonia capacity unaffected by pH is about 0.25 meq/g. Data normalized
with respect to this value appear in Figure 16. Also shown in Figure 16
73
-------
O>
O en
o cr
C_J d)
0.3
0.2
o i
E J
i o.i
OJ
CO
ra
O
OO
0.0
o
o
2.0 4.0
FIGURE 15.
6.0 8.0
Final pH
10.0 12.0
BATCH RESULTS - EFFECT OF pH
ON AMMONIA EXCHANGE
1.0 -
o
0.8
0.6
0.4
0.2 -
0.0
o o Equation 18
o normalized to pH 6.0 o
o
o 2-hr equilibration A
Column to saturation o
10.0 12.0
FIGURE 16. RELATIVE EFFECT OF pH
ON AMMONIA EXCHANGE CAPACITY
74
-------
is the curve predicted from Equation 18 normalized with respect to the
value predicted by Equation 18 at pH 6. Values for this line were
obtained using Q = 1.9 meq/g as determined in the section dealing with
the ion exchange capacity of clinoptilolite. While these data do not
permit a complete and rigorous description of the batch system, they do
demonstrate the validity of the NH3-NH+-Na+-H+ interrelationships
described in Equation 18.
Column Tests
Column tests were run to confirm the results obtained in batch studies.
Runs were made using 3-ft bed depths and exhaustion rates of approxi-
mately 18 BV/hr. Prior to each run, columns were regenerated to the Na
form using 30 BV regenerant containing 0.35 M NaCl (0.17 Ib NaCl/gal)
and sufficient NaOH to raise the pH to 12.5. Column influents were
adjusted to pH 4.0, 6.0, 8.0, 9.5, and 10.0 and exhaustion was continued
until the concentration of ions in the effluent equaled those of the
influent. Ammonia breakthrough curves for these runs are compared in
Figure 17. Ammonia data and effluent pH values are shown in Appendix C.
Influent chemical compositions shown in Table 9, were chosen to be
reasonably representative of a domestic sewage chemically treated with
lime. Since runs were continued until columns were in equilibrium with
the feed, these values are also the equilibrium concentrations. The
TABLE 9
CHEMICAL COMPOSITION FOR COLUMN pH RUNS
PH
4.0
6.0
8.0
9.5
10.0
Concentration, mg/£
NHs-N
21.2
19.5
18.4
18.5
18.4
Na
65.5
55.0
45.3
68.0
97.5
K--
14.6
14.0
12.4
12.7
12.6
Ca
82
67
62
59
38
Mg
3
3
2
3
4
effluent pH, shown in the figures in Appendix C, varied somewhat from
the influent values during some runs, although the equilibrium pH was
that shown in Table 9. High effluent pH at the beginning of the pH 4
run was probably a result of acid soluble impurities in the clinoptilolite
and lime particles not removed after the previous regeneration. The pH
for the pH 10 run was adjusted initially to pH 10.5; however, the
influent pH measured in the head tank and the effluent pH averaged 10.0
This decrease was probably due to precipitation of CaC03 following pH
75
-------
1.0 _
o
o
I
ro
1C
0.8 -
0.6 -
0.4 ~
0.2
0.0
100 200 300 400 500 600 700 800
Throughput, BV
FIGURE 17. EFFECT OF pH ON AMMONIA BREAKTHROUGH
-------
adjustment1and was reflected in the lower influent calcium concentration
for this run. Periodic determination of calcium in filtered and unfil-
tered samples showed that a minimum of 98% of the calcium was soluble.
The greatest ammonia capacity was observed at pH 6. Slightly lower
capacities1at pH 4 and 8 reflect increased competition from hydrogen
ions for exchange sites at pH 4 and a shift in the NHs-NHtf equilibria
at pH 8. Column ammonia capacities for these runs determined from the
ammonia breakthrough curves are shown in Table 10. Capacities were
calculated both for full column saturation and to 1 mg/a NH3-N break-
through. Values for complete saturation, normalized to the capacity at
TABLE 10
COLUMN AMMONIA CAPACITY FOR pH RUNS
Influent pH
4.0
6.0
8.0
9.5
10.0
qMM M to Saturation
IN n L. ~ IN
meq/g
0.48
0.51
0.48
0.37
0.14
qNH4-N to ] mg/* NH3'N
Breakthrough
meq/g
0.27
0.33
0.24
0.06
0.01
pH 6 (0.51 meq/g), are shown in Figure 16. These Values closely follow
the capacities predicted by Equation 18. While the ternary system
described in Equation 18 is less complex than the 6-component system
used for column runs, the effect of pH in the two systems was essentially
the same. Sodium, potassium, calcium, and magnesium ions in the column
system performed the same function as the Na present in the batch
system namely, to replace ammonium ions in the zeolite as the NH4/
total cation ratio decreased at high pH. At
hydrogen ion concentration acted to decrease
which would be exchanged into the zeolite at
indication of relative changes in the ammonia exchange capacity, Equation
18 provides a valid description of the more complex multicomponent
system.
Several observations may be made concerning the data in Figure 16 and
Table 10. Optimum conditions for exhaustion occur between pH 4 and pH
8. Column operation outside this range resulted in a rapid decrease of
ammonia exchange capacity and increased ammonia leakage prior to the
onset of breakthrough. On the other hand, the ineffectiveness of ammonia
exchange at high pH can be utilized advantageously during regeneration.
the zeolite as
low pH, increases in the
the fraction of each ion
a more neutral pH. As an
77
-------
Equation 18 predicts equilibrium ammonia capacities of 0.53 meq/g at
pH 10, 0.08 meq/g at pH 11, and 0.008 meq/g at pH 12 for a solution
concentration of 20 mq/£ NH3-N. While regeneration is not an equili-
brium process, regenerants having sufficient buffer capacity to keep
the pH above 11 throughout the column would result in an increased rate
of ammonia desorption.
SUMMARY
The exhaustion performance of clinoptilolite has been examined using
chemically defined systems containing cation concentrations typical of
domestic sewages. Column tests were continued until the zeolite was in
equilibrium with the influent water. It was demonstrated that the ammonia
exchange capacity of clinoptilolite can be estimated in terms of the
cationic strength of the column influent. From data presented in Figure
12, it is possible to estimate the ammonia exchange capacity of clinop-
tilolite for a particular wastewater having a composition which is not
substantially different from those shown in Table 6. This method
provides a practical means of estimating exchange capacities for design
purposes.
Exhaustion of clinoptilolite in Na and Ca forms revealed that, while the
total ammonia exchange capacity was identical for each form, more
favorable column kinetics resulted in more than twice as much break-
through capacity (to 1 mg/2, NH3-N effluent concentration) for Na-
clinoptilolite. In addition, regeneration was much less efficient using
calcium salts. These results show that the operation of clinoptilolite
on the Na cycle is the preferred mode of operation.
Results of batch and column tests made to elucidate the effect of pH on
ammonia exchange showed that optimum conditions for ammonia exchange
exist between pH 4 and 8 with the ammonia exchange capacity decreasing
rapidly outside this range. A model of the exchange reaction which
included effects of increased hydrogen ion competition at low pH and
NH3-NH£ equilibria at high pH (cf. Equation 18) closely predicted
changes in the ammonia exchange capacity with pH. These results will
help in combining this process with upstream treatment processes in a
way to attain maximum ammonia removals. In addition, the results will
aid in determining column rinsing procedures following high pH regene-
ration which will not impair column performance at the beginning of the
subsequent exhaustion cycle.
78
-------
VIII. REGENERATION STUDIES
GENERAL CONSIDERATIONS
Objectives
Considerations in determining the least cost method of regeneration are
the efficiency with which regenerant is utilized, the volume of regener-
ant required, and the time necessary to complete regeneration. Experi-
ments reported in this chapter were designed to furnish information
necessary for making decisions regarding the choice of regenerant
composition. Objectives of the regeneration study were: 1) to deter-
mine the effect of flow rate on regeneration efficiency, 2) to determine
the most suitable concentration of salt and pH for regeneration, 3) to
determine the effect of attrition caused by exposure of clinoptilolite
to high pH regenerant solutions, and 4) to ascertain the rinse require-
ments of regenerated columns.
In the preceding chapter it was shown that improved performance during
exhaustion resulted from regeneration of clinoptilolite to the Na form.
Therefore, efforts to optimize column regeneration concentrated on the
use of NaCl and NaOH. Sodium hydroxide was used for pH adjustment.
However, because of its lower cost, lime would probably be the most
desirable source of caustic for pH adjustment in full-scale plants.
This substitution will result in little or no change in performance
even with the introduction of some calcium into the exchanger during
regeneration.
Column Operation
All regeneration studies were performed using chemically treated sewage
for the column influent. This work was carried out concurrently with
the process demonstration studies at SERL described in Chapter IX. The
average cation concentrations in the column influent and the range of
values during the test period are shown in Table 11.
Each run commenced with zeolite regenerated to the Na form and continued
until incipient breakthrough of ammonia. Data collected using tap water
containing cation concentrations similar to those in Table 11 (cf. Figures
1 and 2 in Appendix B) indicated that a throughput of approximately 180
BV could be processed prior to ammonia breakthrough. Therefore, columns
were exhausted at a rate of 15 BV/hr for 12 hr and immediately regene-
rated. In this way columns were loaded with approximately the same
amount of ammonia prior to each regeneration. Because of variations in
79
-------
the influent wastewater comoosition, it was not possible to maintain
orecisely the same ammonia loading for each run. Therefore, corrections
were made for the amount of ammonia removed during each cycle in all
calculations concerning regeneration performance.
TABLE 11
COMPOSITION OF COLUMN INFLUENT
FOR REGENERATION STUDIES
Ion
NH3-N
Na
K
Ca
Mg
Average
Concentration
mg/£
19.0
54.0
9.3
71
5
Range
mg/£
13.0 - 22.9
48.3 - 62.1
7.8 - 11.3
48 - 94
1 - 13
Column performance and other data pertinent to the regeneration study
are shown in Table 12. In each regeneration test 30 BV regenerant was
applied to the column. Run designations are those used in the process
demonstration studies at SERL so that the supplementary pilot-plant data
in Appendix E may be referred to using these run numbers. Table 12
includes data describing the exhaustion of columns prior to each
regeneration, the conditions under which each regeneration was made, and
the performance of columns during regeneration. A partial substantiation
of the reliability of these tests is afforded by comparing the ammonia
removed during exhaustion to the quantity of ammonia eluted during the
following regeneration. Rarely was less than 80% of the ammonia removed
during exhaustion recovered during regeneration. This is satisfactory
considering the circumstances under which the tests were conducted and
the volatility of ammonia in the high pH regenerant. While care was taken
in collecting regenerant samples, some loss of ammonia was inevitable.
The low recovery of ammonia in runs 63, 64, and 67 was a result of the
composition of the regenerant used in these runs. Appropriate correc-
tions were made and these are described below. For these runs, in
which ammonia elution was obviously poor, a stronger regenerant was
applied to the column following the 30 BV used for these tests in
order to have a column free of ammonia at the beginning of the next
exhaustion cycle.
80
-------
TABLE 12
COLUMN PERFORMANCE DURING REGENERATION STUDIES
Run
SERL-12
10
3
18
67
27
49
64
16
22
24
26
53
66
51
38
40
46
52
58
33
31
29
44
35
48
56
63
60
65
Exhaustion
Inf.
NH3-N
mg/J,
15.7
15.2
15.2
18.6
20.0
17.5
19.3
22.0
13.0
18.0
16.9
21.0
20.2
22.0
20.0
16.6
19.0
20.0
16.3
22.3
20.0
17.9
18.8
18.1
19.5
18.4
20.8
19.9
22.0
22.9
Eff.
NH3-N
mg/S.
0.56
0.72
1.10
1.09
2.95
0.20
1.97
0.53
0.02
0.29
0.27
0.43
0.98
4.06
3.30
0.28
0.47
0.30
0.47
2.03
0.49
0.11
0.25
0.68
0.31
1.31
1.07
1.82
1.46
4.62
NH3-N
Removed
During Run
equivalents
1.44
1.39
1.34
1.67
1.63
1.65
1.65
2.06
1.24
1.78
1.59
1.97
1.84
1.71
1.60
1.-56
1.77
-1.89
1.51
1.94
1.86
1.70 ~
1.77
1.71
1.83
1.63
1.89
1.73
1.95
1.75
Regenerant3
Flow
(Upflow)
BV/hr
4
7
10
20
30
10
25
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
NaCl
Cone.
Ib/gal
0.049
0.049
0.049
0.049
0.049
0.24
0.10
none
0.049
0.10
0.17
0.24
0.73-
none
0.049
0.10
0.10
0.17
0.24
none
0.049
0.10
0.17
0.17
0.24
0.73
0.73
0.10
O.lOf
0.10f
PH
11.5
11.5
11.3
11.6
11.5
11.5
12.5
11.5
11.5
11.5
11.5
11.5
-11.5
12.0
12.0
12.0
12.0
12.0
12.0
12.5
12.4
12.4
12.3
12.5
12.4
12.5
12.5
8.2
12.2
11.8
NH3-N Eluted
equivalents
1.65
1.36
1.48
1.63
1.02
1.50
1.49
1.26
1.04
1.50
1.56
1.69
1.82
1.62
1.84
1.26
1.86
1.54
1.57
2.05
1.72
1.37
1.35
1.64
1.46
1.61
1.80
0.96
1.84
1.69
Ammonia
Recovery
%
114
98.0
108
97.5
62.8
91.0
90.4
61.5
84.5
84.3
98.0
87.7
99.1
94.1
115
79.0
105
81.5-
104
105
92.4
80.5
76.2
95.8
80.1
99.4
95.5
55.2
106
96.8
Regeneration Efficiency
Regen.
to 95%
Elution
BV
27.0
23.5
27.0
25.4
_
14.7
11.0
_
26.6
26.0
18.2
17.0
16.0
28.0
21.0
16.6
16.8
16.0
17.0
15.3
15.0
10.0
9.2
9.0
9.4
9.0
10.0
51. 6d
14.5
20.0
Na Used
equivalents'5
20.9
18.3
20.9
19.7
55.0
18.9
_
20.7
39.5
47.5
63.2
178
3.1
17.9
26.5
26.7
43.4
64.9
3.7
14.8
17.3
26.2
25.6
37.3
101
112
76. 6d
22.6
31.2
Efficiencyc
%
6.9
7.6
6.4
8.5
-
3.0
8.8
_
6.0
4.5
3.3
3.1
1.0
55.0
8.9
5.9
6.6
4.4
2.3
52.0.
12.6
9.8
6.8
6.7
4.9
1.6
1.7
2.3d
8.6
5.6
aVolume of regenerant used was 30 BV. Includes total Na in regenerant.
c . Eq NH3-N removed during previous exhaustion d, , , . . . . . . .. ., ,_, j
Efficiency = -^ Eq Na used for regeneration x 100' Values obtained by extrapolating available data.
eRegenerant also contained 0.046 Ib CaCl2/gal.
fRegenerant also contained 0.32 Ib CaCl2/gal._
-------
EFFECT OF REGENERATION FLOW RATE
Because exchange rates in clinoptilolite tend to be limited by particle
phase diffusion, it was initially expected that exchange efficiency
might vary considerably with flow rate. Therefore, tests were made to
establish regeneration flow rates which could be used in later experi-
ments. Regenerant was prepared with tap water containing 0.049 Ib NaCl/
gal (0.1 M) and with the pH adjusted to 11.5 with NaOH. Runs were made
using flow rates ranging from 4 to 30 BV/hr. Ammonia concentration
histories for these runs, shown in Figure 18 reveal little difference
in rates ranging from 4 to 20 BV/hr. Points for the 15 BV/hr run were
lower than points for other runs in this range because the influent
ammonia concentration in the previous exhaustion was lower than average
(cf. Table 12). Points for regeneration at 30 BV/hr are much below the
average of the other runs. Only 63% of the ammonia in the zeolite was
eluted in this regeneration indicating that exchange kinetics become
limiting between 20 and 30 BV/hr.
As a further means of elucidating the effect of flow rate on regeneration,
the amounts of ammonia eluted in various volumes of regenerant were
compared. By assuming that essentially all ammonia was eluted from the
clinoptilolite in 30 BV of regenerant, it was possible to normalize the
amount of ammonia eluted in 5, 10, and 20 BV with respect to the amount
removed in 30 BV. This was a good assumption for flows between 4 and
20 BV/hr, but was not valid for regeneration at 30 BV/hr as only 63%
of the ammonia was eluted in 30 BV regenerant. Values for this flow
were corrected using the quantity of ammonia removed during exhaustion.
The percent ammonia eluted in the various regenerant fractions is
shown in Figure 19a. If there was no effect of increasing the flow rate,
the fraction of ammonia eluted in a particular volume of regenerant
should be the same for all flows. These results show that the amount
of ammonia removed per volume of regenerant is approximately constant
up to a rate of 15 BV/hr. A slight reduction was evident at the 20
BV/hr rate after 5 and 10 BV had passed through the column. However,
by the time 20 BV had been passed through the column, no difference was
observed between this flow rate and lesser ones. Performance at 30 BV/
hr was definitely inferior compared to the lower flow rates, resulting
in decreased regenerant utilization.
To determine if a similar relationship existed for different regenerants,
regeneration characteristics were compared at 10 and 15 BV/hr using
0.24 Ib NaCl/gal (0.5 M NaCl) at pH 11.5 and at 15 and 25 BV/hr using
0.10 Ib NaCl/gal (0.2 M NaCl) at pH 12.5. The percentages of ammonia
eluted for these conditions, illustreated in Figure 19b, show little
difference in performance for comparative regeneration conditions.
Ammonia removal at 25 BV/hr using the 0.10 Ib NaCl/gal, pH 12.5 regene-
rant lagged only slightly behind the removal rate at 15 BV/hr. On the
basis of these results it can be concluded that variation in flow rates
in the range of 4 to 20 BV/hr has no significant effect on ammonia
82
-------
oo
CO
200
150
en
100
I
ro
I I I
Regeneration made using
0.049 Ib NaCl, pH 11.5
4 BV/hr (all upflow)
7 BV/hr
10 BV/hr
15 BV/hr
20 BV/hr
30 BV/hr
O Run
A Run
O Run
D Run
O Run
Run
Average
4-20 BV/hr
50 h
10
15 20
Regenerant Volume, BV
25
30
35
FIGURE 18. EFFECT OF FLOW RATE ON REGENERATION
-------
00
-p.
100
80
60
40
20
I
10 20
Flow, BV/hr
30
a. REGENERATION USING 0.049 lb NaCl/gal,
DH 11.5
100
80
-a
a>
= 60
40
20
0
5 BV
A 0.24 lb NaCl/gal, pH 11.5
D 0.10 lb NaCl/gal, pH 12.5
I
I
10 20
Flow, BV/hr
30
b. REGENERATION USING 0.24 lb NaCl/gal
pH 11.5 AND 0.10 lb NaCl/gal, pH 12.5
FIGURE 19. VARIATION OF AMMONIA ELUTION WITH FLOW RATE
-------
elution. Flow rates as high as 25 BV/hr might be used with only minor
impairment of exchange efficiency, but rates of 30 BV/hr or higher result
in a significant loss of efficiency.
OPTIMIZATION OF REGENERANT COMPOSITION
Effect of Salt Strength and pH
Because of the combined effects of pH and salt concentration on the
regeneration of clinoptilolite, tests were made to determine the depen-
dence of regeneration efficiency on these factors. In Chapter VII it
was shown that ammonia sorption was limited at high pH, with only 0.008
meq/g NH3-N held on the zeolite in equilibrium with a 20 mg/£ NH?-N
solution at pH 12. This indicates that regenerant utilization might be
improved using high pH regenerant solutions.
Salt concentrations used in these tests were 0.049, 0.10, 0.17, 0.24,
and 0.73 Ib NaCl/gal corresponding to 0.1, 0.2, 0.35, 0.5, and 1.5 M
NaCl. Tests were also made using high pH solutions containing no NaCl.
Solutions were adjusted to pH 11.5, 12.0, and 12.5 using NaOH. In
raising the pH the NaOH added equaled 0.0015 Ib NaOH/gal for pH 11.5,
0.005 Ib NaOH/gal for pH 12.0, and 0.011 Ib NaOH/gal for pH 12.5. The
regeneration rate was constant at 15 BV/hr. Data for these tests are
sbown in Table 12. Elution curves drawn for each test are shown in
Figures 20 through 25. The completeness of ammonia elution for 5, 10,
and 20 BV of regenerant is shown in Figure 26. As before, the ammonia
eluted in 30 BV was taken as the total amount of ammonia in the clinopti-
lolite and used as the base in calculating the percent ammonia eluted.
Because of the low ammonia recovery for run 64, values for this test
were corrected using the quantity of ammonia removed during exhaustion.
From these results it is evident that a given quantity of ammonia is
eluted in a progressively less regenerant volume as the pH is increased
from 11.5 to 12.5. At a pH of 12.5 practically all ammonia was eluted
from the clinoptilolite in 10 BV of regenerant, whereas at pH 11.5 and
12.0 elution was not nearly as complete in 10 BV. At pH 11.5 elution
approached 100% using 20 BV of regenerant only when the salt concen-
tration was 0.17 Ib NaCl/gal or greater. At pH 12.0 nearly complete
elution of ammonia was achieved in 20 BV using 0.10 Ib NaCl/gal or
greater concentrations. Approximately the same degree of elution at
pH 12.5 was achieved in 10 BV using a salt concentration of 0.10 Ib NaCl/
gal or greater. The effect of pH on regenerant performance is illus-
trated by run 63 shown in Figure 22. This run was made using a regene-
rant solution containing 0.10 Ib NaCl/gal and no caustic (pH 8.2). At
this pH ammonia was eluted from the column at a more or less constant
rate. As the pH of the regenerant was increased, elution curves became
sharper, resulting in the removal of ammonia in much less regenerant
volume.
85
-------
400r
5 10 15 20 25 30
Regenerant Volume, BV
FIGURE 20. AMMONIA ELUTION - NO NaCl
0«
en
£-Run 16, pH 11.5
5 10 15 20 25
Reqenerant Volume, BV
"IGURE 21. AMMONIA ELUTION -
0.049 Ib NaCl/gal
86
-------
Run 22, pH 11.5
Run 63, pH 8.2 _
5 10 15 20 25
Regenerant Volume, BV
FIGURE 22. AMMONIA ELUTION -
0.10 Ib NaCl/gal
600
5 10 15 20 25
Regenerant Volume, BV
FIGURE 23. AMMONIA ELUTION -
0.17 Ib NaCl/gal
30
87
-------
5 10 15 20 25
Reqenerant Volume, BV
FIGURE 24. AMMONIA ELUTION -
0.24 Ib NaCl/gal
1000
Run 48, pH 12.5
Run 56, pH 12.5
5 10 15 20 25
Reqenerant Volume, BV
FIGURE 25. AMMONIA ELUTION -
0.73 Ib NaCl/qal
-------
oo
to
10
0.0 0.05 0.10 0.15 0.20 0.25 0.75 TU)0.05 0.10 0.15 0.20 b725 0.75
0.0 0.05 0.10 0.15 0.20 0.25
Regenerant Strength, Ib NaCl/gal
a. REGENERATION AT pH 11.5 b. REGENERATION AT pH 12.0 c. REGENERATION AT pH 12.5
FIGURE 26. VARIATION OF AMMONIA ELUTION WITH REGENERANT STRENGTH
-------
Several observations may also be made concerning the effect of NaCl
concentration on regeneration. At pH 11.5 ammonia elution was enhanced
with increasing salt concentration up to 0.17 Ib NaCl/gal. Increasing
the salt concentration beyond this value had little effect on elution,
especially with the 20 BV required to remove greater than 95% of the
ammonia from the zeolite. Likewise, increasing the salt concentration
beyond 0.10 Ib NaCl/gal at pH 12.0 and 12.5 had no effect on regeneration
performance. For this range of caustic concentrations, increasing the
NaCl concentration of the regenerant beyond a certain value at a parti-
cular pH had no effect on the rate of ammonia elution. In regenerating
at pH 12.5, no benefit was realized by using a salt concentration greater
than about 0.10 Ib NaCl/gal. This result is surprising because of the
relatively low concentration at which salt addition ceased to improve
ammonia elution. Using conventional ion exchange resins, regeneration
is usually accomplished with a 5 to 10% solution. In this case it was
not beneficial to use a salt concentration greater than about 1%.
The nature of the ammonia elution characteristics observed by varying
regenerant composition suggests that the regenerant pH is the control-
ling factor in determining the volume of regenerant required to remove
ammonia from clinoptilolite. This implies that raising the pH plays a
significant role in increasing the diffusion rate of ammonia from the
solid phase. Based on observations made in these experiments, the
following mechanism of ammonia removal from the solid phase is proposed.
(The reader is referred to Figure 2 for an illustration of the various
parts of the zeolite particle.) Because anions are, to a great extent,
excluded from the intercrystalline channels of the zeolite, no effect of
increasing the solution pH is felt inside the crystallites. However,
hydroxyl ions probably are free to diffuse into the intercrystalline
pores, as these passageways do not contain negatively charged exchange
sites. At this point ammonium and hydroxyl ions are free to react
producing unionized ammonia and water. Due to its smaller size and
reduced affinity for water of hydration, the ammonia molecule thus
produced is capable of diffusing more rapidly to the bulk solution
phase outside the particle. The basic requirement necessary for ion
exchange to occur still exists; namely, that an adequate concentration
of a counter ion (sodium in this case) must be present to replace the
ammonium ion on the exchange site. Once exchange has taken place,
hydroxyl ions act to increase the diffusion rate of ammonia, both by
reducing physical constraints and by keeping the concentration differ-
ential between solid and solution phase ammonium ion concentrations at
a maximum.
Amount of Regenerant Required
The implication of these results to design and operation of ion exchange
plants becomes more apparent when the weight of chemicals and the volume
of regenerant required for regeneration are considered. In conventional
ion exchange practice the amount of regenerant used is frequently
90
-------
expressed as the weight of salt applied per cubic foot of resin. Because
of the importance of pH in addition to salt strength in the elution of
ammonia from clinoptilolite, the amount of caustic used must also be
considered. The cost of chemicals for regeneration will be proportional
to these values. The volume of regenerant required will also be related
to process costs as an indication of the amount of storage space needed,
of the amount of product water that must be used for regenerant makeup
where regenerant is wasted after one use, and of the volume of regenerant
to be processed through a stripping tower if regenerant is reused.
The volume of regenerant required for 95% elution of ammonia per equi-
valent of ammonia in the zeolite bed was calculated for different
regenerant compositions from the data given in Table 12. These values
are shown in Figure 27. The weight of chemicals required for the same
regenerant compositions is shown in Table 13. In all but two instances
lime was used as the caustic for regenerant composition listed in Table
13 as lime would be the least expensive source of caustic. The amount
of lime required was calculated from the amount of NaOH used in the
regeneration tests reported in Table 12 assuming that equivalent amounts
of NaOH and lime would be required for pH adjustment.
Regeneration of clinoptilolite using no NaCl at pH 12.0 or 12.5 must be
made using NaOH for pH adjustment as indicated in Table 13. The compari-
sons of Na- and Ca-clinoptilolite in Chapter VII showed that regeneration
and exhaustion are much more favorable using regenerants containing
sodium. Therefore, regeneration using lime by itself was not considered.
It is also observed from Table 13 that nearly identical amounts of NaOH
are required for regeneration at pH 12.0 and 12.5 using no NaCl 1.06
Ib NaOH/cu ft for regeneration at pH 12.0, and 1.11 Ib NaOH/cu ft at
pH 12.5. However, in Figure 27 the volumes of regenerant required using
these regenerant compositions are approximately 32 gal/eq NH3-N removed
at pH 12.0 and 15 gal/eq NH3-N removed at pH 12.5. Regeneration at pH
12.5 would probably be more favorable because of the smaller regenerant
volume required.
These data were obtained from regenerating 3-ft deep clinoptilolite beds
previously exhausted to a breakthrough NH3-N concentration of approxi-
mately 1 mg/£. Volumes for deeper beds and beds containing a higher
solid phase concentration of ammonia may vary somewhat from these values.
However, a fully exhausted column containing 9.5 eq NH3-N/cu ft regene-
rated using 0.17 Ib NaCl/gal at pH 12.5 required 11.8 gal regenerant/eq
NH3-N in the zeolite. The volume indicated in Figure 27 is 11.1 gal/eq
NH3-N. Corresponding chemical requirements were 2.0 Ib NaCl/eq NH3-N
removed and 0.12 Ib Ca(OH)2/eq NH3-N for the saturated column compared
to 1.87 Ib NaCl/eq NH3-N and 0.11 Ib Ca(OH)2/eq NH3-N calculated from
values in Table 13.
91
-------
i-D
ro
50
o
I
LO
40
O T3
M- OJ
>
-a o
(U E
S- O)
30
CU I
i-
OJ
O) (O
en en
O) O
O r
20
pH 11.5
DH 12.5
0.0 0.10 0.20 0.30 0.40 0.50
Salt Concentration, Ib NaCl/gal
-O
0.60
0.70
FIGURE 27. VOLUME OF REGENERANT REQUIRED FOR
95 PERCENT AMMONIA ELUTION
-------
TABLE 13
AMOUNT OF CHEMICALS REQUIRED FOR REGENERATION
Regenerant
Composition
PH
11.5
12.0
12.5
i
Ib Nad
gal
0.049
0.10
0.17
0.24
0-73
none
0.049
0.10
0.17
0.24
none
0.049
0.10
0.17
0.24
0.73
Regenerant Required
for 95% NH3-N Elution
Ib NaCla
cu ft
13.6
18.5
24.6
26.9
81.9
0
8.1
13.9
18.6
34.8
0
5.1
7.2
12.3
15.8
48.0
Ib Ca(OH)|'b
cu ft
0.38
0.26
0.20
0.16
0.16
c
0.77
0.63
0.50
0.66
d
1.06
0.72
0.72
0.66
0.66
Calculations shown are for 6.6 eq NH3-N/cu ft.
Calculated using the amount of caustic required
to raise the pH of tap water used in this study.
These amounts were 0,0014, 0.046, and 0.010 Ib
Ca(OH)2/gal for pH 11.5, 12.0, and 12.5 regenerants,
respectively.
Regeneration, for these conditions must be made
using NaOH to provide Na in the regenerant. NaOH
required was 1.06 Ib/cu ft.
Regeneration must be made using NaOH. NaOH
required was 1.11 Ib/cu ft.'
93
-------
Efficiency
Although the percent of ammonia eluted with a particular volume of
regenerant is a good indication of the volume of regenerant needed and
the time required for regeneration, it does not directly indicate the
efficiency with which the regenerant is utilized. In this report
regeneration efficiency has been defined as follows:
. . Equivalents NHs-N Removed During Previous Exhaustion ]QQ
tTTiciency, h - Equivalents Na Used for Regeneration
This has also been aptly termed the "regenerant utilization factor" by
Klein [85]. Efficiencies were calculated using the volume of regenerant
required for 95% elution of ammonia from the zeolite. Data used in
making these calculations are shown in Table 12. The efficiency for run
63 could not be calculated, as mentioned previously, because of incom-
plete elution of ammonia from the zeolite. In this case the volume of
regenerant required for 95% ammonia elution was obtained by extrapolating
data for this run shown in Figure 22. This was done on the basis of the
relatively constant ammonia concentration of the spent regenerant.
Efficiencies for runs 64 and 67 also could not be calculated because of
incomplete elution of ammonia. However, for these runs, no assumption
could be made concerning the volume of regenerant required to complete
regeneration.
Efficiencies are shown as a function of regenerant composition in
Figure 28. Efficiencies for pH 12.5 were approximately twice as high
as for pH 11.5. Points for regeneration at pH 12.0 generally lay
between those for pH 11.5 and pH 12.5. The efficiency for run 63 using
0.10 Ib NaCl/gal and an unaltered pH of 8.2 is also shown in Figure 28
for comparison with runs made at higher pH values. Efficiencies achieved
in these experiments were considerably lower than those usually experi-
enced in water softening. This was expected, based on the poorer
exchange kinetics characteristic of clinoptilolite; however, the effi-
ciency calculated here does not reflect the elution of ions other than
ammonia during regeneration. Efficiencies would be correspondingly
higher had these ions been taken into account. In cases where regenerant
is stripped of ammonia and reused, the effective efficiency will approach
100%, as the only sodium ions consumptively used in any cycle will be
the stoichiometric amount needed to replace ions eluted from the zeolite
and the quantity in the regenerant solution lost in transfer between
storage tanks and the ion exchange beds.
An additional consideration when regenerant is reused is the change of
composition of the regenerant solution. In addition to salts added to
make up the regenerant, ions eluted from the clinoptilolite will accu-
mulate in the regenerant solution. In this respect the composition of
the regenerant will differ from that used in these tests. For domestic
94
-------
1.0
Efficiencies Calculated for
Ammonia Elution of 95%
V Regenerant also contained
0.046 Ib CaCl2/gal; pH 12.2
4Regenerant also contained
0.32 Ib CaCl2/gal; pH 11.8
0.0 0.1
0.7
0.2 0.3 0.4 0.5 0.6
Regenerant Strength, Ib NaCl/gal
FIGURE 28. EFFECT OF REGENERANT COMPOSITION ON EFFICIENCY
0.8
95
LIBRARY U.S. ERA
-------
wastewaters calcium will be the most significant ion to accumulate in
the regenerant. To determine the effect of added calcium in the regene-
rant solution, regeneration was carried out in two runs using 0.10 lb NaCl/
gal saturated with lime to which CaCl2 was added. For run 60, 0.046 Ib
CaCl2/gal (0.05 M CaCl2) was added to the regenerant and for run 65,
0.32 Ib CaCl2/gal (0.35 M CaCl2) was added. The resulting pH of these
solutions was 12.2 for run 60 and 11.8 for run 65. Regeneration effi-
ciencies for these runs were 8.6 and 5.6%, respectively. These values,
shown in Figure 28, are close to the efficiencies for runs 38 and 40
in which the regenerant contained 0.10 Ib NaCl/gal at pH 12.0. Thus
the results of these tests were not significantly affected by the presence
of added calcium in the regenerant solution. Regeneration costs based
on the results presented in this chapter are considered in Chapter X.
ATTRITION OF CLINOPTILOLITE IN CAUSTIC SOLUTIONS
Results of the regenerant optimization study showed that clinoptilolite
can be regenerated more effectively at a pH of 12.5 than at 11.5,
indicating that the feasibility of using more concentrated caustic
solutions for regeneration should be examined. However, previous work
by Barrer e_t a_l_. [50] demonstrated that the zeolite framework was not
stable in caustic solutions. Therefore, tests were undertaken to
determine the stability of the clinoptilolite used in this study in
caustic solutions.
Initial tests were made in a batch system. In prder to simulate operating
conditions in which the clinoptilolite is exposed to caustic regeneration
and exhaustion with a solution of near neutral. pH, clinoptilolite samples
were alternately exposed to NaOH and d-istilled,water~ \ One-gram samples
of previously unused 20 x 50 mesh clinoptilolite were added to 12 flasks.
Four flasks were used as controls and exposed only to distilled water.
The remaining eight flasks were cyclically exposed to 100 ma of 2%
NaOH (pH of about 13.3) and distilled water. Samples were shaken in
125-nu erlenmeyer flasks using a Burrell Wrist-Action Shaker (Burrell
Corp., Pittsburgh, Pa.). During each cycle the clinoptilolite was
exposed to NaOH for 2 hr and rinsed with distilled water for 2 hr. When
changing solutions, water was also changed in the control flasks, so the
decanting error would be the same for all flasks. After 5, 7, 25, and
50 cycles, one control and two test flasks were removed for analysis.
Clinoptilolite samples were dried overnight at 105°C and sieved through
20 and 50 mesh sieves. The amount of 20 x 50 mesh material lost is
shown in Table 14. These results show that clinoptilolite was signifi-
cantly attacked by the caustic solution. After exposure to 50 simulated
regeneration cycles, the attrition rate of clinoptilolite was 1.6%/cycle
considering the total weight loss of 20 x 50 mesh material and 0.8%/cycle
when the amount lost in control flasks was subtracted. The latter value
is the loss attributable to the high pH alone assuming that mechanical
attrition in caustic and control samples was equal. However, the total
loss of 20 x 50 mesh material is a better indication of losses which
would occur in actual column operation.
96
-------
TABLE 14
WEIGHT LOSS OF CLINOPTILOLITE EXPOSED TO
2% NaOH (pH 13.3)
Cycles
5
7
25
50
Caustic
Wt. Lossa'b
20x50 Mesh
Material ,g
0.304
0.327
0.506
0.789
Wt. Loss
20x50 Mesh
Material ,%
30.4
32.7
50.6
78.9
Control
Wt. Loss9
20x50 Mesh
Material ,g
0.305
0.312
0.350
0.370
Wt. Loss
20x50 Mesh
Material ,%
30.5
31.2
35.0
37.0
Weight Loss
B/Cycle
Caustic-
Control
0
0.21
0.62
0.84
Caustic
6.1
4.7
2.0
1.6
Initial weight for all samples was 1.000 g.
Weights reported are average of two samples; maximum variation
between two samples of any set was !5%.
Because of the high rate of attrition of clinoptilolite in the presence
of 2% NaOH, further attrition studies were made using small laboratory
columns containing 8.000 or 10.000 g of clinoptilolite. These columns
consisted of 1-cm ID by 21-cm long glass tubes connected to caustic and
rinse solutions using 1/16-in. rubber tubing. Three columns were
connected in parallel to permit testing of three samples for each test
condition. Three columns were connected only to the distilled water
rinse for use as controls. Caustic was fed to the columns from 20-i
glass bottles which were refilled daily. Distilled water for rinsing
was siphoned from a 20-£ bottle which was kept full by continuously
dripping distilled water into the bottle from a distilled water line.
Feed to the columns was alternated between caustic and rinse solutions
at 2-hr intervals using an Intermatic Model T1905 24-hr timer connected
to two 1/8-in. solenoid valves (Hoke model S90A180C, Hoke, Inc., Cresskill,
N. J.). Both valves were the normally closed type and were connected
to be open in alternate 2-hr periods. An illustration of these columns
appears in Figure 29. Flow through the columns was set at approxi-
mately 4 m£/min (17 BV/hr) using screw clamps located at the influent
end of the columns. Columns were operated upflow to minimize plugging.
Effluent from the columns was collected in 8-£ bottles and weighed each
day to determine the volume of liquid passed through each column. The
pH of the effluent from each, column was also measured to insure that
approximately'equal volumes of caustic passed through each column.
These bottles also served to catch any clinoptilolite which might be
washed from the columns if an increase of flow through the column occurred.
97
-------
oo
24-hr
Timer
Caustic
Caustic
a
1 cm ID x 21 cm
glass columns
Solenoid
Valves
'Distilled Water
Waste
Distilled Water
Solenoid
Valves
Screw
Clamps
To Collection
Bottles
To Collection
Bottles
-X-
To Collection
Bottles
FIGURE 29. LABORATORY COLUMNS USED IN ATTRITION STUDY
-------
Columns were inverted every two or three days to prevent the zeolite
from becoming packed and to minimize channeling of flow through the bed.
Results of clinoptilolite exposed to 100 simulated regeneration cycles
at pH 11.5, 12.0, and 12.5 are given in Table 15. The effect of caustic
increased significantly from pH 11.5 to 12.5 as judged by the loss of
zeolite weight/regeneration cycle. Over 100 cycles the total attrition
rate at pH 12.5 was 0.57%/cycle compared to 0.34 and 0.24%/cycle for
exposure to pH 12.0 and 11.5 solutions, respectively. These are signi-
ficantly less than the 1.6%/cycTe measured in batch tests reported in
Table 14. However, the rate of mechanical attrition was also signifi-
cantly higher in.the batch tests. Attrition rates were 0.15, 0.25, and
0.48%/cycle for-exposure to pH 11.5, 12.0, and 12.5 solutions, respec-
tively, when the amount lost in control flasks was subtracted.
Results of tests made to determine the variation of attrition rate with
increasing exposure to caustic are also shown in Table 15. Tests were
50, 100, and 150 cycles in duration. Total attrition rates were 0.56,
0.57, and 0.51%/cycle, resoectively, for 50, 100, and 150 cycles of
exposure. These results indicate that the attrition rate might become
less after exposure to 100 cycles of caustic. This observation is
supported by weight losses of 28.3% from 0 to 50 cycles, 29.1% from 50
to 100 cycles, but only 18.6% from 100 to 150 cycles. Part of the
observed decrease in weight loss from 100 to 150 cycles could be due
to the presence of impurities in the clinoptilolite which constitute
an increasingly larger fraction of the weight left in the column as
clinoptilolite is broken into smaller pieces (or dissolved) and washed
from the column. Barrer [50] reported that a sample of clinoptilolite
Contained 10 to 15% impurities. However, no,attempt was made to measure
impurities present j|n_the zepIfte used in this study. It is also possible
that any harder and" moire "res fs" tan t clinoptilolite that might accumulate
in the zeolite bed.-wi.LL.he. Less desirable as an ion exchanger because
of a lower exchange capacity or poorer exchange kinetics. On the basis
of these results, it cannot be assumed that a lower attrition rate after
100 cycles of exposure to caustic will result in a reduced replacement
rate of clinoptilolite.
These results indicate that zeolite replacement costs will constitute a
significant factor in the cost of ammonia removal using clinoptilolite.
Attrition rates will be approximately 0.25, 0.35, and 0.55%/operating
cycle using regenerants having a pH of 11.5, 12.0, and 12.5, respec-
tively. The attrition rate for exposure to 2% NaOH (pH about 13.3) was
1.6%/cycle measured in a batch system in which samples were constantly
shaken. These values are probably generous estimates of the attrition
rate which would be experienced in full-scale operations because the
exposure times/cycle used in these experiments were longer than the
time required for an actual regeneration cycle. As such, these values
constitute liberal estimates of zeolite attrition for use in the
determination of the least cost regeneration procedure considered in
Chapter IX.
99
-------
TABLE 15
WEIGHT LOSS OF CLINOPTILOLITE IN COLUMN ATTRITION STUDIES
Q
O
pH
11.5
12.0
12.5
12.5
12.5
Cycles
100
100
50
100
150
Caustic
Wt. Loss3
20x50 Mesh
Material ,g
2.436b
2.698C
2.261C
5.738b
6.071c'd
Wt. Loss
20x50 Mesh
Material ,%
24.4
33.8
28.3
57.4
76.0
Control
Wt. Loss3
20x50 Mesh
Material ,g
0.904b
0.904b
0.938C
0.904b
0.716C
Wt. Loss
20x50 Mesh
Material ,%
9.0
9.0
11.7
9.0
9.0
Weight Loss
%/ Cycle
Caustic-
Control
0.15
0.25
0.34
0.48
0.45
Caustic
0.24
0.34
0.56
0.57
0.51
Weights reported are average of three samples; maximum variation among samples of any set
was 11%, excent as noted.
Initial sample weight was 10.000 g.
clnitial sample weight was 8.000 g.
Weight of samples in this set varied by 26%.
-------
RINSE REQUIREMENTS
Removal of salt and.caustic from a freshly regenerated zeolite bed is
necessary before placing the bed back into operation. In this case the
pH of the bed must be reduced before efficient ammonia removal may be
achieved. The effects of pH on ammonia exchange described in Chapter
VII indicated that the pH of g zeolite bed.must be reduced to about 8
in order to permit ammonia removal during exhaustion. However, a reduc-
tion in the waste rinse water volume might be achieved by rinsing to a
pH of 9.5 or 10.0 with some sacrifice of performance during the first
part of the exhaustion cycle. Alternatively, rinsing could be divided
into two steps, with water used for the first step being recycled to
upstream processes and that used for final rinsing being mixed with
product water as long as rinsing is accomplished using ammonia-free
water. The high pH of the first .fraction might be beneficial in
upstream coagulation processes, while the second fraction should not
impair the overall quality of the product water.
In evaluating rinse requirements it was assumed that only the volume of
water necessary to reduce the pH to 1Q would need to be wasted or
recycled for additional treatment. Comparisons of SERL tap water and
product water used for rinsing are shown in Figure 30. It is evident
that the product water, which had a higher buffer capacity, was much
more effective in reducing the column pH". The tap water had an initial
pH of 8.5 and required 5.0 meq/£ of NaOH to raise the pH to 11.5, while
product water had ah initial pH of 7.8 and required 10.0 meq/£ NaOH to
raise the pH to 11.5. While it is unlikely that tap water will be used
for rinsing, these results'dd indicate that rinse requirements will
vary considerably depending DM the,-quality of the rinse water.
Product water was used both alone and in combination with a small volume
of acid applied to the column prior to rinsing-. Rinse volumes required
for reduction of pH to 10.0 are shown in Table 16. The application of
TABLE 16
COMPARISON OF COLUMN RINSING PROCEDURES
Rinse
Tap water3
Product water
Product water preceded
by 1 BV 0.01 M-HC1
Product water preceded
by 2.5 BV 0.01 M HC1
Rinse Volume
to pH 10.0
BV
35
11
8
4
aFor analysis of tap water, see Chapter VI,
101
-------
12.0
11.0
o
no
10.0
9.0
A Tap Water
oe Product Water
D 1 BV 0.01 M HC1; product water
O2. 5 BV 0.01 M HC1 ; oroduct water
Rinse Rate = 11 gal/so ft-min
(uoflovv)
_L
10 20 30 40 50
Rinse Volume, BV
FIGURE 30. COMPARISON OF COLUMN RINSE REQUIREMENTS
60
70
-------
acid to the column prior to the product water rinse further reduced the
rinse requirement. However, the use of acid would probably not be
economically desirable in view of the small rinse volume required using
product water by itself.
SUMMARY
In this chapter factors relating to the regeneration of clinoptilolite
have been investigated. Columns used in the regeneration study were
exhausted to an ammonia breakthrough of approximately 1 mg/£ NH3-N
using a chemically treated wastewater as the column influent. In this
way all columns were loaded with approximately the same amount of ammonia
prior to each regeneration. Regenerant solutions were composed of various
concentrations of NaCl and NaOH. Regenerations made using a variety of
flow rates showed that regeneration performance was not affected signi-
ficantly by a variation of flow rates from 4 to 20 BV/hr. However, a
significant loss of efficiency was observed using flow rates of 30 BV/
hr or higher.
Tests made to determine the effect of regenerant composition on regenera-
tion performance showed that regeneration efficiency increased with
increasing regenerant pH. It was hypothesized that a higher diffusion
rate of ammonia was possible at higher pH values because of the presence
of unionized ammonia. Efficiencies were calculated for the different
regenerant compositions and showed that the highest efficiencies were
obtained using no NaCl and a pH adjusted to either 12.0 or 12.5. Effi-
ciencies for these conditions were 55% at pH 12.0 and 52% at pH 12.5.
Based on the results of these tests, the amounts of regenerant required
for 95% elution of ammonia in the zeolite was calculated. Chemical
weights expressed as Ib/cu ft, are shown in Table 13; the volume of
regenerant required, expressed as gal/eq NH3-N removed, is shown in
Figure 27.
Studies made to determine the attrition of clinoptilolite in the presence
of caustic solutions demonstrated that zeolite replacement must be
considered in estimating the cost of ammonia removal. Attrition rates
of 0.25, 0.35, and 0.55%/cycle were measured for exposure to pH 11.5,
12.0, and 12.5 solutions, respectively, in small columns. A loss of
1.6%/cycle was measured for exposure to a pH 13.3 solution using a
batch method in which samples were shaken constantly.
The rinse requirement was investigated using a tap water, product water,
and product water preceded by various amounts of 0.01 M HC1. While the
use of acid reduced the rinse requirement, this reduction was insuffi-
cient to justify the added expense which would be incurred in purchasing
and storing acid. It was recommended that ammonia-free product water be
used for rinsing with the volume being required to reduce the column pH
to 10.0 returned to upstream treatment processes and the volume required
to further reduce the pH to 8.0 be returned to product water storage.
103
-------
IX. PROCESS PERFORMANCE
GENERAL CONSIDERATIONS
In order to provide data on column performance under operating conditions,
clinoptilolite columns were operated using sewage at SERL, the East Bay
Municipal Utility District (EBMUD), and the Central Contra Costa Sanitary
District (CCCSD). The objectives of these studies were: 1) to deter-
mine ammonia removals which could be expected in cyclic operation of
clinoptilolite, 2) to ascertain the dependence of column performance
upon various operational variables, and 3) to compare the performance
of clinoptilolite using chemically different sewages. To meet these
objectives, runs were made under a variety of conditions. Consequently,
the conduct of the runs at each location and data particular to each
test series are discussed with respect to the test location. Indivi-
dual tests were designed so that runs continued to the approximate time
of ammonia breakthrough. Flow rates were usually adjusted so that
ammonia breakthrough would occur in 12 or 24 hours in order that columns
could be attended to on a regular basis. Because of the necessity of
providing a water which was relatively free of suspended solids for ion
exchange processing, clinoptilolite columns were preceded by various
upstream treatment processes which, in all cases, included chemical
precipitation using lime. Details of the treatment system employed
at each test location are discussed in Chapter VI. In addition, treat-
ment schemes are mentioned briefly in the description of each test
location in this chapter.
STUDIES AT THE SERL PILOT PLANT FACILITY
Process Operation
Tests made at the SERL pilot plant were more numerous than those at
other locations because of the proximity of the facility to the labora-
tory. Consequently, these tests covered a longer period of time and
were designed to furnish information concerning more aspects of column
operation than were studies at other locations. For purposes of discus-
sing these data, runs made have been divided into two parts: Series I
runs made during the summer and early fall which included tests related
to the dependence of column performance on operating variables such as
previous regeneration and flow rate; and Series II runs made under slightly
different conditions during the winter. Various combinations of upstream
treatment processes were used in these studies. In Series I runs,
previous treatment of the column influent included primary sedimentation,
105
-------
activated sludge treatment, and chemical precipitation at pH 11.0;
primary sedimentation, activated sludge treatment, and chemical preci-
pitation at pH 9.5; primary sedimentation followed by chemical preci-
pitation at pH 11.0; or primary sedimentation followed by chemical
precipitation at pH 9.5. In all cases lime was used as the precipitant
in the chemical treatment step. Chemically precipitated effluents were
neutralized by recarbonation, then pumped to the column unit. In
Series II runs previous treatment of the column influent included primary
sedimentation, chemical precipitation at pH 11.0 using lime, and recar-
bonation. Supplementary data concerning pilot-plant operation appears
in Appendix E. Cation compositions of the primary effluent for Series
I and II operation are summarized in Table 17.
TABLE 17
CATION COMPOSITION OF PRIMARY EFFLUENT
DURING SERL PILOT-PLANT RUNSa
Series
I
II
NH3-N
mg/£
20.9
21.3
Na
mg/£
54. 3b
56.3
K
mg/£
9.3b
8.5C
Ca
mg/£
47
44
Mg
mg/a
7
8
pH
7.4
7.3
Values are average of daily composite samples
taken during test period.
Values from analysis of chemically precipitated
effluents.
cSample filtered prior to analysis.
Corresponding average values for column influent and effluent are shown
in Table 18. Values in these tables are averages of daily composite
samples taken during the test periods.
Wastewater Characterization. In addition to daily analyses made to
monitor overall plant performance, tests were run to determine the
nature of the primary effluent of the SERL treatment facility. Raw
sewage was pumped to the pilot plant from a City of Richmond sewer and
consisted primarily of sewage from a residential area. Hourly grab
samples were taken from the effluent launder of the primary sedimen-
tation basin for a 33-hr period. The results of these analyses are
shown in Figure 1 and Table 1 of Appendix D.
Column Operation. All runs began with the columns regenerated with NaCl
and NaOH or NaCl and Ca(OH)2. The exhaustion phase continued for 12 hr/
106
-------
TABLE 18
CATION COMPOSITION OF COLUMN INFLUENT AND EFFLUENT
DURING SERL PILOT PLANT RUNS9
Runsb
1- 8
9-18
19-36
38-46
47-66
Series
I
I
I
I
II
Prior Treatment
Activated sludge;
Pptn. at pH 11.0
Activated sludge;
Pptn. at pH 9.5
Primary effluent;
Pptn. at pH 11.0
Primary effluent;
Pptn. at pH 9.5
Primary effluent;
Pptn. at pH 11.0
Sample
Influent
Effluent
Influent
Effluent
Influent
Effluent
Influent
Effluent
Influent
Effluent
NH3-N
mg/£
14.9
0.67
15.6
0.25
18.8
0.30
19.4
0.34
20.2
1.86
Na
mg/£
57.5
102
58.6
108
52.3
112
52.7
106
54.0
105
K
mg/SL
9.6
2.2
9.4
1.6
9.6
1.9
8.5
0.9
9.3
1.8
Ca
mg/£
86
67
57
52
75
56
51
42
74
57
Mg
mg/£
6
6
13
14
2
4
5
6
4
5
c
PH
7.8
8.0
8.2
8.4
7.7
7.8
7.4
7.6
7.3
7.5
Values are average of daily composite samples taken during each test period.
'individual run data given in Table 19.
'Column influent pH adjusted by recarbonation.
-------
run unless otherwise stated. With but one or two exceptions, flow rates
of 7.5, 10, or 15 BV/hr were used. The bed depth in all cases was 3 ft.
Results of runs made using a synthetic sewage similar to chemically
treated SERL primary effluent reported in Chapter VII indicated that
columns could be operated for 12 hr at 15 BV/hr without exceeding the
ammonia breakthrough capacity. An influent sample for ammonia analysis
was composited over the duration of each run. Four 3-hr composites of
column effluents were taken during each run.
Column Performance Series I
Ammonia removal data and other information relating to individual column
operation are presented in Table 19. The average ammonia removal for
these runs was 97.8% with an average influent NH3-N concentration of
17.6 mg/£ and an average effluent concentration of 0.39 mg/£. The
influent varied from 13.0 to 22.2 mg/n NH3-N and the effluent varied
from 0.02 to 1.90 mg/£ NH3-N, the latter value being measured during the
second run and was probably more the result of inexperience in column
operation than of poor column performance. Effluent concentration
averages were greater than 1 mg/a NH3-N for only three runs out of the
46 made in this part of the study. All of these occurred following
regeneration using a salt concentration of 0.049 Ib NaCl/gal at pH 11.5
(11.0 Ib NaCl/cu ft, 0.34 Ib NaOH/cu ft).
Because the column influent had not been previously filtered or treated
for removal of soluble organics except for partial removals achieved
when upstream treatment included the activated sludge process, attention
was given to evidence of fouling. Here fouling is defined as a loss of
ammonia capacity per unit of zeolite as opposed to physical losses in
weight of material from the ion exchange bed. If fouling occurs it would
result in increasing ammonia concentrations in the column effluent.
Figure 31 shows the effluent ammonia concentrations for these runs
plotted with respect to run number. Increasing run numbers correspond
to increasing use of a particular column. Figure 31 indicates that if
there is any noticeable trend, it is toward a decreasing ammonia leakage,
and thus toward improved column performance. This is partially explained
by the fact that columns were regenerated at a pH of 11.5 during the
first runs and at a higher pH during later runs. It is also possible
that through cyclic use additional pores in the zeolite were opened and
became available for ion exchange.
Statistical analysis of these data using Student's t-test revealed that
the effluent ammonia concentration can be expected to be equal to or
less than 0.97 mg/£ NH3-N 95% of the time. This value is indicative of
ammonia removals which can be achieved when operating columns to incipient
ammonia breakthrough. For full-scale operation where an excess of
regenerant is not applied to the zeolite, higher values might result if
adequate process control is not maintained. In order to closely control
the ammonia concentration of the product water, it would be desirable to
108
-------
TABLE 19
SERIES I OPERATIONAL DATA FOR STUDIES AT SERL PILOT PLANT
SERL
Run
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
38
39
40
41
42
43
44
45
46
Flow3
BV/hr
1
15
15
15
10
15
10
15
7.5b
7.5
10
10
15
10
15
12
15
9
15
10
15
15
10
15
10
15
15
10
15
10
15
10
15
10
15
7.5b
15
10
15'
10
15
15
15
15
15
Col.
No.
2
3
2
1
2
3
2
1
3
2
1
2
3.
2
1
2
3
2
2
1
2
3
2
1
2
2
3
2
1
2
3
2
1
2
3
2
1
2
3
2
1
2
1
2
NH3-N, mg/i
Inf.
15.1
15.2
15.2
15.2
15.2
13.6
13.6
15.6
15.2
15.2
15.7
15.7
15.4
15.4
13.0
13.0
18.6
18.6
18.6
18.9
18.9
16.9
16.9
21.0
21.0
17.5
18.8
18.8
17.9
17.9
20.0
20.0
19.5
19.5
19.4
16.6
20.0
20.0
22.0
19.5
22.2
18.1
18.0
19.0
Eff.
0.77
1.90
1.10
0.12
0.82
0.87
0.45
0.06
0.02
0.72
0.02
0.56
0.12
0.42
0.23
0.02
0.28
1.09
0.28
0.67
0.29
0.11
0.27
0.22
0.43
0.20
0.13
0.25
0.19
0.11
0.23
0.49
0.18
0.31
0.62
0.28
0.17
0.30
' 0.44
0.31
0.15
0.18
0.27
0.47
NH3-N
Removed
%
94.9
87.2
92.7
99.3
94.6
93.4
97.0
99.4
99.9
95.4
99.9
96.2
99.4
97.4
98.5
99.8
98.4
94.1
98.4
96.3
98.4
99.4
98.2
99.0
98.1
98.9
99.5
98.9
98.9
99.4
99.0
97.5
99.0
98.5
96.9
98.2
99.0
98.5
98.2
98.5
99.5
98.9
98.3
97.4
Previous Regeneration
PH
11.2
11.5
11.3
11.5
11.5
11.5
11.5
11.5
11.6
11.5
11.5
11.5
11.5
11.5
11.5
11.5
11.5
11.5
11.5
11.5
11.5
11.5
11.5
11.5
11.5
11.5
11.5
11.5
11.5
12.3
12.3
12.4
12.3
12.4
12.4
12.4
12.5
12.0
12.0
12.0
12.0
11.5
11.4
12.5
Nad
Ib/gal
0.049
0.049
0.049
0.049
0.049
0.049
0.049
0.049
0.049
0.049
0.049
0.049
0.049
0.049
0.049
0.049
0.049
0.049
0.10
0.10
0.10
0.10
0.10
0.10
0.17
0.24
0.17
0.24
0.24
0.17
0.17
0.10
0.10
0.04,9
0.049
0.24
0.24
0.10
0.10
0.17
0.17
0.049
0.17
0.17
1b NaOH
cu ft
0.34
0.34
0.34
0.34
0.34
0.34
0.34
0.34
0.34
0.34
0.34
0.34
0.34
0.34
0.34
0.34
0.34
0.34
0.34
0.34
0.34
0.34
0.34
0.34
0.34
0.34
0.34
0.34
0.34
2.48
2.48
2.48
2.48
2.48
2.48
2.48
2.48
1.02
1.02
1.02
2.48
1.02
1.02
2.48
Ib NaCl
cu ft
11.0
11.0
11.0
11.0
11.0
11.0
11.0
11.0
11.0
11.0
11.0
11.0
11.0
11.0
11.0
11.0
11.0
11.0
22.5
22.5
22.5
22.5
22.5
22.5
38.2
54.0
38.2
54.0
54.0
38.2
38.2
22.5
22.5
11.0
11.0
54.0
54.0
22.5
22.5
22.5
38.2
11.0
38.2
38.2
Prior
Treatment
c
4-> O
-0 CL
CU Q-'
4-> .«i
to OJ
> cnn
i- -0 0.
+J 13
Or +->
01 JO
a
ZJ C
i O
tO -r-
4->
T3 rd
cu )->
4-J -r-
rd CLLO
> -i-
r- U Ol
-M
ra
4J
EL
U
Of
L-
0.0
>1^
s~
ra n=
E n.
E +->
Q_ rd
C
O
+J
rd Ln
.«4_> .
>,.,- 01
i- 0-
rd -i- :r:
E u Q-
i CL)
S- S- +J
Q- CL <0
aAll runs were 12 hr in length except as noted; bRuns 24 hr long; cRegenerant volume was
30 BV for all runs.
109
-------
2.0
0
1.5
1 '
D 7.5 BV/hr
A 10 BV/hr
0 15 BV/hr
CTi
E
(U
CT:
"3
O)
1.0
- Leakage not exceeded 95%
°f time
~~7
0.5
0.0
i D IA in
i
1
10
FIGURE 31.
20 30
SERL Run Number
40
ENVELOPE OF AMMONIA LEAKAGE IN TESTS
AT SERL PILOT PLANT
50
-------
have automatic controls capable of rotating zeolite beds from the
exhaustion to regeneration modes whenever a preset ammonia concentration
in the product water is exceeded. It is believed control could be
achieved using a residual chlorine analyzer. Increasing ammonia concen-
trations in the column effluent would result in an increased chlorine
demand which, when sensed by the chlorine analyzer, would initiate
rotation of ion exchange beds from the service to regeneration mode and
place freshly regenerated beds into the service mode. Inasmuch as
chlorination is likely to be required as a final treatment step for
water to be reused, this feature could be incorporated into plant design
with little additional expense.
Correlation of Column Performance to Operating Variables. Because several
operating variables were changed during the course of the study, it is
difficult to isolate the precise effect of flow rate and conditions of
the previous regeneration on column performance. However, the runs in
Series I were made under relatively uniform conditions judged by the
column influent and effluent compositions. From the wide range of
operating parameters used in these runs, it was evident that some had
a more pronounced effect on column performance than others.
Ion exchange theory predicts that leakage through a column will become
greater as the flow rate is increased. To determine the effect of flow
rate on ammonia leakage, effluent ammonia concentration for the first
90 BV throughput were averaged for each flow rate. It was necessary
to use a throughput of 90 BV because this was the volume processed at
the lowest column flow (7.5 BV/hr) during the standard 12-hr runs. These
results, presented in Table 20, show that effluent ammonia concentrations
were nearly identical for each flow. These results also show that
effluent ammonia concentrations as low as 0.22 to 0.26 mg/a NH3-N can be
consistently produced at these flow rates for a throughput of 90 BV.
These values compare to an average effluent concentration of 0.39 mg/£
NH3-N for the entire length of these runs shown in Table 19. The 90 BV
throughput for which these values were averaged was well short of the
TABLE 20
EFFECT OF COLUMN FLOW RATE ON AMMONIA LEAKAGE
Column Flow Rate
BV/hr
7.5
9
10
12
15
Runs Made at
This Flow
3
1
13
1
26
Average Leakage to
90 BV, mg/£ NH3-N
0.22
0.26
0.20
0.06
0.22
111
-------
ammonia breakthrough in each case, and indicates that exchange kinetics
did not change significantly for this range of flow rates. However, the
leakage would increase nt higher flow rates as the exchange kinetics
must become less favorable and the resulting breakthrough curves flatter.
Another factor which affects column performance is the nature of the
previous regeneration. Figure 32 shows the variation of ammonia leakage
with regeneration level. Values are shown only for runs made at a flow
rate of 15 BV/hr and are effluent averages for the entire length of run
(180 BV throughput). Regeneration levels were calculated for 30 BV of
regenerant used in each run. Thus, for these runs a regenerant salt
concentration of 0.10 Ib NaCl/gal corresponded to a regeneration level
of 22.5 Ib NaCl/cu ft of zeolite. Salt concentrations for other regene-
rant levels may be calculated proportionally from this value. Figure
33 illustrates the effect of regenerant pH on ammonia leakage. Again,
all points are effluent averages for 180 BV throughput at a flow of
15 BV/hr. The relatively high leakage for 11.0 Ib NaCl/cu ft in Figure
32 and for pH 11.5 in Figure 33 reflect high effluent values for runs
2, 3, and 4 (cf. Table 19). The high effluent ammonia value for
regeneration using 22.5 Ib NaCl/cu ft at pH 12.5 appears to be inconsis-
tent with other values in Figures 32 and 33. Because only one regenera-
tion was made under these conditions (run 33), it is quite likely that
this value is not entirely representative of the performance which can
be expected for operation under these conditions.
Both Figures 32 and 33 illustrate that a small improvement in perform-
ance may be achieved by using higher regeneration levels. However, this
improvement is small and probably would not be an important consideration
in process design. These data indicate that while stable column perform-
ance is possible at any of these regeneration levels, the clinoptilolite
is more thoroughly regenerated using higher regeneration levels.
Column Performance Series II
Series I and II runs at SERL were made under approximately the same
conditions except that Series II runs were made in the winter, while
Series I runs were made during the summer and early fall. Prior treat-
ment of the column influent for Series II runs included primary sedi-
mentation, chemical precipitation at pH 11.0 using lime, and recarbonation.
Data summarizing Series II operation at SERL are shown in Table 21.
Ammonia removal during these runs averaged 91.5% which was noticeably
less than removals achieved in Series I runs. Effluent ammonia concen-
trations averaged 1.7 mg/Ji NH3-N for 180 BV throughput with an average
influent concentration of 20.2 mg/a NH3-N. Effluent values ranged from
0.35 to 5.48 mg/£ NH3-N, while influent concentrations ranged from 16.1
to 23.9 mg/c NH3-N. The high effluent concentration for run 66 was a
result of incomplete ammonia elution during the previous regeneration.
112
-------
1.0
01
o>
en
rd
cu
0.5
Exhaustion Flow: 15 BV/hr
ARegenerant
DRegenerant
ORegenerant
pH
pH
pH
11.5
12.0
12.5
0.0
I
I
10 20 30 40 50
Regeneration Level, Ib NaCl/cu ft Zeolite
FIGURE 32. VARIATION OF AMMONIA LEAKAGE WITH AMOUNT
OF SALT USED IN PREVIOUS REGENERATION
60
1.0
01
O)
cr.
re
^.
to
CL>
0.5
0.0
Exhaustion
NaCl Used:
all.
022,
A38,
Flow: 15 BV/hr
Ib/cu ft
O54.0
_L
11.5 12.0
Regenerant pH
12.5
FIGURE 33. VARIATION OF AMMONIA LEAKAGE WITH
pH OF PREVIOUS REGENERATION
113
-------
TABLE 21
SERIES II OPERATIONAL DATA FOR STUDIES AT SERL PILOT PLANT
SEPL
Run
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
66
Flowa
BV/hr
15
15
15
15
15
15
7.5
7.5
15
15
15
15
15
15
15
15
15
15
20
Col.
Mo.
1
2
1
2
1
2
1
2
2
l"
2
1
2
1
2
2
1
2
1
NH3-N, mg/£
Inf.
21.3
13.4
19.3
16.1
20.0
16.3
20.2
20.8
18.6
20.8
17.8
22.3
18.8
22.0
22.0
23.9
19.9
22.0
22.9
Eff.
90 BV
Avq.
1.25
0.45
0.12
0.07
0.29
0.33
0.31
0.27
-
0.45
0.39
0.43
1.15
0.36
0.36
0.53
0.24
0.25
2.42
Eff.
135 BV
Avg.
3.52
0.56
0.26
0.09
1.37
0.31
0.46
0.28
0.67
0.47
0.78
0.77
1.91
0.54
1.01
1.02
0.51
0.23
3.03
Eff.
180 BV
Avq.
5.48
1.31
1.97
0.35
3.30
0.47
0.98
0.40
0.67
1.07
1.46
2.03
-
1.46
-
-
1.82
0.53
4.06
Previous Regeneration
nH
12.5
12.5
12.5
12.2
12.5
12.5
12.0
12.0
12. 5b
11.5
12. 5b
12.4
12. 5b
12.5
12. 5b
12. 5b
12.5
12. 5b
8.2
NaCl
Ib/gal
0.10
0.10
0.10
0.73
0.10
0.73
0.049
0.24
0.10
0.73
0.10
0.73
0.10
0.00
0.10
0.10
0.10C
0.10
0.10
Volume
Regen.
Used,
BV
30
30
30
30
30
30
30
30
13
30
13
30
13
30
13
13
30
13
30
Ib NaCl
cu ft
22.5
22.5
22.5
164
22.5
164
11.0
54.0
9.75
164
9.75
164
9.75
0
9.75
9.75
22.5
9.75
22.5
Ib NaOH or
Ca(_QH)2
cu ft
2.48
2.48
2.48
2.48
2.48
2.48
1.02
1.02
1.08
0.34
1.08
2.48
1.08
2.48
1.08
1.08
2.48
1.08
0
aThrouqhout for all runs was 180 BV.
' Reqenerant oH raised using lime.
cReaenerant also contained 0.046 Ib CaCl2/gal.
-------
In order to obtain a better estimate of ammonia leakage for Series II
runs, the average effluent ammonia concentration for the first 90 and
135 BV throughput was calculated. The effluent average over 90 BV was
0.40 mg/£ NH3-N and for 135 BV was 0.94 mg/£ NH3-N. Effluent values
for the first 90 BV throughput were identical to the overall effluent
average for Series I runs. However, ammonia leakage for Series II runs
increased significantly for throughputs of 135 BV or more. Ammonia
analyses for Series I runs, including results of grab samples taken at
the end of some runs, showed that the effective ammonia capacity to
1 mg/£ NH3-N breakthrough was equaled or exceeded in many runs. Higher
leakages for Series II runs, made at the same flow and for the same
length of time as Series I runs, indicated that some difference existed
between Series I and II runs which resulted in the effective ammonia
exchange capacity of the zeolite being exceeded to a significantly
greater degree during Series II runs. While the influent ammonia concen-
tration was somewhat higher for Series II runs (20.2 mg/a NH3-N for
Series II compared to values of 14.9, 15.6, 18.8, and 19.4 mg/£ NH3-N
for the four phases of Series I operation), higher effluent ammonia values
were not consistently observed for runs having the highest influent
ammonia concentrations. Examination of Table 18 shows that competing
cation concentrations were very similar for all groups of runs made at
SERL. Although these two groups of runs were made at different times
of the year, seasonal temperature fluctuations are not great in the
San Francisco area, and so these differences would not be expected on
the basis of temperature changes. Thus, no explanation for the poor
performance of Series II runs may be offered at this time. However,
these results do indicate the need to identify more completely the
factors which influence ammonia leakage and the ammonia exchange capa-
city of clinoptilolite.
EAST BAY MUNICIPAL UTILITY DISTRICT STUDY
Process Operation
Studies at the East Bay Municipal Utility District were made utilizing
a 20-gpm pilot plant operated by EBMUD personnel for the preparation of
column influent. This treatment system included primary sedimentation,
chemical coagulation using lime, recarbonation, and filtration prior to
ammonia removal.
Wastewater Characterization. The EBMUD plant treats an average waste-
water flow of 80 mgd (average for period July 1, 1969 to June 30, 1970)
collected from Oakland and surrounding communities. A significant
amount of storm runoff enters the system during wet weather. In addition,
the system receives some infiltration of brackish water because of the
proximity of the service district to San Francisco Bay. The results of
analyses run on 2-hr composites of EBMUD primary effluent are shown in
Figure 2 and Table 2 in Appendix D.
115
-------
Column Operation. Columns used in these studies were regenerated using
10 BV of regenerant having a pH of approximately 12.5 and a salt concen-
tration of 0.17 Ib NaCl/gal. This corresponded to a regeneration level
of 12.7 Ib NaCl/cu ft of zeolite and approximately 0.8 Ib NaOH or
Ca(OH)2/cu ft. As in previous studies, a bed depth of 3 ft was used
for all runs. Initially runs were 18 hr long at a flow of 10 BV/hr.
However, most runs were made for 24-hr periods using flow rates ranging
from 5 to 7.5 BV/hr. Supplementary data concerning pilot-plant operation
and column performance appear in Appendix E. Cation analyses of the
primary effluent and the column influent appear in Table 22. It is
evident that the cation concentration of this sewage was significantly
higher than that at SERL described in Tables 17 and 18. Rain which
TABLE 22
CATION COMPOSITION OF EBMUD WASTEWATER
Sample
Primary effluent3
Primary effluentb>c
Lime precipitated
effluentb>c
Column effluent0
NH3-N
mg/£
19.2
11.6
11.5
0.71
Na
mg/a
183
96d
96
153
K
mg/a
33.4
16. 9d
16.9
2.7
Ca
mg/a
39
39
104
88
Mg
mg/£
11
9
7
4
PH
7.1
7.2
7.4
7.5
Analyses made during a period of no rain. Values are
average of 2-hr samples taken over a 24-hr period.
Analyses made during a period of some rain.
cValues are average of daily composite samples taken during
test period.
Values from analysis of precipitated effluent.
occurred throughout the testing period was responsible for many fluctua-
tions in sewage strength. Evidence of this is apparent in comparing
ammonia concentrations of the two primary effluent values - one which
was taken over a 24-hr period during dry weather prior to the beginning
of column operation and the other averaged from daily composite samples
taken during the testing period when much rain occurred. The large
increase in calcium concentration through the chemical treatment step
was partially due to floe carried over from the clarification basin
and partially redissolved after recarbonation.
116
-------
Column Performance
Column performance data for the EBMUD tests are presented in Table 23.
The average ammonia removal for these runs was 94.0% with an influent
average of 11.5 mg/£ NH3-N and an effluent average of 0.71 mg/£ NH3-N.
Poor performance in runs 1 and 2 was probably a result of high competing
cation concentrations in the column influent, since heavy rains had not
yet begun when these runs were made. The cation concentration of a water
will influence column performance in several ways. First, the effective
ammonia capacity of the zeolite will decrease with increasing concen-
trations of competing cations as discussed in Chanter VII. Second, the
leakage of ammonia preceding the actual breakthrough will be greater
for_waters having higher competing cation concentrations. In addition
an increase in ammonia concentration will result in a decreased through-
put to ammonia breakthrough. While the increased ammonia concentration
will increase the effective ammonia exchange capacity of the zeolite,
this increase will not be sufficient to remove the added ammonia.
Regeneration in these tests was planned to be representative of the
procedure which might be used in full-scale operation. From regeneration
tests reported in Chanter VIII, it was found that 10 BV of regenerant
having a salt strength of 0.17 Ib NaCl/gal and a nH of 12.5 would be
required to regenerate these columns. This corresponded to the use of
12.'7 Ib NaCl/cu ft and approximately 0.8 Ib NaOH or Ca(OH)2/cu ft. This
level of regeneration generally resulted in satisfactory column perform-
ance during these runs. However, beginning with run 7, the effluent
ammonia concentration increased in runs made in Column 1 until the
average effluent concentration was 2.19 mg/£ NH3-N for run 11. In this
case successive runs made at 7.5 BV/hr combined with a low regenerant
pH resulted in a buildup of ammonia in the zeolite. Higher regenerant
pH in subsequent runs resulted in more complete ammonia elution from
the column. Lower effluent ammonia concentrations in later runs were a
result of improved regeneration as well as a reduced throughput in runs
15 and 17.
CENTRAL CONTRA COSTA COUNTY SANITARY DISTRICT STUDY
Process Operation
Column influent for these tests was pretreated by partial primary sedi-
mentation (15 min detention time), chemical coagulation at pH 10.5-10.8
using lime, and recarbonation.
Wastewater Characterization. The Central Contra Costa Sanitary District
plant treats an average flow of 20 mgd consisting primarily of sewage
from domestic and commercial origin. Tests at CCCSD were run during dry
weather, so no stormwater flow was present in the sewage during the
period of study. Results of analyses made on 2-hr composites of CCCSD
primary effluent are shown in Figure 3 and Table 3 in Appendix D.
117
-------
TABLE 23
OPERATIONAL DATA FOR STUDIES AT EBMUD PILOT PLANT
CO
Run
EBMUD- 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Flow
BV/hr
10.0
10.0
4.7
5.7
6.1
6.1
7.3
7.5
7.5
5.1
7.5
7.5
7.8
7.7
5.0
5.1
5.1
10.0
10.0
Column
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
NH3-N, mg/£
Inf.
10.7
14.3
5.3
8.5
8.5
8.8
8.1
14.4
12.6
16.5
15.8
16.1
15.1
13.3
11.2
9.7
12.0
8.3
9.7
Eff.
1.33
3.2
0.41
0.32
0.26
0.31
0.51
0.29
0.83
0.62
2.19
0.19
0.56
0.31
0.35
0.76
0.25
0.35
0.55
NH3-N
Removed
%
87.8
77.6
92.4
96.5
96.5
96.6
93.8
97.9
93.6
96.4
86.1
98.8
96.0
97.7
96.4
91.8
98.3
95.2
93.8
Length
of Run
hr
18
18
24
24
24
24
24
24
24
24
23
23
23
24
22
24
24
24
22
Previous
Regeneration9
PH
12.5
12.5
12.2
12.2
12.2
12.2
12.2
12.4
12.4
12.2
12.2
12.4
12.4
12.4
12.5
12.8
12.8
12.8
12.6
Caustic'3
Lime
Lime
NaOH
NaOH
NaOH
NaOH
NaOH
NaOH
NaOH
NaOH
NaOH
NaOH
NaOH
NaOH
NaOH
NaOH
NaOH
NaOH
Lime
Regeneration with 10 BV regenerant containing 0.17 Ib NaCl/gal which corresponded
to the use of 12.7 Ib NaCI/cu ft.
bCaustic added for oH adjustment was approximately 0.8 Ib NaOH or Ca(OH)2/cu ft.
-------
Column Operation. All runs made in this study were 24 hr in duration at
a flow of approximately 5 BV/hr. The clinoptilolite depth was 3 ft.
Regeneration was made using 20 BV of regenerant containing 0.17 Ib NaCl/
gal at a pH of 12.0. This corresponded to a regeneration level of
25.4 Ib NaCl/cu ft of zeolite and of approximately 0.6 Ib Ca(OH)2/cu ft.
This amount of regenerant was calculated based on the results of regene-
ration tests reported in Chapter VIII. In all runs regenerant solutions
were made using lime as the source of caustic. Data concerning overall
pilot-plant performance appear in Appendix E. The average cation concen-
trations of the waste at various stages of treatment based on daily
composite samples are shown in Table 24.
TABLE 24
CATION COMPOSITION OF CCCSD WASTEWATER9
Sample
Partially settled
sewage
Lime precipitated
effluent
Column effluent
NH3-N
mg/a
21.4
19.5
0.50
Na
mg/a
108
107
135
K
mg/£
11.6
11.5
3.0
Ca
mg/a
58
70
60
Mg
mg/a
9
6
6
pH
7.2
7.1
7.3
Values are average of daily composite samples taken
during test period.
Column Performance
Column performance during the
Poor performance in run 6 was
influent due to an inadequate
sample of column influent for
CCCSD study is summarized in Table 25.
a result of an increase in pH of the column
reserve of C02. The 24-hr composite
this period had a pH of 9.1 (cf. Appendix
E). The four-column effluent samples taken over the period of the run
contained ammonia concentrations of 0.64, 1.23, 13.0, 22.1, and 24.2
mg/a NH3-N compared to an influent concentration of 22.9 mg/£ NH3-N.
This indicates that the pH increased throughout the run resulting in a
progressive deterioration of effluent quality. The concentration of the
last effluent sample exceeded the influent concentration average
indicating that desorption of ammonia may have occurred during part of
the run. This incident illustrates the importance of column influent
pH control in achieving high levels of ammonia removal. In all runs at
SERL this was readily achieved by the feedback control of the lime
feeding equipment and careful surveillance over recarbonation.
119
-------
TABLE 25
COLUMN PERFORMANCE FOR STUDIES AT CCCSDC
CCCSD
Run
1
2
3
4
5
6
Flow5
BV/hr
6.7
5.7
5.0
5.0
5.0
5.0
Column
2
1
2
1
2
1
NH3-N, mg/«,
Inf.
18.5
19.1
20.2
20.1
19.3
22.9
Eff.
1.18
0.52
0.26
0.25
0.27
9.25C
NH3-N
Removed
%
93.5
97.5
98.5
99.0
98.5
59.9
Regeneration in all runs was with 20 BV regenerant
having a salt concentration of 0.17 Ib NaCl/gal and a
pH of 12.0. This corresponded to the use of 25.4 Ib
NaCl/cu ft and approximately 0.6 Ib Ca(OH)2/cu ft.
All runs were 24 hr long.
cEffluent pH averaged over length of run was 9.1.
Disregarding run 6, ammonia removal in this study averaged 97.5% with an
average effluent concentration of 0.50 mg/a NH3-N and an average influent
concentration of 19.5 mg/£ NH3-N. Effluent concentrations ranged from
0.25 to 1.18 mg/a NHs-N, and influent concentrations ranged from 18.5 to
20.2 mg/fc NH3-N. The high effluent concentration observed in run 1
follows the pattern of the tests made at other locations. For the initial
runs made at SERL, high concentrations in the first few runs were origi-
nally attributed to inexperience in column operation. However, the
recurrence of high values in initial tests at other locations implies
that other factors were involved. Since all clinoptilolite used was
dry prior to the beginning of tests in each location, there is a possi-
bility that several days are required for water to completely displace
air initially present in the zeolite pores. Air present in the pore
spaces would effectively block the entrance of cations to exchange sites
and reduce the exchange capacity accordingly.
OVERALL COLUMN PERFORMANCE
Clinoptilolite columns were operated in three different locations using
chemically treated wastewaters. The overall level of performance
achieved in these tests was an ammonia removal of 95.7%. The effluent
ammonia concentration average for all runs except run 6 at CCCSD was
120
-------
0.75 mg/a NH3-N. This average is for the total throughput for all runs
which ranged from 120 to 180 BV except for one run which was 90 BV in
length. Effluent averages for individual locations were 0.39 mg/£ NH3-N
for Series I runs at SERL, 0.71 mg/n NH3-N for runs at EBMUD, and 0.50
mg/a NH3-N for runs at CCCSD excluding CCCSD run 6. An average column
influent pH of 9.1 for run 6 at CCCSD resulted in an ammonia removal of
only 60%, illustrating the importance of pH control in achieving' high
removals of ammonia. Runs made in these studies were designed to end
short of ammonia breakthrough. However, high effluent concentrations
were experienced in Series II runs at SERL. In these runs the effluent
ammonia concentration averaged 1.7 mg/a NH3-N for 180 BV throughput,
but only 0.40 mg/a NH3-N for the first 90 BV and 0.94 mg/a NH3-N for the
first 135 BV indicating that higher ammonia removals could have been
achieved had these runs been stopped sooner. Removals achieved in these
tests are indicative of performance which could be expected from full-
scale plants.
Changes in the cation composition of water treated using clinoptilolite
should be considered because of the effect these changes might have on
potential reuse of the product water. Cation compositions for wastes at
the various test locations are included in Tables 18, 22, and 24.
Inspection of'these values shows that, on an overall average, sodium
increased from 78 to 125 mg/a because regeneration was made using sodium
as the counter ion. Potassium decreased from an average of 11.7 mg/£ in
the column influent to 2.3 mg/a in the effluent. The calcium concentra-
tion of the influent averaged 80 mg/a and decreased to 67 mg/a through
the column. However, practically no decrease was observed in runs at
CCCSD due to the use of lime for regenerant preparation. In tests
where NaOH was used for regenerant preparation, the total hardness was
reduced from 232 to 190 mg/£ as CaC03. This reduction might be bene-
ficial depending on the particular waste treatment objectives. However,
no softening would be realized if lime were used as a regenerant.
Overall influent and effluent magnesium concentrations did not differ
over the period of these tests. While some removal was noticed in runs
at EBMUD, small increases in magnesium through the column at SERL offset
this in the overall average. This was probably due to entrapment of
precipitated Mg(OH)2 particles in the column during regeneration which
subsequently redissolved during exhaustion. This precipitate was
probably a result of fine particles in the regenerant solution which
did not settle before the regenerant was used. This is corroborated by
the small rise in pH observed in all test locations. The rise in pH,
which averaged 0.2 pH units, was probably also due to exchange of hydrogen
into the zeolite and to a shift in the NH3-NH4 equilibrium caused by the
removal of ammonium ions from solution.
121
-------
X. DESIGN CONSIDERATIONS AND COST ANALYSIS
Because the primary objective of engineering design is to find the most
economically feasible solution to a particular problem, it is imperative
that design and economic considerations be discussed together. It is
the purpose of this chapter to aoply the results obtained in the experi-
mental phases of the study to the design of a clinoptilolite ammonia
exchange plant and to estimate treatment costs associated with the
construction and operation of such a plant.
A necessary precondition for the estimation of process costs is the
existence of reasonably accurate data describing process performance
for all conditions to be considered in the cost analysis. Even when
such data are available, much time would be required to analyze all
possible alternative methods of process operation to determine the
least cost method of operation. An additional consideration which may
significantly affect the estimated cost of a unit process is the possi-
bility of complementary operation of several processes in the treatment
system. However, in order to take advantage of savings which might
result from such operation, it is necessary to know the specific processes
to be included in the total design. In order to keep this analysis
generally applicable to as many design situations as possible, a minimum
of assumptions have been made concerning other treatment processes which
might be present in the treatment system. On the other hand, it should
be kept in mind that the ammonia exchange process is most likely to
operate in conjunction with processes which efficiently remove organic
carbon and phosphorus and that these are most likely to include a chemical
precipitation step and probably, but not necessarily, a biological one.
As a further point, the removal of nitrogen by a zeolite exchanger pro-
vides the additional benefit of a high degree of clarification by filtra-
tion. It is likely that objectives of effluent clarity would be desired
in any system where high performance removal of carbon, phosphorus, and
nitrogen was required.
PROCESS DESIGN AND OPERATING PROCEDURE
The design of a clinontilolite ion exchange unit will be illustrated by
considering a 10-mgd waste flow containing an influent ammonia concen-
tration of 20 mg/£ NH3-N. The general properties and operating conditions
for clinoptilolite are summarized in Table 26. These values were the
optimum operating conditions determined during the experimental phases
of the study. Desirable particle sizes and exhaustion flow rates were
determined in previous studies by Mercer [13,14,37]. Suggested bed
depths were based on breakthrough curve characteristics and headless
123
-------
measurements. The gross and net particle specific gravities shown
Table 26 reflect the internal oorositv of the individual c
particles.
net particle specific gravities shown in
porosity of the individual clinoptilolite
TABLE 26
PROPERTIES AND OPERATING CONDITIONS FOR CLINOPTILOLITE
Item
Total exchange capacity9
Mi 11i equi valents/gram
Equivalents/liter
kilograins/cu ft
Physical properties
Gross particle specific gravity
Net particle specific gravity
Bulk density, Ib/cu ft
Proposed operating conditions
Exhaustion flow rate, BV/hr
Bed depth, in.
Regenerant strength, Ib NaCl/gal
Regenerant flow rate, BV/hr
Rinse volume, gal/cu ft
Backwash rinse rate, 50% bed
expansion, gal/sq ft-min
Headless, ft/ft at 5.6 gal/sq ft-min
Value
1.9b
1.4
30.6
1.6
2.4
473
15
36-72 c
d
15
80
11
0.7
Ammonia exchange capacity will usually be less
than this value depending on water composition.
Dry weight.
Greater depths might be used, but allowance must
be made for greater headless.
dSee Figures 27, 34, and 35, and Table 13.
Most literature concerning the design of ion exchance installations has
been based on the use of cylindrical, steel exchange vessels. While
this type of unit is feasible for softening and demoralization of
industrial process and boiler feed water, gravity flow units constructed
from reinforced concrete would be less costly for treating a flow of
10 mgd. Concrete is suitable for use as ion exchange vessels for this
process as it is resistant to both NaCl and lime [86]. The use of this
124
-------
type unit in municipal water softening plants has been reported by Hughes
and Crane [87]. An example of this type of construction is the 200-mgd
F.E. Weymouth Filtration and Softening Plant located at LaVerne, Calif.
[88,89]. Although gravity ion exchange units are mentioned in water
treatment texts, little information is given concerning specific design
criteria [90,91]. However, basic construction details of this type of
structure should be much the same as for rapid sand filters.
Because the cost of ion exchange processing bears a direct relation to
the level .of performance required, the effluent water quality is an
important consideration in estimating process costs. While the treat-
ment objectives concerning nitrogen removal will vary with each specific
case, it is reasonable to assume that at least 90% removal of nitrogen
will be required where the presence of nitrogen impairs water quality.
As the basis of this estimate, the performance objective will be 95%
removal of ammonia assuming the influent concentration to be 20 mg/a
NH3-N. According to the process performance data in Chapter IX, this is
a realistic objective which can be consistently met in cyclic operation.
However, for design purposes complete ammonia removal by the exchanger
will be assumed. This will simplify calculations somewhat and will
result in a more conservative design. For a situation in which less
treatment is required, split treatment could be used to produce a desired
effluent concentration. In this case treatment costs would be propor-
tionately less.
In order to calculate the size of the ion exchange unit needed, the
ammonia capacity of the clinoptilolite must be determined from the
characteristics of the influent water. While this is most accurately
determined in pilot studies, the ammonia capacity of clinoptilolite can
be estimated from Figure 12 if the cationic strength of the was'tewater
is known. Assuming that the influent water has a cationic strength of
0.006 moles/ji, the breakthrough ammonia capacity of .the clinoptilolite
will be approximately-0.25 meq/g for a 3-ft bed; the capacity to satu-
ration will be approximately 0.44 meq/g. A greater effective ammonia
capacity can be realized by increasing the depth of the zeolite bed.
However, increased bed depths result in greater headless and require
that units be built to accommodate the greater head requirement. The
use of a 6-ft bed would result in greater ammonia capacity per unit of
exchanger and while requiring a deeper Structure, the additional cost
would be nominal. Assuming that 3 ft of the zeolite bed wi11 have an
ammonia exchange capacity equal to 0.25 meq/g and that the remaining
3 ft will have a capacity equal to 90% of the saturation capacity or
0.40 meq/g, the 6-ft bed will have an effective capacity of 0.32 meq/g
(equivalent to 6.6 eq/cu ftiand 5.1 kgr/cu ft).
The zeolite volume required to treat a 10-mgd waste flow at 15 BV/hr ;
(1.9 gpm/cu ft) is 3650 cu ft. 'Assuming complete removal of ammonia,
the throughput to ammonia breakthrough is 165 BV and a run length of
11 hr. Allowing-^ hr down time per cycle for regeneration and rinsing
(one to one and one quarter hours should actually be sufficient), the
125
-------
zeolite volume must be increased proportionately to 4300 cu ft to accom-
modate the total design flow. Using four units, each having the dimen-
sions 12 ft x 15 ft x 6 ft deep, the total zeolite volume is 4320 cu ft.
Regeneration procedure and the choice of regenerant composition will be
discussed later in this chapter. However, for purposes of determining
the volume of reusable water produced by the process, it is assumed that
no more than 15 BV of regenerant will be required if regenerant is not
reused and that only 0.5 BV of regenerant will be consumed per cycle if
regenerant is reused. Column rinsing requirements are summarized in
Table 16. To reduce the pH to 10, 11 BV of product water are required
which would be returned to upstream treatment processes. Further rinse
water required would be returned to product water storage as the slightly
higher pH and NaCl concentration of this water should not be detrimental
to product water quality. Because both regeneration and rinsing are
accomplished upflow, no provision was made for a separate backwashing
step prior to regeneration. The effective throughput is 150 BV/cycle
when regenerant is wasted and 165 BV/cycle when regenerant is reused.
The liberal estimate of down time per cycle includes sufficient time
for reprocessing of rinse water returned to upstream treatment. The
volume of product water produced is 9.1 mgd when regenerant is wasted
and slightly less than 10.0 mgd when regenerant is reused. Operating
characteristics of this plant are described in Table 27.
TABLE 27
OPERATING CHARACTERISTICS FOR A 10-mgd CLINOPTILOLITE
ION EXCHANGE FACILITY
Operational Feature
Value
Design Flow Rate
Ammonia Exchange Capacity
Zeolite Volume
Ion Exchange Vessels
Cycle Time
Throughput
15 BV/hr
6.6 eq/cu ft (0.32 meq/g,
5.1 kgr/cu ft)
4300 cu ft
4-12 ft x 15 ft x 6 ft
deep, reinforced concrete
gravity flow units
13 hr including an 11 hr
exhaustion cycle and 2 hr
for regeneration, rinsing
and reprocessing of rinse
water
165 BV/cycle; effective
throughput of 150 BV/cycle
when regenerant wasted,
165 BV/cycle when regenerant
reused
126
-------
Several economies of operation might be realized depending on other
processes included in the total waste treatment system. If lime were
used as the precipitant in an upstream precipitation-clarification
process, lime costs would be appreciably less than those used in this
analysis. On-site recalcination of lime usually is feasible where more
than 100 tons of lime are used per week. An additional advantage of
having a high pH upstream precipitation process would be the use of the
high pH effluent for regenerant makeup, thus reducing lime requirements
for the ion exchange process. For climates where heating in the regene-
rant stripping process would be required in winter, waste heat from lime
recalcining or carbon regeneration furnaces or the gas from sludge
digesters might be used for that purpose.
COST OF AMMONIA REMOVAL
Cost of Ion Exchange Processes
Although the literature concerning specific applications of ion exchange
is extensive, little has been written concerning the cost of ion exchange
processes. In addition, practically all information available has dealt
with small-scale ion exchange softening and demineralization plants. Ion
exchange vessels for these installations are characteristically steel
and rubber-lined steel pressure vessels as opposed to the reinforced
concrete gravity units considered in this analysis. An extensive
discussion of the cost of ion exchange processes in pressurized containers
has been prepared by Sanks [92] and Sanks and Kaufman [93,94]. Other
discussions concernin0 thi design and economic considerations of ion
exchange" units have b§en published by Kunin [44], Peak and David [95],
and Monet [96,97]. Cost estimating guidelines have been published by
Nelson [98] and the Office of Saline Water [99] and were used where
applicable in this analysis. Costs of treatment processes analyzed by
Smith [100], Smith and McMichael [74 ], and Koenig [101] were also
utilized for cost information.
Cost of Regeneration
In addition to information found in Chapter VIII, costs of salt and
lime, zeolite replacement, and stripping of regenerant solutions are
needed to estimate the cost of regeneration. Included in the cost of
regeneration are chemical costs, all zeolite replacement costs, and all
costs associated with regenerant stripping. Costs of chemical and
regenerant storage facilities are included in capital costs which are
considered separately.
Chemical prices were obtained from manufacturers in the San Francisco
area. Because chemical prices vary appreciably, depending on the
location and quantity used, it was necessary to make several assumptions
regarding the availability of chemicals. Transportation charges
127
-------
considerably increase the cost of chemicals. Furthermore, it is diffi-
cult to accurately estimate transportation charges without knowing the
origin and destination of the haul and other factors specific to each
locality. However, in order to obtain prices for the purposes of cost
estimation, transport distances were assumed and an attempt was made to
apply average transport charges for these distances. Transportation
charges for 300 miles were used for NaCl shipment. Because lime is
manufactured in a greater number of locations across the country, only
200 miles transport of lime was assumed. Commercial grades of chemicals
are satisfactory for regeneration and prices were based on this grade of
chemicals. The base price of "stack run" NaCl was quoted as $12.90/ton
(Leslie Salt Co., San Francisco, California). When freight charges of
$15/ton were added, the total price of salt was $27.907ton or $1.30/cwt.
The price of lime was $21/ton and the cost of shipping was estimated to
be $12/ton. Assuming the delivered product to be 90% pure, the effec-
tive lime cost was $36.70/ton of CaO or $1.83/cwt. The price of NaOH
was $81/ton of NaOH as 50% liquid and the price of 300 miles shipping
was estimated to be $15/ton of liquid, bringing the total cost of NaOH
to $in/ton of NaOH or $5.50/cwt.
Estimation of clinoptilolite replacement costs was complicated by the
fact that no large market currently exists for this material. The
Baroid Division of National Lead Company, Houston, Texas currently
markets clinoptilolite only in a -4 mesh size for $75/ton (equivalent
to $1.75/cu ft) f.o.b. Newberry, California [102]. While it is possible
that clinoptilolite might be made available in a 20 x 50 mesh size if
a firm commitment for purchase of a large quantity of the zeolite were
made, no price projections for the 20 x 50 mesh material are available
at this time. Therefore, it was assumed that -4 mesh clinoptilolite
would be crushed to the desired 20 x 50 mesh size at the treatment
plant. A crushing and screening unit consisting of a twin loader
impactor crusher, a vibrating screen, conveyor, and 2000 cu ft storage
was estimated to cost $40,000 including installation. This was amortized
for 10 yr at 5% and added to labor costs (4 man-hr/day) for a total
crushing cost of $0.62/cu ft of product zeolite for a 10-mgd plant.
This is based on a replacement rate of 50 cu ft of zeolite/day and
represents perhaps twice the expected rate of replacement.
Shipping charges for transporting clinoptilolite to the treatment site
are highly speculative, both because of uncertainty concerning the
treatment plant location and because of the possibility of commercial
development of zeolite deposits other than the one at Hector, California.
For this analysis transportation of 400 miles from Hector, California
was assumed. This resulted in a shipping cost of $0.29/cu ft of -4 mesh
zeolite. Costs of zeolite replacement at other locations will vary from
this figure if transport over a significantly greater distance is involved.
Based on a 50" recovery of clinootilolite as 20 x 50 mesh material,
replacement clinoptilolite would cost $4.70/cu ft of product. Costs of
chemicals and clinoptilolite replacement are summarized in Table 28.
128
-------
TABLE 28
COST OF CHEMICALS AND CLINOPTILOLITE
Material
Sodium Chloride
Base price
300 miles transportation
Total cost
Lime (CaO)
Base price
200 miles transportation
Total cost, 90% pure
Total cost per unit of
usable material
Sodium Hydroxide (50% liquid)
Base price
300 miles transportation
Total cost
Clinoptilolite3
Base price
400 miles transportation
Crushing and storage
Total cost per unit of
product
Units
$/ton
$/ton
$/ton
$/ton
$/ton
$/ton
$/ton
$/ton
$/ton
$/ton
$/cu ft
$/cu ft
$/cu ft
$/cu ft
Cost
12.90
15.00
27.90
21.00
12.00
33.00
36.70
81.00
30.00
m . oo
3.50
0.58
0.62
4.70
aBased on 50% recovery of zeolite as 20 x 50 mesh
material.
Estimates for stripping of regenerant solutions were based on stripping
costs given by Smith and McMichael [74, 100]. Because little work has
been published concerning air stripping of concentrated ammonia solutions,
it was assumed that the cost of stripping ammonia from sewage could be
applied to regenerant stripping without serious error. A cost of
$0.10/1000 gal:of regenerant was estimated to cover the total cost of
regenerant stripping.
Regeneration costs were calculated using the regenerant requirements
shown in Figure 27 and Table 13. Zeolite replacement rates were
129
-------
calculated on the basis of attrition rates measured in Chapter VIII
0.25%/cycle for regeneration at pH 11.5, 0.35%/cycle at pH 12.0, and
0.55%/cycle at pH 12.5. Two cases were considered: 1) wasting of
regenerant after one use, and 2) reuse of regenerant by stripping ammonia
from the regenerant solution. For no regenerant reuse, lime requirements
were calculated for a regenerant makeup water having an alkalinity of
80 mg/£ as CaC03. When regenerant is reused, the pH of the regenerant
will drop both as ammonia from the zeolite is absorbed by the regenerant
solution and as the regenerant is passed through the stripping tower.
While the level to which the pH will drop is dependent on several factors
and cannot be accurately estimated, for this analysis it was assumed
that the regenerant pH would have to be raised from 7 or 8 to the
desired value. Salt requirements for regenerant reuse were assumed to
be equal to the stoichiometric amount of ammonia eluted from the zeolite
plus a loss of 0.5 BV of regenerant/cycle.
The cost of regeneration is shown in Figures 34 and 35 for regenerant
wasting and regenerant reuse, respectively. Costs are shown both as
$/eq NH3-N removed and as $/1000 gal of treated water for a waste
containing 20 mg/i NH3-N. The costs in Figure 34 are heavily dependent
on the salt concentration because a high proportion of regeneration
costs are due to salt used for regenerant makeup. The prices estimated
for regeneration using no NaCl at pH 12.0 and 12.5 were made on the
basis of using NaOH for pH adjustment. As shown in Chapter VII, the
use of lime alone for regeneration would require an excessive volume of
regenerant. In addition the effective ammonia capacity of clinoptilolite
is more than twice as great when the zeolite is regeneranted using a
regenerant containing sodium. The cost of regeneration when regenerant
is used only once was nearly the same for regeneration at pH 12.5 using
NaOH and for regeneration at pH 12.0 or 12.5 using lime for pH adjust-
ment and 0.049 Ib NaCl/gal. While the use of slightly less salt might
result in lower treatment costs, some NaCl must be added to prevent the
volume of regenerant required from becoming excessively large as would
be the case if calcium were the only cation in the regenerant. The
least cost method of regeneration when regenerant is not reused was
achieved using 0.049 Ib NaCl/gal at pH 12.5 with lime as the source of
caustic. However, the cost of regeneration using this regenerant
composition ($0.092/1000 gal) was only slightly less than regeneration
using only NaOH at pH 12.0 or 12.5 ($0.097/1000 gal). Because of the
uncertainties in making a cost estimate of this sort, these costs may
change when chemical prices and transportation costs in a specific
location are considered.
Regeneration costs when regenerant reuse is considered (cf. Figure 35)
are relatively insensitive to changes in the salt strength of the
regenerant because of the large fraction of salt which is saved for
reuse. While no one factor dominated the total cost of regeneration in
this case, zeolite replacement costs constituted a much larger propor-
tion of the total regeneration cost and resulted in lower regeneration
costs as the regenerant pH was reduced. The least cost of regeneration
130
-------
0.0 0.1 0.2 0.7 0.8
Renenerant Salt Concentration, Ib NaCl/cial
FIGURE 34. COST OF REGENERATION - NO REGENERANT REUSE
fcO-
0.0 0.1 0.2 0.7 0.8
Reqenerant Salt Concentration, Ib NaCl/gal
FIGURE 35. COST OF REGENERATION -
REUSE OF REGENERANT
131
-------
when regenerant is reused was achieved using either 0.10 or 0.17 Ib NaCI/
gal at pH 11.5 and was estimated to cost $0.042/1000 gal for a wastewater
ammonia concentration of 20 mg/£ NH3-N. The greater volume of the less
concentrated regenerant resulted in somewhat greater costs for lime
addition and regenerant stripping which balanced the greater cost of-
NaCI addition associated with the use of the more concentrated solution.
While regeneration at pH values less than 11.5 was not studied, it is
doubtful that the use of a lower pH regenerant would lead to reduced
regeneration costs, as the increased regenerant volumes required would
result in greater costs for regenerant stripping. In addition, zeolite
replacement will constitute a less significant fraction of regeneration
costs as attrition rates become less at reduced pH levels. These obser-
vations are corroborated by the very small differences estimated for
minimum regeneration costs at pH 11.5 and 12.0. The minimum cost of
regeneration at pH 12.0 was $0.043/1000 gal achieved using either 0.10
or 0.17 Ib NaCl/gal. This cost is only slightly more than the minimum
regeneration cost at pH 11.5 of $0.042/1000 gal.
The small differences in these costs suggest that the determination of
the least cost of regeneration in any particular case will be signifi-
cantly influenced by assumptions made in this analysis which will not
be valid in all cases. Costs of regenerant stripping and zeolite
replacement will be affected by geographic and climatic considerations
as well as by specific agreements which can be made regarding the
purchase of clinoptilolite. (Crushing of clinoptilolite at the mine,
if it could be arranged, would result in considerable savings in trans-
portation and operating costs.) Other factors including the volume of
regenerant which will have to be replaced after each use and the exact
amount of lime required to raise the regenerant pH to the desired level
can accurately be determined only from further studies in which problems
of regenerant reuse are considered. However, as the basis of this
analysis it is believed that regeneration costs of $0.042/1000 gal using
a regenerant containing either 0.10 or 0.17 Ib NaCl/gal at pH 11.5
constitutes a realistic estimate of regeneration costs for the conditions
outlined above.
Capital Costs
The basic cost of gravity flow ion exchange units was estimated from
costs of sand filters given by Smith [ion], Smith and McMichael [ 74],
and Koenig [101]. It was assumed that the greater depth requirements
for an ion exchange unit would be compensated for by the higher rate of
flow per unit of surface area compared to that used in sand filters.
While the bed depth in this case is 6 ft compared to about 2 or 3 ft
for filtration, less headless would be expected in an ion exchange unit
even if the ion exchange medium were also used as a filter due to run
lengths of only 11 hr. A cost of $580,000 was estimated for the ion
exchange units including influent pumps and piping for influent, effluent,
and rinsing waters. This cost was current as of July 1971, having been
updated using the Engineering News-Record Construction Cost Index [103].
132
-------
Piping for regeneration was estimated to cost $26,000. Two 2100-gpm
pumps for regeneration were estimated to cost a total of $30,000,
including installation-
Regenerant handling facilities were estimated for storage of one day's
regenerant volume for the case of regenerant wasting and one-half day's
requirement for regenerant reuse. Facilities for chemical storage were
provided for a 10-day supply of chemicals. The total cost of these
facilities was estimated to be $192,000 where regenerant is wasted and
$143,000 where regenerant is reused.
Because ion exchange equipment is amenable to automation, fully auto-
matic instrumentation and control is desirable as this will substantially
reduce the amount of labor required. Limon and Calise [104] stated that
complete automation and instrumentation of a plant adds less than 5 to
10% to the installed cost of the facility. Because the percentage cost
of instrumentation is nearly independent of the plant size, 5% was added
to the installation cost to cover automation of the plant. A contengency
allowance of 10% was added to the total investment, and an engineering
fee of 10% was added to the sum of the investment cost plus contingency
allowance.
The magnitude of amortization charges depends on the useful life of the
ion exchange equipment and the interest rate. Sanks [92] concluded that
the life of ion exchange equipment is probably not less than 15 years,
while the Office of Saline Water [99] specified a life of 20 years with
shorter lives assumed for some equipment. Because of the innovative
nature of this process, a life of 15 years was assumed in this analysis.
Interest rates for municipal bonds have risen sharply in the last few
years from a relatively stable interest of about 4% for the preceding
15 years. Bonds listed in the Bond Buyer's Index averaged 6.0% in
July 1971. While it seems likely that interest rates will not remain
at this high level over a long period, it is the opinion of analysts
that the return to a long term rate of 4% is equally unlikely because
of techniques which are being used to manage the nation's economy [105].
For this analysis a rate of 5% was assumed. A summary of the cost
analysis for the two plant designs is nresented in Table 29.
Operating Costs
The amount of labor needed to maintain and operate the plant was esti-
mated to be 2-1/2 persons per 8-hr day. Distribution of this labor
included 1-1/2 persons during the day shift and 1/2 person for each of
two night shifts. Labor required for operation and maintenance of the
stripping tower where regenerant is reused and for crushing of replace-
ment clinoptilolite was included in regeneration costs. Labor charges
were based on a cost of $10,000/man-yr. Maintenance was calculated on
the basis of the OSW recommendation of 0.0015% of the total plant invest-
ment [99]. Power costs were based on a rate of $0.01/kw-hr and a
133
-------
TABLE 29
COST SUMMARY - 10-mgd CLINOPTILOLITE ION EXCHANGE PLANT
Item
Capital Costs
Ion Exchange Units
Piping exclusive of influent,
effluent, and rinsing
Pumping exclusive of influent
Regenerant handling
Instrumentation
Contingencies
Engineering
Total capital investment
exclusive of regenerant
stripping where regenerant
is reused
Capital Cost (15 yr life,
5% interest)
Operating Costs
Labor
Maintenance
Power
Regeneration
Total Operating Cost
Operating Cost
Total Cost
Units
$
$
$
$
$
$
$
$/1000 gal
$/day
$/day
$/day
$/day
$/day
$71000 gal
$71000 gal
Costs
No Regenerant
Reuse
580,000
26,000
30,000
192,000
$828,000
41,500
$869,500
87,000
$956,500
95,500
$1,052,000
0.028
70
15
55
920
$1 ,060
0.106
$0.134
Regenerant
Reuse
580,000
26,000
30,000
143,000
$779,000
39,000
$818,000
82,000
$900,000
90,000
$990,000
0.026
70
15
55
420
560
0.056
$0.082
134
-------
continuous power requirement of 310 hp. Regenerant costs were calculated
for the use of 0.049 Ib NaCl/gal at pH 12.5 where regenerant is wasted
and 0.10 Ib NaCl/gal at pH 12.5 where regenerant is reused.
Cost Summary
A summary of the cost analysis presented in Table 29 reveals that ammonia
removal using clinoptilelite will cost $0.134/1000 gal when regenerant is
used once and wasted, and $0.082/1000 gal when regenerant is reused.
Where regenerant is wasted after one use, appropriate disposal costs
must be added to this figure. In cold climates where regenerant is
reused, heating costs for the air stripping process must be considered.
However, as pointed out previously, it is felt that waste heat from
other processes could be utilized for this purpose with little additional
expense being incurred.
The results of this cost analysis may be compared to analyses made for
similar installations at the South Tahoe Public Utility District [76]
and Blue Plains [78] treatment plants. Design for a proposed clinopti-
lolite ion exchange facility at the STPUD was based on the use of
twelve 900-cu ft reactors operating at a flow of 6 BV/hr. Regeneration
was accomplished using a two-stage process with provision for regenerant
reuse. The total cost of this process was estimated to be $0.158/1000
gal. The major difference in this cost and the one 'obtained in the
present analysis was due to the higher capital costs associated with
lower flow rates and the increased expenditure for regeneration facili-
ties. Costs estimated for proposed clinoptilolite ion exchange facili-
ties at the Blue Plains plant in Washington, D. C. were $0.097/1000
gal for a 300-mgd plant and $0.103/1000 gal for a 240-mgd plant [78].
The design in this case was based on a flow of 18 BV/hr and the use of
two columns in series during the exhaustion cycle. Provisions were made
for regenerant reuse and recovery of ammonia by absorption in sulfuric
acid. Although costs for this estimate were very similar to those
obtained herein, capital costs were somewhat higher and chemical costs
somewhat lower than ones obtained in this analysis. The higher capital
costs were probably due to the use of two beds in series and costs
incurred in providing ammonia recovery equipment. The higher chemical
costs obtained in the present analysis were probably due in part to the
more detailed regeneration performance data obtained in this study.
Costs for ammonia removal using clinoptilolite have greater meaning
when compared to costs of other nitrogen removal methods. Costs for
ammonia removal by chlorination discussed in Chapter V were estimated
to be $0.075/1000 gal for chlorine costs alone. However, increasing
concern for the toxicity of wastewater effluents might make this method
undesirable, even if total treatment costs for chlorination were
favorable compared to other methods. Costs for nitrification-denitrifi-
cation processes were estimated to be $0.08/1000 gal by McCarty [9] and
$0.12/1000 gal for the Blue Plains treatment plant in Washington, D. C.
135
-------
[78]. Thus ammonia removal using clinoptilolite is at least comparable
to these methods on the basis of cost alone. In addition it is believed
that there are advantages to the ion exchange process which make it
desirable compared to other processes. It is believed that process
stability and its insensitivity to toxicants constitute a significant
advantage over biological processes. The relatively insignificant
volume of liquid and solid wastes generated by the process will be an
advantage when disposal of these wastes is considered. Finally, chemical
changes in the product water accompanying the removal of ammonia are much
less likely to be objectionable than those introduced by chlorination.
136
-------
XI. ACKNOWLEDGMENTS
This investigation was sponsored in part by Grant No. 17080 DAR from
the Office of Water Quality, Environmental Protection Agency. The work
was performed by Mr. John H. Koon under the direction of Professor
Warren J. Kaufman. Mr. Warren Schwartz served as Project Officer for
the Environmental Protection Agency. The assistance of Messrs. Phil
Palmer and Ronald Swindell, and Miss Kristen Tarr in performing analyses
and conducting experimental work throughout the period of the study is
gratefully acknowledged. Appreciation is likewise expressed to Miss
Janice Sanders and Mr. Paul Graham for their assistance. The authors
are especially indebted to Mrs. Cecelia Dove for her valuable help in
preparation of the manuscript.
Portions of the experimental phase of this study were performed in
conjunction with other work performed concurrently with this project.
The assistance of Mr. Arnold B. Menar, also supported by this contract,
and Mr. Larry A. Esvelt were much appreciated. The generous cooperation
of the engineers and technicians of Special District No. 1 of the East
Bay Municipal Utility District, the Central Contra Costa Sanitary District,
and the Sanitary Engineering Division of the Bechtel Corporation was
appreciated. The successful completion of this project would have been
impossible without the help and cooperation of these persons.
137
-------
XII. REFERENCES
1. McGauhey, P. H., et^ al_. Eutrophication of Surface Waters - Lake Tahoe:
Bioassay of Nutrient Sources. Lake Tahoe Area Council, FWPCA Progress
Report for Grant No. WPD 48-01 (Rl), May 1968.
2. News Release, Washington, D. C. U. S. Department of Interior, May 17, 1970.
3. Bain, Richard C., and Herbert E. Pintler. Effects of the San Joaquin
Master Drain on Water Quality of the San Francisco Bay and Delta -
Appendix, Part C. Nutrients and Biological Response. FWPCA, Southwest
Region, August 1968.
4. Kaiser Engineers. San Francisco Bay-Delta Water Quality Control Program,
Final Report to the State of California. San Francisco, June 1969.
5. U. S. Public Health Service Drinking Water Standards. Public Health
Service Publication No. 956. U.S. Department of Health, Education, and
Welfare, Public Health Service, Washington, D. C. 1962.
6. McKee, J. E., and H. W. Wolf. Water Quality Criteria. 2nd ed. Publ.
No. 3-A. State Water Quality Control Board, Sacramento, State of
California, The Resources Agency, 1963.
7. Stern, Gerald. Removal of Nitrogen from Wastewater. Internal Report.
Taft Water Research Center, Federal Water Pollution Control Admin.,
Cincinnati, August 1966. (This document is unavailable. However, much
information contained in this report is also found in Nitrogen Removal
from Wastewaters, Report No. ORD-17010, FWQA, Cincinnati, Ohio, Oct. 1970.)
8. Rohlich, G. A. "Methods for the Removal of Phosphorus and Nitrogen
from Sewage Plant Effluents," in Advances in Water Pollution Research,
W. W. Eckenfelder, ed. New York: Pergamon Press, 1964, pp. 207-214.
9. McCarty, Perry L. "Biological Processes for Nitrogen Removal -Theory
and Application," presented at the 12th Sanitary Engineering Conference'
Univ. of Illinois, Urbana, 111., February 12, 1970.
10. Samples, William R. "Removal of Nitrogenous Compounds from Wastewaters,"
in Water Reuse. Chemical Engineering Progress Symposium Series,
Amer. Institute of Chemical Engineers, 6^(78):223-229, 1967.
11. Barth, Edwin F., Richard C. Brenner, and Ronald F. Lewis. "Chemical-
Biological Control of Nitrogen and Phosphorus in Wastewater Effluent,"
J. Water Pollution Control Fed., 40(12):2040-2054, 1968.
12. Ames, L. L. "Zeolitic Removal of Ammonium Ions from Agricultural and
Other Wastewaters," 13th Pacific Northwest Industrial Waste Conference
Proceedings, Washington State University, Pullman, Washington,
April 1967.
139
-------
13. Mercer, B. >!., L. L. Anes, C. J. Touhill, W. J. Van Slyke, and
P.. B. Dean. "Ammonia Removal from Secondary Effluents by Selective
Ion Exchange," J. 1,'ater Pollution Control Fed., 42_:R95-R107,
February 1970.
14. Mercer, B. W., R. C. Arnett, and R. B. Dean. "Optimization of Column
Performance for Ammonia Removal from Wastewater Using Selective Ion
Exchange," Wastewater Reclamation and Reuse Workshop Proceedings.
Sanitary Engineering Research Laboratory, University of California,
Berkeley, June 25-27, 1970.
15. Helfferich, Friedrich. Ion Exchange. New York: McGraw-Hill, 1962.
16. Daniels, Farrington, and Robert A. Alberty. Physical Chemistry.
3rd ed., New York: Wiley, 1966.
17. Boyd, G. E., J. Schubert, and A. W. Adamson. "The Exchange Adsorption
of Ions from Aqueous Solutions by Organic Zeolites. I. Ion-exchange
Equilibria," J. Amer. Chem. Soc., 69_(11): 2818-2829, 1947-
18. Walton, Harold F. "Ion Exchange Equilibria," in Ion Exchange - Theory
and Application, ed. by Frederick C. Nachod. New York: Academic Press,
1949.
19. Baetsle, Leo H. Notes on the Physical Chemistry of Ion Exchange.
Radiological Health Sciences Program, School of Public Health, University
of California, Berkeley, 1963.
20. Vermeulen, Theodore, Gerhard Klein, and Nevin K. Hiester. "Adsorption
and Ion Exchange," Section 16. Chemical Engineers' Handbook. 5th ed.,
New York: McGraw-Hill, in press.
21. Glueckauf, E. "Principles of Operation of Ion Exchange Columns," in
Ion Exchange and Its Application. Society of Chemical Industries,
London, 1955.
22. Vermeulen, Theodore. "Separation by Adsorption Methods," in Advances
in Chemical Engineering, vol. 2. ed. by Thomas B. Drew and John W. Hoopes.
New York: Academic Press, 1958.
23. Boyd, G. E., A. W. Adamson, and L. S. Myers. "The Exchange Adsorption
of Ions from Aqueous Solutions by Organic Zeolites. II. Kinetics,"
J. Amer. Chem. Soc. , 69^:2836-2848, 1947.
24. Vermeulen, Theodore. "Theory for Irreversible and Constant-Pattern
Solid Diffusion," Industrial and Engineering Chemistry, 45_(8): 1664-1670,
1953.
25. Hiester, Nevin K., Shirly B. Radding, Richard I. Nelson, Jr., and
Theodore Vermeulen. "Interpretation and Correlation of Ion Exchange
Performance Under Nonlinear Equilibria," A.I.Ch.E. Jour., 2(3):404-411,
Sept. 1956.
140
-------
26. Thqmas, Henry'C. "Heterogeneous Ion Exchange in a Flowing System,"
J. Amer. Chem. Soc.. 66(10) :1664-1666, 1944.
27. Hiester, Nevin K., and Theodore Vermeulen. "Saturation Performance of
Ion Exchange and Adsorption Columns," Chem. Eng. Progress, 48(10):505-
516, Oct. 1952.
28. Hall, Kenneth R., Lee C. Eagleton, Andreas Acrivos, and Theodore
Vermeulen. "Pore- and Solid-Diffusion Kinetics in Fixed-Bed Adsorption
Under Constant-Pattern Conditions," Industrial and Engineering Chemistry,
5_(2):212-223, May 1966.
29. Dranoff, Joshua, and Leon Lapidus. "Equilibrium Ternary Ion Exchange
Systems," Industrial and Engineering Chemistry, 49_(8): 1297-1302, Aug. 1957.
30. Klein, Gerhard, Daniel Tondeur, and Theodore Vermeulen. "Multicomponent
Ion Exchange in Fixed Beds General Properties of Equilibrium Systems,"
Industrial and Engineering Chemistry Fundamentals, 6_(3): 339-351, Aug. 1967.
31. Tondeur, Daniel, and Gerhard Klein. "Multicomponent Ion Exchange in
Fixed Beds Constant Separation-Factor Equilibrium," Industrial and
Engineering Chemistry Fundamentals, _6(3):351-361, Aug. 1967.
32. Gleason, George H., and Alfred C. Loonam. "The Development of a Chemical
Process for Treatment of Sewage," Sewage Works J., 5_(l):61-73, 1933.
33. Gleason, George H., and Alfred C. Loonam. "Results of Six Months'
Operation of Chemical Sewage Purification Plant," Sewage Works J.>
6.(3):450-458, 1934.
34. Nesselson, Eugene J. Removal of Inorganic Nitrogen from Sewage Effluent.
Ph.D. Thesis. University of Wisconsin, Madison, Wis., 1954.
35. Smith, Clinton E., and Robert L. Chapman. Recovery of Coagulant, Nitrogen
Removal, and Carbon Regeneration in Waste Water Reclamation. Final report
to Federal Water Pollution Control Admin, for Grant WPD-85. South Tahoe
Public Utility District, South Lake Tahoe, Calif., June 1967.
36. Slechta, Alfred F., and Gordon L. Gulp. "Water Reclamation Studies at
the South Tahoe Public Utility District," J. Water Pollution Control Fed.,
39(5):787-814, May 1967.
i
37. Battelle Memorial Institute. Ammonia Removal -from Agricultural
Runoff and Secondary Effluents by Selected Ion Exchange. Report No.
TWRC-5, Federal Water Pollution Control Administration, Cincinnati,
Ohio, March 1969.
38. Amphlett, C. B. Inorganic Ion Exchangers. Chapter 3. "The Zeolites."
New York: Elsevier, 1964.
141
-------
39. Sherry, Howard S. "The Ion Exchange Properties of Zeolites," Chapter
3 in Ion Exchange, vol. 2. ed. by Jacob A. Marinsky. New York:
Marcel Dekker, 1969.
40. Marshall, C. E. The Colloid Chemistry of the Silicate Minerals. New
York: Academic Press, 1949.
41. Barrer, R. M. "Ion Exchange and Ion Sieve Processes in Crystalline
Zeolites," J. Chem. Soc. London, 2342-2350, 1950.
42. Ames, L. L. "Mass Action Relationships of Some Zeolites in the Region
of High Competing Cation Concentration," Amer. Mineralogist, 48:
868-882, 1963.
43. Barrer, R. M. "Crystalline Ion Exchangers," Proc. Chem. Soc. London,
99-112, 1958.
44. Kunin, Robert. Ion Exchange Resins, 2nd ed. N-w York: Wiley, 1958.
45. Schaller, W. T. "The Mordenite-Ptilolite Group; Clinoptilolite, A
New Species," Amer. Mineralogist, 17:128-134, 1932.
46. Kostov, Ivan. Mineralogy. London: Oliver and Boyd, 1968.
47. Berman, Harry. "Constitution and Classification of the Natural Silicates,'
Amer. Mineralogist, ^:342-408, 1937.
48. Pabst, Adolf. Minerals of California. Bulletin No. 113. Division of
Mines, Calif. Dept. of Natural Resources, Calif. State Printing Off.,
Sacramento, Calif., Feb. 1938.
49. Mumpton, Frederick A. "Clinoptilolite Redefined," Amer. Mineralogist,
45:351-361, 1960.
50. Barrer, R. M., R. Papadopoulos, and L. V. C. Rees. "Exchange of Sodium
in Clinoptilolite by Organic Cations, J. Inorg. Nucl. Chem., 29:
2047-2063, 1967.
51. Ames, L. L., L. B. Sand, and S. S. Goldich. "A Contribution on the
Hector, California Bentonite Deposit," Econ. Geol. 53^22-37, 1958.
52. Ames, L. L. "Some Zeolite Equilibria with Alkali Metal Cations,"
Amer. Mineralogist, 49:127-145, 1964.
53. Sheppard, Richard A. "Zeolites in Sedimentary Deposits of the United
States A Review," in Second International Conf. on Molecular Sieve
Zeolites, Amer. Chem. Soc., Washington, D. C., 1970, pp. 428-459.
54. lijima, Azuma. "Composition and Origin of Clinoptilolite in the Nakano-
sawa Tuff of Kumoi, Hokkaido," in Second International Conf. on Molecular
Sieve Zeolites, Amer. Chem. Soc., Washington, D. C. 1970, pp. 540-547.
142
-------
55. Minato, Hideo, and Minoru Utada. "Clinoptilolite from Japan," in
Second International Conf. on Molecular Sieve Zeolites, Amer. Chem.
Soc., Washington, D. C., 1970, pp. 535-539.
56. Inoue, Yoriteru. Investigation of Deep Hell Injection of Radioactive
Wastes. Ph.D. Thesis. University of California, Berkeley, 1962.
57. Ames, L. L. "Effect of Base Cation on the Cesium Kinetics of Clinopti-
lolite," Amer. Mineralogist, 47:1310-1316, 1962.
58. Howery, Darryl G., and Henry C. Thomas. "Ion Exchange on the Mineral
Clinoptilolite," J. Phys. Chem., 69(2):531-537, 1965.
59. Frysinger, Galen R. "Cesium-Sodium Ion Exchange on Clinoptilolite,"
Nature, 194:351-353, 1962.
60. Ames, L. L. "The Cation Sieve Properties of Clinoptilolite," Amer.
Mineralogist, 45:689-700, 1960.
61. Duolite Ion Exchange Manual. Diamond Shamrock Chemical Company,
Redwood City, California, 1969.
62. Ames, L. L. "Cation Sieve Properties of the Open Zeolites Chabazite,
Mordenite, Erionite and Clinoptilolite," Amer. Mineralogist, 46:1120-
1131, 1961.
63. Merkle, A. B., and M. Slaughter. "The Crystal Structure of Heulandite,"
Amer. Mineralogist, 52:273-276, 1967-
64. Ames, L. L. "Some Zeolite Equilibria with Alkaline Earth Metal Cations,"
Amer. Mineralogist. 49_:1099-1110, 1964.
65. Pauling, Linus. The Nature of the Chemical Bond. Cornell University
Press, Ithaca, N. Y., 1948.
66. Jenny, H. "Studies on the Mechanism of Ionic Exchange in Colloidal
Aluminum Silicates," J. Phys. Chem., 36_(2):2217-2258, 1932.
67. Grim, Ralph E. Clay Mineralogy. New York: McGraw-Hill, 1968.
68. Ames, L. L. "Zeolite Cation Selectivity," Canadian Mineralogist.
8:325-333, 1965.
69. Ames, L. L. "Kinetics of Cesium Reactions with Some Inorganic Cation
Exchange Materials," Amer. Mineralogist. 47:1067-1078, 1962.
70. Ames, L. L. "Self-Diffusion of Some Cations in Open Zeolites," Amer.
Mineralogist, 50:465-475. 1965. "
71. Baker, Gerald C. "Water Softening by Base Exchange," J. Amer. Water
Works Assoc., 11:128-149, 1924.
143
-------
72. Barrer, R. M., and M. B. Makki. "Molecular Sieve Sorbents from
Clinoptilolite," Canadian J. Chem., 42; 1481 -1487, 1964.
73. Gulp, Russell L. "Nitrogen Removal by Air Stripping," in Wastewater
Reclamation and Reuse Workshop Proceedings, Sanitary Engineering
Research Laboratory, University of California, Berkeley, June 25-27,
1970.
74. Smith, Robert, and Walter F. McMichael. Cost and Performance Esti-
mates for Tertiary Wastewater Treating Processes. Report No. TWRC-9,
Federal Water Pollution Control Admin., Cincinnati, Ohio, June 1969.
75. Farrell, J. B. "Ammonia Nitrogen Removal by Stripping with Air," in
Nitrogen Removal from Wastewaters. Report No. ORD-17010, Federal
Water Quality Admin., Cincinnati, Ohio, Oct. 1970.
76. Wilson, Jerry C., and David R. Evans. Phase I Engineering Design
Report - Supplementing Ammonia Stripping with Further Nitrogen Removal
by Selective Ion Exchange and Breakpoint Chiorination, presented to
Federal Water Quality Admin., Project No. L-145.98, South Tahoe Public
Utility District, South Lake Tahoe, Calif., April 1970.
77. Dean, R. B. "Removal of Ammonia Nitrogen by Selective Ion Exchange,"
in Nitrogen Removal from Wastewaters. Report No. ORD-17010, Federal
Water Quality Admin., Cincinnati, Ohio, Oct. 1970.
78. Bechtel Corporation. Preliminary Cost Estimates for a Blue Plains
Advanced Waste Treatment Plant. Prepared for the Federal Water Quality
Admin., Bechtel Corp., San Francisco, Calif. July 10, 1970.
79. Mercer, B. W., and L. L. Ames. "The Adsorption of Cesium, Strontium,
and Cerium on Zeolites from Multication Solutions," A.E.C. Document
HW-78461, Hanford Atomic Products Operation, Richland, Washington,
Aug. 1963.
80. Standard Methods for the Examination of Water and Wastewater. 12th
ed. New York: Amer. Public Health Assoc., 1965.
81. FWPCA Methods for Chemical Analysis of Water and Wastes. Federal Water
Quality Admin., Cincinnati, Ohio, Nov. 1969.
82. Jenkins, David. Analytical Methods. Sanitary Engineering Research
Laboratory, University of California, Berkeley, March 1966.
83. Berry, T. F. Crushing Tests on a Sample of Clinoptilolite. Mines
Branch Investigation Report IR 61-66, Dept. of Mines and Tech. Surveys,
Mineral Processing Div., Ottawa, June 1961.
84. Helfferich, F. A. "Ion Exchange Kinetics," in Ion Exchange, Jacob A.
Marinsky, ed., New York: Marcel Dekker, 1969.
85. Klein, Gerhard. Private communication, 1971.
144
-------
86. "Effect of Various Substances on Concrete and Protective Treatments,
Where Required." Portland Cement Association, Skokie, Illinois,
1968.
87. Hughes, W. J., and H. B. Crane. "Trends in Municipal Zeolite Water
Softening," J. Amer. Water Works Assoc., 22_(1):68-81, Jan. 1930.
88. Streicher, Lee. "Operating Experiences at La Verne Softening Plant,"
J. Amer. Water Works Assoc.. 37(5):482-488, May 1945.
89. Montgomery, James M., and William A. Aultman. "Water Softening and
Filtration Plant of the Metropolitan Water District of Southern
California," J. Amer. Water Works Assoc., 32(1):l-24, Jan. 1940.
90. Water Quality and Treatment. 3rd ed. Prepared by the American Water
Works Assoc., Inc., New York: McGraw-Hill, 1971.
91.' Water Treatment Plant Design. American Water Works Assoc., Inc.,
New York, 1969.
92. Sanks, Robert L. Partial Demineralizatton of Brackish Water by
Ion Exchange. Ph. D. Thesis, University of California, Berkeley,
1965.
93. Sanks, Robert L., and Warren J. Kaufman. "Partial Demineralization
of Brackish Water," Proc. ASCE, J. Sanitary Eng. Div. . 92_(SA6):
95-111, Dec. 1966.
94. Sanks, Robert L., and Warren J. Kaufman. "Recycling of Saline
Wastes," Proc. 3rd International Conf. on Water Pollution Research,
1966. .
95. Peak, R., and M. David. "Costs of Cation Exchange Equipment," Chem.
Eng. Progress. 53_:37J-40J, Jan. 1957.
96. Monet, G. P. "Cost of Ion Exchange," Chemical Engineering, 57(3):
106-107, March 1950. '
97. Monet, G. P. "Economic Considerations," Chapter Six of Ion Exchange
Technology. 2nd ed. Edited by F. C.' Nachod and'Jack Schubert.
New York: Academic Press, Inc. 1956.
98. Nelson, W. L. "Cost-imating," A technical Manual published'by the
Oil and Gas Journal, The Petroleum Publishing Co., Tulsa, 1957.
99. Office of Saline Water, U. S1. Dept. of Interior. A Standardized
Procedure for Estimating Costs of Saline Water Conversion.1 March 1956.
100. Smith, Robert. "Cost of Conventional and Advanced Treatment of
Wastewater," J. Water Pollution Control Fed., 49_(9): 1546-1574, Sept.
1968.
145
-------
101. Koenig, Louis. "The Cost of Water Treatment by Coagulation, Sedi-
mentation, and Rapid Sand Filtration," J. Amer. Water Works Assoc.,
59(3):290-336, March 1967.
102. Offeman, R. E. Technical Development Coordinator, Baroid Division,
National Lead Co., Houston, Texas. Private communication, May 13,
1971.
103. Engineering News-Record. ]j86_(27): 76, July 8, 1971.
104. Limon, Leonard, and V. J. Calise. "Automation of Boiler Feedwater and
Process Water Treatment Plants," Tappi, 39(11):68A-86A, Nov. 1956.
105. Ostrov, Michael H. Dean Witter and Co., Inc., Berkeley, California.
Private communication, July 1971.
146
-------
XIII. GLOSSARY
Symbol Definition
A, B Cations.
a Valence of ion A; activity.
b Valence of ion B.
BV Bed volume; equal to the gross volume occupied by the exchange
material.
c Solution phase concentration, meq/£.
CQ Total solution concentration of cations, meq/£.
CCCSD Central Contra Costa Sanitary District.
d Particle diameter.
D.p Film diffusion coefficient, sq cm/sec.
D Solid diffusion coefficient, sq cm/sec.
D Pore diffusion coefficient, sq cm/sec.
EBMUD East Bay Municipal Utility District.
F Volumetric flow rate, ma/sec or BV/hr.
h Height of bed, cm or ft.
I+ Cationic strength, moles/£.
1C Selectivity coefficient, Equation 3. Ion A in solution phase
displacing ion B in solid phase.
K Equilibrium constant.
eq ^
m. Concentration of cation i, moles/£.
N Number of transfer units, dimensionless; Equation 9.
NH3-N Total ammonia nitrogen concentration, mg/£ or meq/£.
NH^-N Unionized ammonia nitrogen concentration, mg/£ or meq/£.
147
-------
Symbol Definition
NH4-N Ammonium ion concentration expressed as N, mg/a or meq/2..
q Solid phase concentration, meq/g.
Q Total exchange capacity, meq/g.
S Column cross-sectional area, sq cm or sq ft.
SERL Sanitary Engineering Research Laboratory.
t Time, sec.
T Throughput parameter, equal to ratio of meq fed to column to
meq total column capacity.
v Volume of zeolite bed, cu cm or cu ft.
x Dimensionless fluid phase concentration (x = c/C ).
y Dimensionless solid phase concentration (y = q/Q).
z. Valence of cation i.
Z Zeolite exchange site.
A
aB Separation factor, Equation 5. Ion A in solution phase
replacing ion B in solid phase.
£ Voids ratio
p. Bulk density of zeolite bed, g/cu cm.
* Correction factor for pore diffusion, Equation 7-
y Correction factor for solid diffusion, Equation 8.
* Equilibrium value.
148
-------
XIV. APPENDICES
Page
A. Column Construction Details
Figure 1: Details of Columns in SERL Column Unit 152
Figure 2: Details of Column Unit Used in EBMUD and
CCCSD Studies 153
Figure 3: Detail of Surface Wash Apparatus 154
B. Concentration Histories - Effect of Water Composition on
Ammonia Exchange Capacity
Figure 1: Effect of Water Composition on Ammonia
Exchange Capacity Run 2 156
Figure 2: Effect of Water Composition on Ammonia
Exchange Capacity - Run 3 156
Figure 3: Effect of Water Composition on Ammonia
Exchange Capacity - Run 4 157
Figure 4: Effect of Water Composition on Ammonia
Exchange Capacity - Run 5 157
Figure 5: Effect of Water Composition on Ammonia
Exchange Capacity - Run 6 . . , 158
C. Ammonia Concentration Histories Effect of pH on Ammonia
Exchange
Figure 1: Ammonia Concentration History - pH 4.0 160
Figure 2: Ammonia Concentration History - pH 6.0 160
Figure 3: Ammonia Concentration History - pH 8.0 161
Figure 4: Ammonia Concentration History - pH 9.5 161
Figure 5: Ammonia Concentration History - pH 10.0 .... 162
D. Wastewater Characterization
Figure 1: Waste Characterization - SERL Primary
Effluent 164
149
-------
Page
Table 1: Composition of SERL Primary Effluent 167
Figure 2: Waste Characterization EBMUD Primary
Effluent 168
Table 2: Composition of EBMUD Primary Effluent 170
Figure 3: Waste Characterization CCCSD Primary
Effluent 171
Table 3: Composition of CCCSD Primary Effluent 173
E. Supplementary Operating Data
Table 1: SERL Series I Supplementary Data 176
Table 2: SERL Series II Supplementary Data 181
Table 3: EBMUD Supplementary Data 183
Table 4: CCCSD Supplementary Data 187
Table 5: Biochemical Oxygen Demand 188
Table 6: Organic Plus Ammonia Nitrogen 189
150
-------
APPENDIX A
COLUMN CONSTRUCTION DETAILS
151
-------
INFLUENT (DOWNFLOW
OPERATION 11" FROM
CONSTANT HEAD2 TANK
-y- PVC PLUG
ETHYLENE
VINYL ACETATE
TUBINGv
t
PVC BALL VALVE
y PVC UNION
WASTE
-5- PLEXIGLASS FLANGE- 7"DIAM.
2 WITH 6 HOLES DRILLED EQUALLY
x l- HEX. HD. BOLTS (6)
8 OUTLETS SPACED 6"APART UP COLUMN
- WALL , PLEXIGLASS
NYLON "SWAGELOK
CONNECTOR
y PVC PLUi
EFFLUENT
(DOWNFLOW OPERATION
TO RATE CONTROLLER,
y PVC BALL VALVE
^ HOLES DRILLED ^
BETW. CENTERS
j|"x2|- HEX HD. BOLTS (6)
RUBBER GASKET
NYLON SCREEN, 210/t MESH
RUBBER GASKET
PVC UNION
y- PVC BALL VALVE
REGENERANT
INFLUENT
FIGURE I. DETAILS OF COLUMNS IN SERL COLUMN UNIT
152
-------
PRESSURE GAUGE
NEOPRENE
HOSE
3"
-f- PVC UNION
-^- PVC BARBED HOSE
CONNECTOR
FITTING FOR SURFAQE.WASH
y PLEXIGLASS
2 FLANGED _Ji
7'V
PVC BALL VALVE
DRILL - , 6 HOLES
x l- HEX.HD. BOLTS
--"- LIP CUT INTO FLANGE
NYLON SCREEN.
210 fJ- MESH
8-12 MESH RIVER
GRAVEL
NYLON SCREEN
MESH
_
-|x 1 HEX.HD. BOLTS (6)
RUBBER GASKET
*2j. HEX.HD. BOLTS (6)
4 SINGLE UNION
PVC BALL VALVE
|- 90° ELBOW
PVC
RUBBER GASKET
PVC UNION, -|
NEOPRENE HOSE
-- PVC BARBED HOSE
CONNECTOR
FIGURE 2. DETAILS OF COLUMN UNIT USED IN EBMUD
AND CCCSD STUDIES
153
-------
X
S.S- QUICK-RELEASE
FITTING
BRASS SHUT-OFF
COCK *
RUBBER 0-RING
IMPERIAL FITTING
+- PVC CAP
I" PLEXIGLASS ROD
ALL HOLES
^ DRILL
(b
f
WELDED ACROSS
END OF TUBE
~^T
i
I
l
/;. |
l
i
1
i
~' i
i
r
1
1
1
:: 1
1
i
1
1
o 1
1
f
1
1 '
:ld
j'i__L
3
./ **
t
i
4
t
I
r .
4
~r
i
L,-
i
4
31
I
^J"
J_
r
4
I
ki
n 3-
PLEXIGLASS
FLANGE
1-0. D. x4'-0" STAINLESS
STEEL TUBE
16
FIGURE 3. DETAIL OF SURFACE WASH APPARATUS
154
-------
APPENDIX B
CONCENTRATION HISTORIES - EFFECT OF WATER
COMPOSITION ON AMMONIA EXCHANGE CAPACITY
155
-------
C"
O)
£Z
o
QJ
U
o
o
5.0 _
4.0 -
3.0 -
2.0
1.0 -A
0.0
i i i i
Pbv = 5800 g
F - 15.0 BV/hr
400 600
Throughout, BY
800
1000
FIGURE 1. EFFECT OF WATER COMPOSITION ON AMMONIA
EXCHANGE CAPACITY - RUN 2
5.0
I I I
I I
i I
Pbv - 6000 g
F = 15.0 BV/hr
(C0)Ca = 3.3 meq/£
0.0
0.0
200
FIGURE 2.
400 600
Throughput, BV
1000
EFFECT OF WATER COMPOSITION ON AMMONIA
EXCHANGE CAPACITY - RUN 3
156
-------
7.0
PbV '= 4770 g
F = 18.9 BV/hr
= 3.4 meq/£
200
400 600
Throughput, BV
800
10.0
FIGURE 3. EFFECT OF WATER COMPOSITION ON AMMONIA
EXCHANGE CAPACITY - RUN 4
8.0
o?
D
, 6.0
fO
i-
4->
c:
QJ
CJ
c
O
4.0
2.0
0.0
I I I I I I I
Pbv = 5050 g
F = 17.2 BV/hr
(CQ)Ca = 6.9 meq/i
oHg '
O J n r-i
(Co)NH3~N
(CQ)K=0.38
0 200 400 600 800
Throughout, BV
1000
1200
FIGURE 4. EFFECT OF WATER COMPOSITION ON AMMONIA
EXCHANGE, CAPACITY - RUN 5
157
-------
sr
o
ro
s_
-i->
c
cu
u
c:
o
o
1 1 1 1 1
Pbv = 5500 g
F = 15.0 BV/hr
(C )N = 7.60 meq/a
-------
APPENDIX C
AMMONIA CONCENTRATION HISTORIES - EFFECT OF
pH ON AMMONIA EXCHANGE
159
-------
6.0
5.0
4.0
T T
Pbv = 4680 g
F = 19.2 BV/hr
0.0
200
FIGURE
400 600
Throughout BY
AMMONIA CONCENTRATION HISTORY
800
oH 4.0
o.
o
o
o
1.0
0.8
0.6
0.4
0.2
0.0
Pbv = 4770 g
F = 18.9 BV/hr
0
200
800
400 600
Throughout BV
FIGURE 2. AMMONIA CONCENTRATION HISTORY - ~\\ 6.0
1000
160
-------
I 1 I 1 1 1
= 5040 g
F = 17.9 BV/hr
400 600 800
Throughput BV
1000
FIGURE 3. AMMONIA CONCENTRATION HISTORY - oH 8.0
^10.0
t 9.0
UJ
1.0
0.8
o
"^ 0.6
ft
1 0.4
0.2
0.0
v = 4950 g
F = 18.1 BV/hr
J 1 I J
200
400 600 800
Throughput - BV
1000
1200
1200
FIGURE 4. AMMONIA CONCENTRATION HISTORY - oH 9.5
161
-------
11.0
10.0
Pbv = 4860 g
F = 18.4 BV/hr
200
400 600
Throughout BV
800
1000
rI3URE 5. AMMONIA CONCENTRATION HISTORY - oH 10.0
162
-------
APPENDIX D
WASTEWATER CHARACTERIZATION
163
-------
8.0 L
O,
200
0815 1215
1615 2015 2415
Time, 1-27-71
0415 0815
to 1-28-71
1215 1615
a. ALKALINITY, pH
-14
J 6
.0
,0
,0
.0
,0
0
0
0
0
1 1
-Total Phosphorus
Total Soluble
Phosphorus
0815 1215 1615 2015 2415 0415 0815 1215 1615
Time, 1-27-71 to 1-28-71
b. TOTAL AND TOTAL SOLUBLE PHOSPHORUS
1. WASTE CHARACTERIZATION - SERL PRIMARY E-LUEN1
164
-------
O)
CD
o
CD
s-
o
Organic plus
Nitrogen
Ammonia Nitrogen
0815 1215 1615 2015 2415 0415 0815 1215 1615
Time, 1-27-71 to 1-28-71
c. AMMONIA NITROGEN, ORGANIC PLUS AMMONIA NITROGEN
300
o
o
o
c:
res
o
CO
o
OO
o>
"O
c:
O)
r-
200
100
Solids
_L
'815 1215 1615 2015 2415 0415 0815 1215 1615
Time, 1-27-71 to 1-28-71
d. SUSPENDED SOLIDS, BOD, COD
FIGURE 1 (Continued). WASTE CHARACTERIZATION - SERL
PRIMARY EFFLUENT
165
-------
0815 1215 1615 2015 2415 0415 0815
Time, 1-27-71 to 1-28-71
CALCIUM, MAGNESIUM
1215 1615
e.
80
70
50
40
30
20
10
0
0815 1215
T
e
on
to
re
£
Q.
-Fotassium (Hltered)
I
1615 2015 2415 0415 0815
Time, 1-27-71 to 1-28-71
f. SODIUM, POTASSIUM
1215 1615
rHURE 1 (Continued). WASTE CHARACTERIZATION
PRIMARY EFFLUENT
SERL
166
-------
TABLE 1
COMPOSITION OF SERL PRIMARY EFFLUENT
Analysis
24-hr
Average
Concentration3
COD, mg/a
BOD, mg/a
Suspended Solids, mg/£
Alkalinity, mg/a as CaC03
Soluble Phosphorus, mg/a
Total Phosphorus, mg/a
Org. + Ammonia Nitrogen, mg/a
NH3-N, mg/£
Sodium, mg/£
Potassium (filtered), mg/a
Calcium, mg/£
Magnesium, mg/a
pH
239
86
66.5
262
7.8
9.2
26.5
18.8
51.1
6.8
47
12
7.6
aAverage for first 24 samples of 33 samples
take^.
167
-------
600
=<
~~'-^
en
E. 500
Q
S 400
300
o
-o
200
r 100
?200
_L
Suspended Solids
_L
JL
_L
J_
1600 2000 2400 0400 0800
Time, 11-12-70 to 11-13-70
a. SUSPENDED SOLIDS, COD
1200
D.
7.5
7.0
6.5
o
<_>
rcJ
OK1
CD
300 -
200
100
03
CC
_L
_L
1200 1600 2000 2400 0400
Time, 11-12-70 to 11-13-70
0800
1200
b. ALKALINITY, pH
FHURE 2. WASTE CHARACTERIZATION - EBMUD PRIMARY EFFLUENT
168
-------
1200 1600 2000 2400 0400
Time, 11-12-70 to 11-13-70
c. AMMONIA, SODIUM, POTASSIUM
0800
1200
1200
1600
0800
2000 2400 0400
Time, 11-12-70 to 11-13-70
d. CALCIUM, MAGNESIUM
FIGURE 2 (Continued). WASTE CHARACTERIZATION - EBI'IUD
PRIMARY EFFLUENT
1200
169
-------
TABLE 2
COMPOSITION OF EBMUD PRIMARY EFFLUENT
Analysis
24-hr
Average
Concentration
COD, mg/t
Suspended Solids, mg/z
Alkalinity, mg/£ as CaC03
NH3-N, mg/£
Sodium, mg/£
Potassium, mg/a
Calcium, mg/£
Magnesium, mg/£
PH
484
94
160
19.2
183
33.4
39
12
7.1
170
-------
<=*
O)
0
O
O
300,
200
o
GO
100
-a
c.
0)
ex
oo
0
1400
_L
_L
_L
1800 2200 0200 0600
Time, 4-6-71 to 4-7-71
a. SUSPENDED SOLIDS, COD
1000
1400
8.0
7.5
7.0
CO
8 400
o
* 300!
200
rj 100
re
1400 1800 2200 0200 0600
Time, 4-6-71 to 4-7-71
1000
1400
b. ALKALINITY, pH
FIGURE 3. WASTE CHARACTERIZATION - CCCSD PRIMARY EFFLUENT
171
-------
160
140
cr
E
- 120
O
I/O
03
O
100
T 80
CO
E
13
O
Q_
o>
c:
re
60
40
Sodi urn-
lei um
Ammonia
Nitrogen
1800
2200 0200 0600
Time, 4-6-71 to 4-7-71
1000
1400
c. Ammonia, Sodium, Potassium, Calcium, Magnesium
FIGURE 3. WASTE CHARACTERIZATION - CCCSD PRIMARY EFFLUENT
172
-------
TABLE 3
COMPOSITION OF CCCSD PRIMARY EFFLUENT
Analysis
24-hr
Average
Concentration
COD, mg/£
Suspended Solids, mg/a
Alkalinity, mg/£ As CaC03
NH3-N, mg/i
Sodium, mg/a
Potassium, mg/a
i
Calcium, mg/a
Magnesium, mg/&
pH
222
83
281
20.9
121
10.8
77
8
7.7
173
-------
APPENDIX E
SUPPLEMENTARY OPERATING DATA
-------
TABLE 1
SERL SERIES I SUPPLEMENTARY DATAa
Run
1
2,3
4,5
6,7
8
Date
1970
8-25
8-26
8-27
8-28
8-29
Treatment
Pri. Eff.
Act. Sludge
Ppt. Eff.
Col. Eff.
Pri. Eff.
Act. Sludqe
Ppt. Eff.
Col. Eff.
Pri. Eff.
Act. Sludge
Ppt. Eff.
Col. Eff.
Pri. Eff.
Act. Sludge
Ppt. Eff.
Col. Eff.
Pri. Eff.
Act. Sludge
Ppt. Eff.
Col. Eff.
COD
mg/£
248
107
27
25
260
79
24
21
132
99
27
25
229
78
19
18
272
62
26
24
SS
mg/fc
76
32
1.2
1.4
90
49
1.2
<1.0
36b
49
<1.0
-------
TABLE 1 (Continued)
SERL SERIES I SUPPLEMENTARY DATA9
Run
9,10
11,12
13,14
15,16
17,18
Date
1970
9-1
9-2
9-3
9-4
9-5
Treatment
Pri. Eff.
Act. Sludge
Ppt. Eff.
Col. Eff.
Pri. Eff.
Act. Sludge
Ppt. Eff.
Col. Eff.
Pri. Eff.
Act. Sludge
Ppt. Eff-.
Col. Eff.
Pri. Eff.
Act. Sludge
Ppt. Eff.
Col. Eff.
Pri. Eff.
Act. Sludge
Ppt. Eff.
Col. Eff.
COD
mg/a
278
94
44
39
297
77
35-
31
238
64
33
31
252
' 85
34
32
SS
mg/a
84
4.1
1.0
84
37
2.9
1.2
69
19
-------
00
TABLE 1 (Continued)
SERL SERIES I SUPPLEMENTARY DATA3
Run
19
21,22
23,24
25,26
27
Date
1970
9-16
9-17
9-18
9-19
9-20
Treatment
Pri. Eff.
Ppt. Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
Col. Eff.
Pri. Eff.
Pot. Eff.
Col. Eff.
COD
mg/£
238
79
70
255
77
62
256
73
62
294
61
47
SS
mg/£
65
5.6
6.4
76
5.7
6.7
89
6.7
6.6
95
7.7
5.3
Turb.
JTU
210
49
45
200
38
41
270
76
43
30
28
PH
7.3
8.1
8.2
7.3
8.0
8.0
7.4
7.9
8.0
7.3
7.8
7.8
Tot. P
mg/£
10.8
0.3
0.3
11.1
0.2
0.2
11.1
0.2
0.2
13.9
0.3
0.2
Tot.
Sol. P
mg/£
8.7
0.3
0.3
8.7
0.2
0.2
8.7
0.2
0.2
11.3
0.2
0.2
NH3-N
mg/£
18.5
18.6
0.28
19.4
18.9
0.48
20.8
16.9
0.19
21.6
21.0
0.32
17.5
0.20
Na
mg/£
51.8
104
51.8
106
48.8
115
50.6
110
61.7
123
K
mg/£
9.8
4.1
10.2
2.0
9.6
1.6
9.4
1.1
11.3
4.7
Ca
mg/£
48
70
50
52
76
58
50
75
46
38
71
52
43
60
40
Mg
mg/£
7
3
3
2
2
4
4
2
3
10
3
5
5
2
3
Treatment: primary sedimentation, chemical precipitation at pH 11.0, ammonia sorption.
-------
TABLE 1 (Continued)
SERL SERIES I SUPPLEMENTARY DATAa
Run
28,29
30,31
32,33
34,35
36
Date
1970
9-22
9-23
9-24
9-25
9-26
Treatment
PH. Eff.
Ppt. Eff.
Col. Eff.
PH. Eff.
Ppt. Eff.
Col. Eff.
PH. Eff.
Ppt. Eff.
Col. Eff.
PH. Eff.
Ppt. Eff.
Col. Eff.
PH. Eff.
Ppt. Eff.
Col. Eff.
COD
mgA
272
94
90
247
72
68
265
70
64
256
84
73
201
70
51
SS
mg/£
76
13
4.4
66
5.9
4.5
79
3.0
3.0
22
1.5
3.3
37
1.4
8.8
Turb.
JTU
310
44
32
310
80
32
310
49
29
310
18
29
PH
7.2
7.6
7.7
7.3
7.7
7.9
7.6
7.3
7.6
7.5
7.3
7.5
7.4
7.3
7.4
Tot. P
mg/£
12.6
0.2
0.2
10.6
0.1
0.1
12.6
0.2
0.2
10.8
0.4
0.3
Tot.
Sol. P
mg/£
9.8
0.2
0.2
8.2
0.1
0.1
9.6
0.2
0.2
8.3
0.4
0.3
NH3-N
mg/£
17.2
18.8
0.19
20.5
17.9
0.15
22.8
20.0
0.36
21.2
19.5
0.25
24.8
19.4
0.62
Na
mg/i
58.0
128
48.3
113
52.0
102
47.9
106
K
mg/£
8.8
0.6
9.0
0.6
9.5
1.6
8.8
1.2
Ca
mg/£
82
51
47
83
58
48
88
78
47
77
64
43
64
60
Mg
mg/a
3
3
4
3
4
4
4
4
5
4
5
8
4
5
treatment: primary sedimentation, chemical precipitation at pH 11.0, ammonia sorptlon.
-------
TABLE 1 (Continued)
SERL SERIES I SUPPLEMENTARY DATA9
Run
38
39,40
41,42
43,44
45,46
Date
1970
9-29
9-30
10-1
10-2
10-3
Treatment
Pri. Eff.
Ppt. Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
Col. Eff.
COD
mg/a
251
110
102
233
94
80
275
103
90
245
94
68
235
76
47
SS
mg/£
62
4.4
5.8
6.7
4.1
8.8
84
11.5
7.6
71
3.1
5.5
72
3.8
7.9
Turb.
JTU
270
38
50
340
90
51
290
36
35
37
25
PH
7.4
7.8
7.8
7.4
7.2
7.4
7.4
7.3
7.6
7.3
7.3
7.6
7.3
7.4
7.5
Tot. P
mg/£
11.7
1.0
0.7
10.2
0.9
0.6
11.3
0.7
0.6
10.4
0.6
0.5
Tot.
Sol. P
mg/a
9.4
0.1
0.7
8.7
0.8
0.6
9.4
0.7
0.5
8.6
0.5
0.5
NH3-N
mg/a
16.4
17.6
0.28
20.7
20.0
0.24
19.8
20.8
0.38
19.6
20.2
0.42
16.6
18.5
0.37
Na
mg/a
56.3
80.5
49.5
99.0
51.8
126
53.0
111
53.0
113
K
mg/a
8.8
1.1
8.4
0.9
8.8
0.7
8.4
0.9
8.0
0.8
Ca
mg/a
50
54
56
41
50
53
40
48
27
42
53
34
46
50
39
Mg
mg/a
4
5
8
9
8
11
8
8
11
8
4
4
7
7
6
Treatment: primary sedimentation, chemical precipitation at pH 9.5, ammonia sorption.
-------
TABLE 2
SERL SERIES II SUPPLEMENTARY DATA3
00
Run
47,48
49,50
51,52
53
54
55
Date
1971
2-17
2-18
2-19
2-20
2-21
2-23
Treatment
Pri. Eff.
Ppt. Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
Col.. Eff.
Pri. Eff.
Ppt. Eff:
Col. Eff.
COD
mg/a
198
98
100
216
100
95
196
83
87
218
68
70
278
77
69
296
92
115
SS
mg/£
46
1.4
3.0
51
1.1
2.0
42
3.3
2.9
84
5.5
9.6
101
6.8
6.8
102
5.0
3.6
Turb.
JTU
230
230
24
16
240
240
130
330
PH
7.5
7.3
7.6
7.4
7.0
7,5
7,4
7.0
7.,5
7.6
7.1
7.4
7 . 4
6". 7
7.2
7.4
7.0
7.4
Tot. P
mg/a
9.4
0.7
0.5
9.2
0.6
0.5
Tot.
Sol. P
mg/a
8.0
0.7
0.5
7.9
0.6
0.5
NH3-N
mg/£
19.7
19.8
3.4
19.5:
17.9
1.16
20.0
18, 2
1.89
20.9
20.2
0.98
Z4.5
20.8
0.60
17.4
18.6
0.67
Na
mg/£
50.6
84.0
58. 6b
62.1
100
49.4
109
105
48.3
112
59.8
112
K
mg/£
10.5
3.5
8. 6b
10.2
1.6
8.6
1.6
1.6
7.8
0.4
10.2
0.5
Ca
mg/a
56
70
62 ..
.45
66
57
54
65
53
54
66
59
43
62
55.
41
69
62
Mg
mg/a
4
3
5
11
6
8
,,4
5
6
5
4
5
7
6
5
11
4
5
aTreatment: primary sedimentation, chemical precipitation at pH 11.0, ammonia sorption.
Sample composited over period 2-17 to 2-21.
-------
00
ro
TABLE 2 (Continued)
SERL SERIES II SUPPLEMENTARY DATAe
Run
56,57
58,59
60,61
62,63
64
66
Date
1971
2-24
2-25
2-26
2-27
2-28
3-1
Treatment
Pri. Eff.
Ppt. Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
Col. Eff.
COD
mg/£
238
75
77
242
80
81
236
68
88
250
78
71
276
82
81
SS
mg/Ji
67
4.8
2.8
59
4.2
4.4
66
3.8
7.2
66
5.8
8.8
108
7.6
11.0
90
4.8
15
Turb.
JTU
290
330
360
310
340
300
PH
7.4
7.4
7.7
7.5
7.3
7.7
7.4
7.0
7.6
7.2
7.2
7.3
7.1
7.1
7.4
7.2
7.3
7.8
Tot. P
mg/Ji
9.8
0.7
0.6
8.5
0.5
0.5
Tot.
Sol. P
mg/£
8.4
0.5
0.5
7.8
0.4
0.3
NH3-N
mg/£
18.1
18.3
1.26
20.6
20.6
2.38
21.6
22.0
1.11
22.8
21.9
1.42
23.1
22.0
0.53
21.1
22.4
4.06
Na
mg/£
55.2
105
55. 2b
49.4
100
54.0
106
52.9
126
58.6
102
K
mg/£
9.0
0.8
8.6b
10.2
0.8
8.2
1.6
9.0
0.4
9.4
1.6
Ca
mg/fc
45
75
67
52
74
62
40
76
74
34
80
35
35
86
58
32
94
44
Mg
mg/£
7
3
3
5
3
5
12
4
3
11
3
8
11
1
12
2
treatment: primary sedimentation, chemical precipitation at pH 11.0, ammonia sorption.
Sample composited over period 2-23 to 2-28.
-------
TABLE 3
EBMUD SUPPLEMENTARY DATA9
00
CO
Run
1
2
3
4
5
6
7
Date
1970
11-26
11-27
11^30
12-1
12-2
12-3
12-4
Treatment
Pri. Eff.
Ppt. Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
Col. Eff.
COD
mg/Si
346
166
127
311
116
109
200
92
51
266
94
98
273
112
84
409
113
95
414
158
99
SS
mg/a
83
26
12
70
17
8.0
88
30
7.7
180
8.1
13.6
100
7.9
6.6
190
14
5.4
227
58
4.2
pH
7.0
7.2
7.2
8.1
7.1
7.0
7.5
6.9
6.9
7.3
7.1
7.3
7.1
7.6
7.3
7.0
7.2
7.1
7.2
7.2
7.1
Tot. P
mg/£
4.0
2.3
0.9
5.6
0.9
0.9
4.4
0.8
0.5
7.1
0.8
0.7
7.5
0.7
Tot
Sol. P
mg/£
2.7
1.7
0.7
3.7
0.6
0.7
1.9
0.3
0.3
3.7
0.4
0.3
3.7
<0.1
NH3-N
mg/£
12.3
10.7
1.33
14.8
14.3
3.2
5.7
5.3
0.41
8.8
8.5
0.32
9.7
8.5
0.26
9.2
8.8
0.31
9.7
8.1
0.51
Na
mg/a
110
110
92.0
131
78.2
120
K
mg/a
12.9
2.7
12.5
2.3
12.1
2.7
Ca
mg/a
31
96
96
38
84
98
39
90
43
120
92
33
88
96
36
110
86
Mg
mg/a
9
2
6
8
9
6
10
10
9
6
8
15
14
5
11
5
4
treatment: primary sedimentation, chemical coagulation, filtration, ammonia sorption,
-------
TABLE 3 (Continued)
EBMUD SUPPLEMENTARY DATA9
CO
Run
8
9
10
11
Date
1970
12-8
12-9
12-10
12-11
Treatment
Pri. Eff.
Ppt. Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
Col. Eff.
COD
mg/a
353
309
196
352
175
161
374
182
165
453
197
175
SS
mg/£
90
136
18
72
5.0
2.5
84
81
12
120
15
8.1
PH
7.0
7.4
7.4
Tot. P
mg/a
7.0
0.7
0.5
7.6
0.8
0.8
Tot.
Sol. P
mg/£
4.7
0.3
0.3
3.3
0.2
0.3
NH3-N
mg/a
12.6
14.4
0.29
13.2
12.6
0.83
16.1'
16.5
0.62
15.2
15.8
2.19
Na
mg/a
117
177
105
136
129
192
117
184
K
mg/a
21.1
3.5
25.0
3.5
23.5
3.5
27.4
3.5
Ca
mg/a
33
97
87
37
100
97
44
94
32 .
110
96
Mg
mg/£
13
4
6
11
3
8
4
17
5
4
treatment: primary sedimentation, chemical precipitation-, filtration, ammonia sorption,
-------
TABLE 3 (Continued)
EBMUD SUPPLEMENTARY DATA3
O3
cn
Run
12
13
14
15
Date
1970
12-14
12-15
12-16
12-17
Treatment
Pri. Eff.
Ppt. Eff.
A.C. Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
A.C. Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
A.C. Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
A.C. Eff.
Col. ff
COD
mg/a
22
339
162
57
76
352
169
100
242
140
72
56
SS
mg/a
84
9.9
8.5
2.4
147
8.4
1.7
2.1
93
14
3.0
3.4
70
27
9.1
3.3
PH
7.3
8.2
8.4
8.3
7.1
7.9
7.9
8.0
7.2
8.0
8.1
8.0
7.2
7.2
7.9
8.4
Tot. P
mg/a
7.5
0.6
0.5
0.4
4.5
1.4
0.7
0.4
Tot.
Sol. P
mg/a
4.6
0.5
0.4
0.4
2.8
0.8
0.8
0.3
NH3-N
mg/a
16.1
0,19
16.7
15.1
14.8
0.56
-12.6
13.3
0.31
10.5
11.2
11.1
0.35
Na
mg/a
138
184
143
184
117
193
95.5
168
K
mg/a
10.6
3.5
18.8
3.1
20.7
3.5
17.2
1.6
Ca
mg/a
35
.104
75
47
44
36
102
105
70
Mg
mg/a
8
9
4
7
3
8
3
4
3
Treatment: primary sedimentation, chemical precipitation, filtration, carbon
tion. ammonia sorotion.
sorption, ammonia sorption.
-------
TABLE 3 (Continued)
EBMUD SUPPLEMENTARY DATA9
Run
16
17
18
19
Date
1970
12-19
12-20
12-21
12-22
Treatment
Pri. Eff.
Ppt. Eff.
Resin Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
Resin Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
Resin Eff.
Col. Eff.
Pri. Eff.
Ppt. Eff.
Resin Eff.
Col. Eff.
COD
mg/£
243
135
90
77
243
86
67
60
147
49
34
28
247
134
64
81
SS
mg/£
52
32
7.7
11
86
17
3.3
2.8
76
16
8.2
4.6
Turb.
JTU
112
46
23
15
64
17
DH
7.3
7.3
7.2
7.3
7.4
7.6
8.1
8.3
7.0
7.9
7.8
7.9
Tot. P
mg/£
6.4
0.7
1.6
1.3
4.9
1.0
0.8
0.7
Tot.
Sol. P
mq/£
3.6
0.3
1.3
1.0
2.7
0.6
0.6
0.5
NH3-N
mg/£
11.3
9.7
0.76
12.0
12.0
11.5
0.25
8.1
8.3
8.6
0.35
10.5
9.7
9.8
0.55
Na
mg/£
102
126
107
138
78.2
113
106
133
K
mg/£
19.1
1.1
16.8
2.0
6.2
1.6
Ca
mg/£
41
104
98
95
36
91
82
42
98
90
52
126
116
106
Mg
mg/£
9
4
4
5
9
6
4
3
4
2
3
3
2
2
treatment: primary sedimentation, chemical coagulation, filtration, macroporous resin
sorption, ammonia sorption.
-------
TABLE 4
CCCSD SUPPLEMENTARY DATA3
00
Run
1
2
3
4
5
6
Date
1971
4-26
4-27
4-28
4-29
4-30
5-1
Treatment
Part. Set. Sew.
Ppt. Eff.
Col. Eff.
Part. Set. Sew.
Ppt. Eff.
Col. Eff.
Part. Set. Sew.
Ppt. Eff.
Col. Eff.
Part. Set. Sew.
Ppt. Eff.
Col. Eff.
Part. Set. Sew.
Pot. Eff.
Col. Eff.
Part. Set. Sew.
Pot. Eff.
Col. Eff.
COD
mg/£
77
68
262
72
63
320
64
330
68
52
284
74
54
260
104
58
SS
mg/A
9.6
8.7
118
11
7.5
141
9.1
142
6.5
6.9
123
9.6
6.8
108
33
9.3
PH
6.8
6.9
7.2
6.7
6.8
7.5
6.8
7.3
6.5
6.9
7.2
6.3
6.8
7.4
9.1
9.0
Tot. P
mg/£
12.7
1.5
1.3
13.3
1.0
0.6
Tot.
Sol. P
mg/£
10.5
0.9
0.9
9.8
0.2
0.5
NH3-N
mg/A
18.5
1.18
21.1
19.1
0.52
18.8
20.2
0.26
23.1
20.1
0.25
23.5
19.3
0.27
18.0
22.9
9.3
Na
mg/ji
95
143
112
98
138
113
113
117
115
139
102
102
139
103
113
138
K
mg/£
10.9
3.9
10.9
10.9
3.1
11.7
12.5
11.7
11.7
3.9
11.7
11.7
3.9
12.1
11.7
2.3
Ca
mg/£
62
53
64
74
78
50
72
57
71
76
54
71
74
57
70
48
Mg
mg/£
11
8
8
4
1
14
2
9
3
5
11
4
3
11
3
4
treatment: partial sedimentation, chemical coagulation at pH 10.5-10.8, filtration,
ammonia sorption.
-------
TABLE 5
BIOCHEMICAL OXYGEN DEMAND
mg/£
Location
SERL-I
SERL II
EBMUD
CCCSD
Date
8-28-70
9-3-70
9-17-70
9-24-70
10-1-70
2-18-71
2-25-71
3-8-71
11-25-70
12-2-70
12-11-70
12-16-70
4-27-71
4-30-71
PH. Eff.
gga
121a
108
121
127
92
96
121
260
185
217
167
112C
130C
Ppt. Eff.
3
5
30
32
65
43
26
47
136
56
127
112&
29
20
Col. Eff.
3
2
28
27
40
49
30
43
123
48
121
99
29
14
Primary effluent received activated sludge treatment
prior to chemical precipitation; BOD of activated sludge
effluent equalled 53 mg/A on 8-28-70 and 24 mg/x, on 9-3-70.
Precipitated effluent passed through 3 ft of activated
carbon prior to ammonia sorption; BOD of carbon effluent
equalled 85 mg/Ji.
cPartially settled sewage.
188
-------
TABLE 6
ORGANIC PLUS AMMONIA NITROGEN
mg/2,
Location
SERL-I
SERL II
EBMUD
CCCSD
Date
8-29-70
9-3-70
9-18-70
9-22-70
9-30-70
2-18-71
2-25-71
11-24-70
12-1-70
12-3-70
12-8-70
12-10-70
12-15-70
12-17-70
12-20-70
12-22-70
4-25-71
4-27-71
4-29-71
4-30-71
Pri. Eff.
33,8a
31. 2a
29.9
27.7
28.0
28.1
30.0
34.6
18.1
23.4
31.4
32.9
27.6
20.0
21.3
20.2
35. 6d
32. 8d
36. 4d
36. 7d
Ppt. Eff.
18.1
18.7
23.5
22.0
23.2
24.6
23.5
31.1
15.7
14.4
23.9
24.9
22. 5b
16. 9b
18. 8C
17. 1C
22.4
29.4
24.9
23.8
Col. Eff.
6.1
7.5
5.3
2.1
3.4
5.5
5.0
10.2
4.5
3.1
7.4
6.6
2.7
3.1
2.8
3.5
9.0
3.6
2.9
2.4
Primary effluent received activated sludge treatment prior
to chemical precipitation; Org.-N + NH3-N of activated sludge effluent
equalled 20.5 mg/£ on 8-29-70 and 20.2 mg/jj, on 9-3-70.
Precipitated effluent passed through 3 ft of activated
carbon prior to ammonia sorption;Org.-N + NH3-N of carbon effluent
equalled 18.5 mg/£ on 12-15-70 and 15.5 mg/£ on 12-17-70.
cPrecipitated effluent passed through 3 ft of macroporous
resin prior to ammonia sorption; Org.-N + NH3-N of resin effluent
equalled 18.1 mg/£ on 12-20-70 and 15.5 mg/£ on 12-22-70.
Partially settled sewage.
189
-------
SELECTED WATER
RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
I. Report No.
2.
4. Title
Optimization of Ammonia Removal by Ion Exchange
Using Clinoptilolite,
7. Authoi(s)
Koon, John H., and Kaufman, Warren J.
9. Organization
California, University of; Berkeley, Sanitary Engineering
Research Laboratory
3. Accession No.
w
5. Report Date
6.
8. Performing Organization
Report No.
10. Project No.
11. Contract/Grant No.
17080 DAR
13. Type of Report and
Period Covered
12. Sponsoring Organization Environmental Protection Agency
75. Supplementary Notes
SERL Report No. 71-5, September 1971, 189 + ix p, 35 fig,29 tab, 105 ref,
5 append.
16. Abstract
The zeolite ion exchanger clinoptilolite was investigated with the objective of
optimizing its application to ammonia removal from wastewaters. The study included
multiple cycle pilot plant operations at three municipal sewage treatment plants. Parti-
cular attention was given to cation interference with exhaustion performance and with
minimum cost regeneration.
The ammonia capacity of clinoptilolite was found to be nearly constant over the pH
range of 4 to 8, but diminished rapidly outside this range. In regeneration the pH was
critical in determining the NaCl requirements, a higher pH favoring lesser amounts of
salt. However, at a pH over 12.5 zeolite attrition became excessive and exchanger makeup
contributed significantly to operating costs.
An average ammonia removal of 95.7% was obtained in demonstration studies on three
municipal wastes having an NHs-N content of about 20 mg/n. The cost of ammonia removal
using clinoptilolite for a 10-mgd plant operating under these conditions was estimated
to be $0.082/1000 gal. Ammonia removals down to less than 0.5 mg/£ NH3-N is technically
feasible, but only with shorter exhaustion runs and greater regenerant requirements.
This report was submitted in fulfillment of Grant No. 17080 DAR between the
University of California and the Environmental Protection Agency. Partial support was
provided by the University.
i?a.Descriptors *cafjon exchange, *Ammonia, *Zeolites, *Tertiary Treatment, Nitrogen,
Waste Water Treatment, Water Reuse, Cost Analysis.
*Clinoptilolite, *Advanced Wastewater Treatment, Chemical-Physical Treatment
of Wastewater.
17c.CO WRR Field & Group Q5D
18. Availability
19. Security Class.
(Report)
20. Security Class.
(Page)
21. No. of
Pages
22. Price
Send To :
WATER RESOURCES SCIENTIFIC INFORMATION CENTER
U.S. DEPARTMENT OF THE INTERIOR
WASHINGTON, D. C. 20240
Abstractor John H. Koon
\mstitution California, University of; BerkelPV
WRSIC 102 (REV. JUNE 1971)
SP 0 9 13.261
6U.S. GOVERNMENT PRINTING OFFICE: 1972 484-486/273 i-3
------- |