Environmental Protection
Agency
Environmental EPA-eOQ/2-?@-Q§6
Laboratory June 197B
Cincinnati OH 4S268
Research and Development
High Rate Nutrient
Removal for
Combined Sewer
Overflows
Bench Scale and
Demonstration
Scale Studies
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-78-056
June 1978
HIGH RATE NUTRIENT REMOVAL FOR COMBINED SEWER OVERFLOWS
Bench Scale and Demonstration Scale Studies
by
C. B. Murphy, Jr., Orest Hrycyk, William T. Gleason
O'Brien & Gere Engineers, Inc.
Syracuse, New York 13221
Grant No. S-802400
Project Officer
Chi-Yuan Fan
Storm and Combined Sewer Section
Wastewater Research Division
Municipal Environmental Research Laboratory (Cincinnati)
Edison, New Jersey 08817
MUNICIPAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed by the Municipal Environmental Research
Laboratory, U.S. Environmental Protection Agency, and approved for
publication. Approval does not signify that the contents necessarily
reflect the views and policies of the U.S. Environmental Protection Agency,
nor does mention of trade names or commercial products constitute endorse-
ment or recommendation for use.
-------
FOREWORD
The U.S. Environmental Protection Agency was created because of in-
creasing public and governmental concern about the dangers of pollution to
the health and welfare of the American people. Noxious air, foul water,
and spoiled land are tragic testimony to the deterioration of our natural
environment. The complexity of that environment and the interplay between
its components require a concentrated and integrated attack on the problem.
Research and development is that necessary first step in problem
solution and it involves defining the problem, measuring its impact, and
searching for solutions. The Municipal Environmental Research Laboratory
develops new and improved technology and systems for the prevention,
treatment, and management of wastewater and solid and hazardous waste
pollutant discharges from municipal and community sources, for the
preservation and treatment of public drinking water supplied and to
minimize the adverse economic, social, health, and aesthetic effects of
pollution. This publication is one of the products of that research; a
most vital communications link between the researcher and the user
community.
One of the many areas of concern in the water pollution field is the
effect of combined sewer overflows on the water quality of the receiving
streams. This report is the culmination of a study which piloted a unique
concept in treating combined sewer overflows for the removal of the
macronutrients, phosphorus and ammonia nitrogen. The high rate physical/
chemical system consisted of in-line chemical addition, mixing, coagulation
and flocculation, filtration and ion exchange. The treatment system
described herein, lends itself to situations where land availability is at
a minimum and storage or conveyance of combined sewer overflows to existing
treatment sites is impractical.
Francis T. Mayo
Di rector
Municipal Environmental Research
Laboratory
-------
ABSTRACT
A high rate physical/chemical treatment system has been evaluated for
the removal of suspended solids and the macronutrients, phosphorus and
nitrogen, from combined sewer overflow. The system utilized a single unit
process concept consisting of in-line chemical addition, coagulation,
flocculation, high rate filtration and ion exchange.
The results of this program have demonstrated that the simultaneous
removal of suspended solids, phosphorus, and ammonia-nitrogen from a
combined sewer overflow is feasible using the high rate unit process concept.
Suspended solids removal ranged from 90 - 100 % with alum and polymer
dosages of 220 mg/1 and 1 mg/1, respectively.
Alum dosages resulting in Al/P molar ratios greater than 1.4 were
effective in reducing the total inorganic phosphorus (TIP) concentration
from greater than 10 mg/1 to less than 0.9 mg/1.
The clinoptilolite was effective in reducing the ammonia level to
below the detectable levels of 0.02 mg/1 NHs-N. Influent NHo-N concentra-
tions ranging from 0.2 to 16 mg/1 were reduced below detectable levels during
the initial period of contact. The effective exchange capacity of clinop-
tilolite was found to range from 0.20 to 0.64 meq NH.^-N/g clinoptilolite as
the influent NH3-N concentration ranged from 7.5 to 16 mg/1, respectively.
This study was jointly supported by the United States Environmental
Protection Agency, Grant Number S-802400, and the Department of Drainage
and Sanitation of Onondaga County, New York.
Bench scale work was commenced in January 1974. Construction of
demonstration scale facilities began shortly after the completion of the
bench scale work in 1975. The demonstration scale facility was started up
in June 1975 and terminated in August 1976.
IV
-------
CONTENTS
Foreword i i i
Abstract iv
Figures vi
Tables vii
Abbreviations and Symbols viii
Acknowl edgement ix
1. Introducti on * 1
2. Conclusions 3
3. Recommendati ons 7
4. Bench Scale Study - Results and Discussion
Medi a Selection 8
Coagulation & Flocculation Studies 9
Clinoptilolite Evaluation 10
Phosphorus Removal 14
Ammoni a Removal 15
5. Demonstration Pilot Plant-Experimental Design & System
Description
Design Philosophy 21
Equipment Description 25
6. Demonstration Pilot Plant-Results and Discussion
Combined Sewer Overflow Application 30
Dry Weather Creek Fl ow Treatment 47
7. Regeneration Evaluation
Clinoptilolite Regeneration 49
Alum Regeneration 53
8. Projected System Costs 55
References 60
Glossary 62
-------
FIGURES
Number Page
1 Ammonia Removal Efficiency 16
2 Effect of Clinoptilelite Depth on Ammonia Removal Efficiency 17
3 Effect of Clinoptilolite Particle Size on Ammonia Removal
Efficiency 19
4 Effect of Clinoptilolite Particle Size and Depth on Ammonia
Removal Efficiency 20
5 Maltbie Street Site Plan 26
6 Process Flow Diagram 27
7 Microwedge Strainer Assembly for Filters 29
8 Suspended Sol ids Removal Run #2 30
9 Suspended Sol ids Removal Run #5 32
10 Suspended Solids Removal Run #22 32
11 Suspended Solids Removal Run #28 33
12 Suspended Solids Removal Run #36 33
13 Percent Suspended Solids Removal vs Influent Suspended Solids 34
14 Percent Suspended Solids Removal vs Influent Suspended Solids 35
15 Eff 1 uent vs Appl i cation Rats 36
16 Percentage of Phosphorus Remaining 37
17 Phosphorus Removal 37
18 Phosphorus Removal Run #9 38
19 Percentage of Phosphorus Remaining 38
20 Residual Phosphorus Concentration 39
21 Ammonia-Nitrogen Removal Breakthrough Curve 40
22 Ammonia-Nitrogen Removal Breakthrough Curve.First Te'sting" s'eVYeV.*.'.'.'. 41
23 Ammonia-Nitrogen Removal Breakthrough Curve Second Testing Series 42
24 Ammonia-Nitrogen Removal Breakthrough Curve Third and Fourth
Testing Series ; 45
25 Zeolite Regeneration System 49
26 Ammonia Recovery from Thermal Regeneration of Clino 51
27 X-ray Diffraction Patterns of Clinoptilolite.... 52
VI
-------
TABLES
Number Page
1 Effective Exchange Capacity of Clinoptilelites From Various
Deposits 4
2 Polyelectrolyte Evaluation 9
3 Effective Exchange Capacities of Clinoptilolite Samples 12
4 Effective Exchange Capacities of Clinoptilolite Samples 12
5 Effect of Pretreatment on the EEC of Clinoptilolite 14
6 Effect of Clinoptilolite Depth on Ammonia Removal Efficiency 16
7 Effective Exchange Capacity of Clinoptilolite Under Constant
Flow Conditions 18
8 Clinoptilolite Particle Size Distribution After Backwashing 24
9 Velocity Gradients for High and Low Energy Mixers 28
10 Sel ected Data for Combi ned Sewer Overf 1 ow Runs 31
11 Additional Analyses for A Combined Sewer Overflow 46
12 Effective Adsorption Capacity of Thermally Regenerated Clino 51
13 Operating Characteristics For a 10 MGD Combined Sewer Overflow
Nutrient Removal Facility 56
14 Capital Costs For 10 MGD Nutrient Removal System for CSO 56
15 Annual Cost of Nutrient Removal Facility 58
vn
-------
LIST OF ABBREVIATIONS AND SYMBOLS
ABBREVIATIONS
BV/hr Bed volume per hour
cfm Cubic feet per minute
clino Clinoptilolite
CSO Combined sewer overflow
F Coli Fecal coliform bacteria
fps/ftp Feet per second per foot
gpm/ft Gallons per minuted per square foot
Kg/cm2 Kilogram per square centimeter
KWK 2 Kilowatt hour
1/min/m Liter per minute per square meter
m3/min Cubic meters per minute
meq/g Mi 11iequivalents per gram
mgd Million gallons per day
mg/1 Milligrams per liter
ml Mi Hi liter
mm Millimeter
N Normality
psi Pounds per square inch
SS Suspended solids
TIP Total inorganic phosphorus
TOC Total 9rganic carbon
EEC Effective ammonia exchange capacity
SYMBOLS
Al
A1(OH)3
A1P04
A12(S04)3
Ba
Ca
C02
Cr
Cu
Fe
HC1
HCOg
Aluminum
Aluminum hydroxide
Aluminum phosphate
. 14 H20 Alum
Barium
Calcium
Carbon dioxide
Chromium
Copper
Iron
Hydrochloric Acid
Bicarbonate
H2S04 Sulfuric Acid
K Potassium
Mg Magnesium
Na Sodium
NaCl Sodium Chloride
NaOH Sodium Hydroxide
Na3P04 Sodium phosphate
NH3-N Ammonia as nitrogen
P Phosphorus
P04-3 Phosphate
PgOs Phosphorus pentoxide
Zn Zinc
viii
-------
ACKNOWLEDGMENTS
This study was cosponsored by the U.S. Environmental Protection
Agency (USEPA) and the County of Onondaga. A great deal of support
and technical assistance was received from Storm and Combined Sewer
Section (Edison), Municipal Environmental Research Laboratory,
Cincinnati, U.S. Environmental Protection Agency. Special thanks
are due to Richard Field, Chief, and Chi-Yuan Fan, USEPA Project
Officer and Staff Engineer of the Section, for their helpful
guidance over a long period of time.
The project was administrated under the direction of John M.
Karanik, Grant Director, County of Onondaga, and Frank J. Drehwing,
Vice President, O'Brien and Gere Engineers Incorporated. Technical
supervision was provided by C. B. Murphy, Jr., Managing Engineer,
and Orest Hrycyk, Project Manager.
The bench-scale studies were performed in the pilot testing
area of the offices of O'Brien and Gere Engineers, and the assistance
of all personnel who performed countless chemical analyses is acknow-
ledged.
IX
-------
SECTION 1
INTRODUCTION
Most older cities in the United States have a combined sewer system
that serves dual purposes: (1) continuous conveyance of domestic sewage
from households, commercial and industrial institutions, and (2) conveyance
of intermittent storm runoff during times of precipitation, snow melt, and
street washings. During times of heavy precipitation, the increased flow
in the sewers can result in severe hydraulic overloading of the wastewater
treatment plant. The result of this overloading can be reduced treatment
efficiency and the possibility of a biological system washout.
To prevent such occurrences, combined sewer systems were designed with
system relief points which divert the excess flow from the conveyance and
treatment system to .a holding basin temporarily or the nearest body of
surface water. This places a severe wet weather pollutional load on the
receiving streams. Since precipitation cleanses homes, cars, streets, com-
mercial, industrial, and agricultural areas, stormwater runoff contains
substantial quantities of nutrients, organics, salts, solids, heavy metals,
oils, pesticides, herbicides, etc. Several studies which characterized the
chemical and physical nature of stormwater runoff have singled out untreated
stormwater overflows as a significant threat to surface water quality. (1)
(2) (3)
Organic material, both soluble and insoluble, exerts an oxygen demand
in the receiving stream which can create oxygen limiting conditions in the
water column, thereby impairing the level of aquatic life. Excess quantities
of organics, and more importantly, excess inorganic nutrients can accelerate
the onset of eutrophication, seriously disrupting the character of the
aquatic system. Increases in suspended matter, heavy metals, and in
the bacterial population all exert a detrimental effect on the receiving
stream.
This study is primarily concerned with the evaluation of a single unit
process for the removal of suspended solids and the macronutrients (phospho-
rus and nitrogen) from the combined sewer overflows.
Uttormark, et al. (4) estimate that runoff from urban areas has a
higher nutrient concentration than runoff from other non-point sources such
as agricultural areas, forests, and wetlands. Murphy (5) in a study of the
nutrient levels in Onondaga Lake, Syracuse,. N.Y., and its tributaries
estimates that of the total inorganic phosphorus entering the lake, 71
percent is from municipal continuous discharges and 19 percent is from
municipal intermittent (wet weather) sources such as combined sewer overflows.
-------
Phosphorus has long been suspected as a limiting nutrient for sustaining
algal growth. In lake waters slight increases in the phosphorus concen-
tration are known to overstimulate the growth of algae resulting in algal
blooms. However, these nuisance blooms can be controlled and in some cases,
eliminated. Murphy (5) noted that a reduction of the total inorganic
phosphorus concentration in Onondaga Lake from 1.74 mg/1 to 0.74 mg/1
resulted in a decline and eventual elimination of the bluegreen APHANIZOMENON.
which in previous years was dominant in the late summer and early fall
growing seasons. It was also noted that as a result of reducing the phospho-
rus concentration, the phytoplankton diversity index increased, reflecting
a higher degree of stabilization within the phytoplankton community.
Nitrogen, like phosphorus,is one of the macronutrients required for
algal growth. However, only a few species of algae can satisfy their
requirement by direct use of molecular nitrogen. Typically, most algae use
ammonia, nitrite or nitrate,as a nitrogen source for their synthesis of
proteins and other forms of organic nitrogen. As in the case with phospho-
rus, an overabundance of nitrogen (ammonia, nitrite, or nitrate) in a
surface water can trigger an algal bloom.
The objective of the project was to determine the effectiveness and
feasibility of dual use of a high rate physical chemical process for removal
of the suspended solids, phosphorus, and ammonia nitrogen from combined
sewer overflows and creek flow during wet-weather and dry-weather, re-
spectively.
The project consisted of two phases: (1) a bench scale study, and (2)
a demonstration scale study. Both studies used in-line coagulation and
flocculation, high rate filtration and ion exchange. Due to several oper-
ational restrictions at the demonstration site, the larger scale study used
the two media in separate contactors rather than in a single contactor.
The soluble and particulate phosphates were precipitated and entrained in an
alum floe and removed along with other suspended and coagulated matter in
the primary filter medium, while the soluble ammonia and short chain organic
nitrogen species were removed by ion exchange mechanisms in the secondary
filter medium, a natural occurring zeolite, clinoptilelite.
The demonstration scale study was designed to verify the results of
the bench study and to develop treatment unit costs.
-------
SECTION 2
CONCLUSIONS
1. The results of this program have demonstrated that the sequential
removal of ammonia and phosphorus from a combined sewer overflow is
feasible using a high rate unit process. The phosphorus is removed
via alum addition, coagulation, and filtration. Ammonia removal is
accomplished via an exchange/adsorption mechanism using a natural
occurring zeolite, clinoptilolite, which also serves as a polishing
filter.
2. The in-line addition of alum and polymer and subsequent in-line
coagulation, flocculation and high rate filtration, at a surface
application rate of 407 1/min/m2 (10 gpm/ft2), through a dual
media process consisting of granular coal material and a naturally
occurring zeolite, produced a high quality effluent. An alum
dosage of 110 mg/1 when combined with a polymer dosage of 1.0 mg/1
was capable of reducing the influent suspended solids level from 50
to 400 mg/1 to below 40 mg/1.
3. The added alum was effective in reducing phosphorus levels in the
applied wastewater. An aluminum to phosphorus molar ratio of 1.4
or greater resulted in effluent phosphorus concentrations below 0.5
mg/1.
4. High rate filtration of the combined sewer overflow with alum addition
resulted in several additional benefits. Test results indicated that in
addition to the reductions in phosphorus and suspended matter, reductions
in TOC, heavy metals and fecal coliform could also be expected. TOC
reductions averaged 50%, whereas, heavy metal reductions averaged 99+%
for chromium and iron, and 50% for zinc and copper. In a number of cases,
a 2.5 log reduction in fecal coliforms was also realized.
5. Effective backwashing was achieved by preliminary air scouring at
0.11 m3/min (4 cfm) for 8 minutes. Backwashing at 814 1/min/m2 (20 gpm/ft2)
was commenced during the final three minutes of air scouring and
continued for 15 minutes. With the use of air scour, 4 to 8 percent
of the filtered volume was required for backwashing.
6. Clinoptilolite, a naturally occurring zeolite, is capable of reducing
the ammonia levels in combined sewer overflows to less than 0.02
mg/1 ammonia nitroaen (NHs-N). Utilizing fresh clinoptilolite
influent NH3-N concentrations as low as 0.2 mg/1 can be reduced to
below the detectable limits during the initial contact period.
-------
Clinoptilelite's ammonia removal capacity was determined to be a
function of the NH3-N concentration in the contacting solution. The
effective exchange capacity of clinoptilolite was found to range from
0.20 to 0.64 meq NH^-N/g of clinoptilolite as the influent NHa-N
concentration ranged from 7.5 mg/1 to 16 mg/1 NH3-N, respectively.
7. Improvement in the effective exchange capacity (EEC) of clinoptilolite
was observed when decreasing the flow rates from 14 BV/hr to 2.8 BV/hr.
The EEC's were determined to be 0.64 and 0.42 meq NHs-N/g clinoptilolite
for the 2.8 and 14 BV/hr flow rates, respectively. Although the flow
rate decreased by a factor of 5, there is only a 48% increase in EEC.
This percentage increase in the EEC does not justify designing a system
at the lower flow rate.
8. The clinoptilolite has the ability to reduce any high peaks in the
influent NF^-N concentration even after the effective exchange capacity
has been exceeded. However, ammonia leaching from the clinoptilolite
occurs when the influent NHs-N concentration drops below the equilibrium
concentrati on.
9. Considerable variability in the effective exchange capacity is exhibited
by various types of clinoptilolite. Evaluations of samples from eight
deposits in western United States revealed a wide variation in the EEC.
EEC values are presented in Table 1. Hector clinoptilolite (San
Bernardino County, CA), was used exclusively in this project, since it
was the only clinoptilolite that was commercially available in the
quantities required.
TABLE 1: EFFECTIVE EXCHANGE CAPACITY OF CLINOPTILOLITES FROM VARIOUS DEPOSITS
Effective Exchange Capacity
Origin of meq NH3-N/g
Clinoptilolite Sample Clinoptilolite
1.
2.
3.
4.
J5.
6.
7.
8.
Malheur County, Oregon
Owyhee County, Idaho
Grant County, New Mexico
San Bernardino County, California (Hector)
Lander County, Nevada
Fremont County, Wyoming
Maricopa County, Arizona
Washoe. County, Nevada
0.390
0.343
0.331
0.311
0.250
0.187
0.162
0.146
NH^-N cone, of contacting solution = 21 mg/1
10. The effect of chemical pretreatment was evaluated for five of the eight
clinoptilolite samples listed in Table 1. Clinoptilolite samples 1 and
7 in Table 1 exhibited an increase in the EEC after treatment with 6N HC1
followed by a caustic brine rinse, whereas, samples 2, 4 and 8 exhibited
a decrease in the EEC.
-------
11. Phosphorus removal results indicated that in-line static mixers are
effective in providing the necessary high energy mixing required for
the initial aluminum and phosphorus contact.
12. A1:P molar ratios above 1.4 were necessary to consistently obtain
effluent TIP concentrations within the range of 0.2 to 0.9 mg/1.
Effluent TIP concentrations increased dramatically as the A1:P molar
ratio decreased below 1.4.
13. Treatment of creek water resulted in ammonia and phosphorus removal
efficiencies similar to those obtained when treating a combined sewer
overflow.
14. Thermal regeneration of clinoptlolite is a feasible alternative to
chemical regeneration. The optimum regeneration temperature is 500°C
to 540°C with a regeneration time of one hour. Lower temperatures do
not result in complete clinoptilolite regeneration and temperatures
greater than 550°C appear to cause some structural alterations which
affect the ammonia exchange capacity.
15. Thermal regeneration of clinoptilolite is an energy intensive operation,
and as such, is more expensive than conventional regeneration. It is
estimated that the energy cost for thermal regeneration of clinoptilolite
(with 40% moisture) adds an additional $0.023/m3 ($0.088/1000 gal) to
treatment costs for a wastewater containing 10 mg/1 NHs-N. According
to Koon and Kaufman (6), the least cost of chemical regeneration was
achieved when the regenerant was reused. This cost was estimated to be
$0.013/m3 ($0.050/100 gal) for a wastewater containing 20 mg/1 NH3-N.
Reducing the influent NH3(N) concentration to 10 mg/1 would increase the
resultant cost.
Thermal regeneration can be cost-effective if a source of waste heat,
such as that liberated from a carbon regeneration facility, is in
reasonable proximity to a treatment system using clinoptilolite.
16. Clinoptilolite which was subjected to 3 cycles of regeneration at 500°C
and subsequent ammonia uptake still retained greater than 97% of its
original EEC.
17. Essentially, 90% of the ammonia adsorbed onto the clinoptilolite is
removed after one hour of thermal regeneration at 500°C. This was
verified by using argon gas to transport the desorbed ammonia to a trap.
18. Thermal treatment of waste alum sludge at 1000°C had no significant
effect on the amount of phosphorus that was reclaimed with the alum.
This indicates that little Pz®5 was ^ost upon thermal regeneration.
Thermally reclaimed aluminum sulfate from waste alum sludge appears to
be unsuitable for use as a coagulant where phosphorus removal is
required.
19. The high rate nutrient removal system was also found to be successful in
reducing nutrient concentrations from the dry-weather flow of a tributary
-------
Onondaga Creek, discharging to Onondaga Lake. Multiple application would
make better use of the wet-weather facilities and would provide nutrient
reductions preceding critical periods of phytoplankton production.
20. The capital costs for a nutrient removal facility designed to treat 10
mgd of CSO are estimated to be $4.9 million (ENR Construction Cost
Index 2573).
21. Assuming 100 overflow occurrences are treated per year, the annual
operating, maintenance, and amortization costs of a 10 mgd nutrient
removal facility with thermal regeneration of spent clinoptilolite are
expected to be $0.7 million.
-------
SECTION 3
RECOMMENDATIONS
1. The ammonia removal application of a wide range of natural zeolites
including clinoptilolite should be evaluated for the purpose of
establishing a source of material having a high ammonia exchange
capacity and optimal resistance to mechanical degradation.
2. A long term study utilizing the configured nutrient removal
process should be conducted to determine variability in ammonia
removal efficiency, mechanical attrition and the effects of oil
and greases and divalent metal cation species under system optimized
operating conditions.
3. The operation of the nutrient removal process should be evaluated
under a dual dry-.and wet-weather operation. The dry-weather application
should involve a tertiary treatment application on conventional
sanitary wastewater.
4. The biological regeneration of spent clinoptilolite should be evaluated.
Particular attention should be given to the definition of optimized
aeration requirements, contact time, b"4oregeneration of kinetics,and
physical attrition of the exchange capacity.
5. The thermal regeneration of spent clinoptilolite should be investigated
on a scaled up basis with particular emphasis on establishing design
parameters.
-------
SECTION 4
BENCH SCALE STUDY RESULTS AND DISCUSSION
MEDIA SELECTION
Prior to any treatment evaluations it was necessary to determine the
optimum particle size of Anthrafilt (trade name for anthracite filter
media) to be used with a -20+50 mesh clinoptilolite. This particle size
of clinoptilolite was selected because it offered the highest surface arear
and exchange capacity, ,at an acceptable headless when used in a column. A
-20+50 mesh material is one which passes through a No. 20 U.S. Standard
Sieve, aperture opening 0.84 mm (0.033 in.), but is retained on a No. 50
Sieve, aperture opening 0.30 mm (0.012 in.).
Compatability evaluations were conducted by backwashing a 15.25 cm
(6 in.) by 3.65 m (12 ft.) plexiglas column filled with both media, and
observing the particle distribution pattern after settling. That anthra-
filt which produced a minimum interphase with the clinoptilolite would be
used in the bench scale testing. Anthrafilt sizes investigated were
numbers 1-1/2, 2 and 3 whose particle sizes were 0.85 to 95 mm, 2.38 to
4.76 mm and 4.76 to 7.94 mm, respectively.
These tests revealed the incompatabi1ity of the -20+50 mesh clinop-
tilolite with any of the aforementioned Anthrafilt sizes. Upon backwashing,
the lighter clinoptilolite particles sifted upward through the larger
Anthrafilt particles which were settling, resulting in an inversion of the
media. Since it was not possible to find an Anthrafilt size that was
compatable with the -20+50 mesh clinoptilolite, the clinoptilolite particle
sizes were increased and tested with the aforementioned Anthrafilt sizes.
Additional testing revealed that a -4+20 (<4.76 mm >0.84 mm) mesh clinop-
tilolite and No. 1-1/2 Anthrafilt resulted in no media inversion and
produced only a 2.5 cm (1 in.) zone of media integration after backwashing.
Although these two media were found to be compatible, it was felt
that using them in an actual filter operation would not be desirable.
The disadvantages would include the fact that the small particle sizes
found in the No. 1-1/2 Anthrafilt increase the probability of the floe
blinding the media in a shorter timeframe thereby resulting in higher
head losses and shorter filter runs. Secondly, the use of the -4+20
mesh clinoptilolite would result in a dramatic decrease in the ammonia
removal efficiency, since according to other investigations (7) it was
determined that the ammonia removal capacity of clinoptilolite is inversely
proportional to the particle size.
-------
It was for the above reasons that a search for a filter medium,
with a density less than Anthrafilt was initiated. Several synthetics
(plastics) were evaluated as to their specific gravity and compatibility
with clinoptilolite. The plastic pellets, used primarily for injection
molding and extrusions, were usually in the form of .316 cm (0.125 in.)
diameter cylinders by .316 cm (0.125 in.) long and were comparable to No.
2 Anthrafilt. Even though all the plastics were lighter than Anthrafilt,
each formed a zone of integration with the clinoptilolite. However, a
resin marketed by General Electric Plastics Division as NORYL proved to
be compatible with clinoptilolite, as it formed only a very slight
integrated interphase with -20+50 mesh clinoptilolite, and no integrated
interphase with -16+30 mesh clinoptilolite.
As a result of this investigation, the pilot runs were conducted
using Noryl instead of Anthrafilt as the primary medium and -16+30 mesh
clinoptilolite was selected as the secondary medium. The -16+30 clinop-
tilolite had an effective size of 0.615 mm and a uniformity coefficient
of 1.76.
COAGULATION AND FLOCCULATION STUDIES
Polyelectrolyte Selection
A total of 23 polyelectrolytes were evaluated for their ability to
produce a dense floe and a clear supernatant. The results from dosing
an alum floe with 10 mg/1 of various polyelectrolytes are shown in Table 2..
TABLE 2. POLYELECTROLYTE EVALUATION
Pol vmer
Dow Purifloc
A23
C31
C32
C41
Nil
N12
N17
N20
Calgon
CAT- Floe
CAT- Floe 8
CAT- Flow LV
WT 3000
Nalco 71D09
Magnifloc 905N
Supernatant
Floe Char- (after 15 min.
acteristic Settling)
Large Fibrous
Fibrous
Fibrous
Fibrous
Fi brous
Fibrous
Fluffy
Fi brous
Fluffy
Fluffy
Fluffy
Large Fibrous
Fibrous
Fluffy
Clear
Cloudy
Cloudy
Clear
Clear
Clear
Cl oudy
Clear
Cloudy
Cloudy
Cloudy
Clear
Clear
Cl oudy
Settling
Characteristic
Rapid
Floating Floe
Floating Floe
Floating Floe
Rapid
Rapid
Very Slow
Slow
Very Slow
Very Slow
Very Slow
Very Rapid
Rapid
Slow
-------
The polyelectrolytes that were evaluated at a dosage of 5 mg/1 are
listed below:
Atlasep 3A3
Calgon WT 3000
Gendrin
Magnifloc 560C
Magnifloc 836A
Nalcolyte 675H
Nalcolyte 670H
Purifloc Nil
R&H C7
At the concentration of 5 mg/1, the criterion for selecting a
polymer was based on the resulting floe characteristics and its settling
rate, since all of the polymers, except for Magnifloc 560C, resulted in
a clear supernatant after a 20 minute settling period. "The three
polymers with the fastest settling floe, in decreasing order were:
Atlasep 3A3, Nalcolyte 71D09 and Calgon WT 3000. However, lowering the
final poylmer concentration to 2 mg/1, produced cloudy supernatants in
all of the samples except the one containing the Atlasep 3A3.
On the basis of the above information, Atlasep 3A3 was chosen over
the other polyelectrolytes.
Alum Feed Concentration
Three concentrations of alum feed were evaluated: 10, 5 and 1%. The
flocculation energy of the system was varied by changing the length of.
coagulation coil and/or the flow rate. Varying the concentration of the
alum feed solution appeared to have no effect on the formation of floe <
particles, nor on the efficiency of phosphorus removal. Similar conclusions
were arrived at when analyzing the data concerning the energy, of floe-
culation. No related change in TIP removal or floe formation was noticed
in response to a change of the flocculation energy.
This leads one to conclude that there are other more determining
factors which influence the efficiency of TIP removal and floe formation
during alum addition, coagulation, and flocculation. Alternately,.if
these parameters do affect coagulation and flocculation, they do not do
so in the range of concentrations and values investigated in this study.
CLINOPTILOLITE EVALUATION
Clinoptilolite Selection
Since Clinoptilolite1s ammonia removal efficiency is dependent on
its effective ammonia adsorption capacity, it is important to determine
those Clinoptilolite deposits having the highest ammonia adsorption
capacity. Upon the acquisition of this information, the ammonia exchange
system can be optimized. To date many of the studies in the literature,
10
-------
including this bench and demonstration scale study, have utilized clinop-
tilolite mined near Hector California. Although the Hector deposit is the
only one that is commercially active, there are several known sizable
deposits in this country. Therefore, it becomes necessary that work be
done to establish information concerning the various clinoptilolite
deposits. Preliminary investigations to determine the effective ammonia
exchange capacity for various clinoptilolite samples were conducted.
Eight samples of clinoptilolite were obtained from different areas of
western United States. The samples.were from Grant County, New Mexico;
Owyhee County, Idaho; Fremont County, Wyoming; Lander County, Nevada;
Washoe County, Nevada; Malheur County, Oregon; Maricopa County, Arizona;
and San Bernardino County, California, (Hector).
The samples were sieved to -30 +40 mesh (standard screen scale) and
then washed with demineralized water to remove any fine material. After
drying at 105°C for 18 hours, the samples were again sieved to a -30 +40
mesh size distribution.
In order to determine an effective exchange capacity for each of the
eight samples, static equilibrium tests were performed utilizing an
equipoise shaker. Two grams of each sample were placed in a one-liter
container to which 500 ml of primary wastewater (previously passed through
a 200 mesh screen) was added. In all tests the ammonia-nitrogen concen-
tration of the wastewater was approximately 20 mg/1. The applied waste-
water was analyzed for NH^-N, TKN, K, Ba, Ca, Na, Fe and Mg. .The clinop-
tilolite sample and applied wastewater were shaken on the equipoise shaker
for two hours after which a sample of the wastewater in each container was
taken and quantitatively analyzed for each of the above parameters. The
clinoptilolite was then filtered and contacted with an additional 500 ml
of wastewater. A maximum of three repetitions of the above procedure was
conducted. The resulting integrated reduction in NH3-N yielded the
effective adsorption capacity for each clinoptilolite sample.
The data from the static tests indicated that a substantial
difference exists in the effective exchange capacities exhibited by clinop-
tilolite samples acquired from the various deposits. Results of the first
and second series of tests are presented in Table 3 and 4, respectively.
Although influent NH3-N concentrations did vary slightly from 18 mg/1 to
21 mg/1 between the two tests, the relationship between the clinoptilolite
exchange capacities did not show any variation. The clinoptilolite'
samples which showed the highest effective exchange capacity were obtained
from Malheur County, Owyhee County, and Grant County. Those samples
exhibiting the lowest effective exchange capacity were obtained from
Maricopa County and Washoe County. The Malheur County sample had an
effective exchange capacity which was 170% greater than that determined
for the sample obtained' from Maricopa County in the first series and over
160% greater in the second series. In the initial exchange capacity
evaluations, the Malheur County sample had a measured effective exchange
capacity of approximately 0.377 meq NH3-N/g compared to only 0.140 meq
NH3~N/g of clinoptilolite acquired from Maricipa County deposits. In the
second series of evaluations only two runs (two hours each) were made, and
11
-------
as a result of the reduced contact time, the clinoptilolite samples did not
exhibit their optimum effective exchange capacity. However, the higher
applied NHs-N concentration of 21 mg/1 resulted in an increase in the
exchange capacity of the Malheur County clinoptilolite which was determined
to be 0.390 meq/g. The Maricopa County clinoptilolite sample exhibited
an exchange capacity of 0.162 meq/g under the higher NHs-N wastewater
concentration.
TABLE 3. EFFECTIVE EXCHANGE CAPACITIES OF CLINOPTILOLITE SAMPLES
Series 1 Concentration NHs-N applied - 18 mg/1 NHs-N
NH3-N applied per contact - 9 mg
Total NHs-N applied - 27 mg
meq NH3-N/g
Sample mg NH3-N Adsorbed Clinoptilolite
Malheur County
Grant County
Hector Clinoptilolite
Owyhee County
Lander County
Fremont County
Maricopa County
Washoe County
10.57
8.05
8.01
6.92
6.31
5.24
3.90
3.80
0.377
0.288
0.286
0.247
0.225
0.187
0.139
0.140
TABLE 4. EFFECTIVE EXCHANGE CAPACITY OF CLINQPTILOLITE SAMPLES
Series 2 Concentration NHs-N applied - 21 mg/1
NHs-N applied per contact - 9 mg
Total NHs-N applied - 27 mg
meq NHs-N/g
_ Sample _ mg NHj-N Adsorbed _ Clinoptilolite
Malheur County 10.92 0.390
Owyhee County 9.61 0.343
Grant County 9.28 0.331
Hector Clinoptilolite 8.72 0.331
Lander County 7.00 0.250
Fremont County 5.25 0.187
Maricopa County 4.55 0.162
Washoe County 4.10 0.146
12
-------
Regarding the other parameters tested, it was observed that the
clinoptilolite samples having the highest effective ammonia exchange
capacity exhibited a tendency to desorb sodium while the samples with the
lowest capacities adsorbed sodium. One could conclude that the natural
occurring clinoptilelites in the sodium form may be the most effective
NH3-N exchangers.
The results indicated that for the eight clinoptilolite samples tested,
four, Malheur, Grant, Owyhee, and Hector, had higher effective capacities
while those of Maricopa, Washoe, Fremont, and Lander had substantially
lower exchange capacities.
Clinoptilolite Activation
Clinoptilolite is rarely found in a pure state when it is mined;
there are usually some impurities associated with it. Depending on the
origin of the clinoptilolite, the impurities may consist of metal oxides,
carbonates or even other zeolites. The presence of these impurities can
impair clinoptilelite's ability to remove ammonia from a contacting solution.
The removal of other zeolites from clinoptilolite can be very difficult,
and costly; however, a number of other impurities can possibly be eliminated
by chemical treatment.
The effect of chemical pretreatment was evaluated for five clinopt-
ilolite samples obtained from different deposits located in western United
States. The procedure consisted of contacting 2 g samples of clinop-
tilolite with a 2N solution of hydrochloric acid, (HC1), followed by treat-
ment with a caustic-brine solution (0.5 g/1 HaOH + 20 g/1 Nad).' Following
this treatment, the samples were rinsed with deionized water and then
subjected to effective exchange capacity evaluations.
The results from the activation tests are presented in Table 5. In
some cases pretreatment increased the effective exchange capacity as was
the case in clinoptilolite samples obtained from deposits in Malheur County,
Oregon and Maricopa County, Arizona. The most significant increase in the
effective exchange capacity was observed with the sample from Maricopa
County, Arizona. Its effective exchange capacity almost doubled (0.162 to
0.309 meq NHs-N per g clinoptilolite) after treatment with caustic-brine
solution. The.Malheur County sample exhibited an increase of approximately
22 percent (0.390 to 0.475 meq NHa-N per g clinoptilolite) upon activation.
The remaining clinoptilolite samples from deposits contained in Owyhee County,
Idaho; Washoe County, Nevada; and Hector clinoptilolite from San Bernandino
County, California showed a slight decrease in their effective exchange
capacity after the caustic-brine chemical treatment. The most
significant decrease in the effective exchange capacity occurred for the
Hector clinoptilolite sample which exhibited a decrease from 0.311 to
0.273 meq NHs-N per g of clinopt4-lolite. This represents a 12
percent decrease in the effective exchange capacity.
13
-------
TABLE 5. EFFECT OF PRETREATMENT ON THE EEC OF CLINOPTILOLITE
10.9 mg NH3-N Per Contact Period
Total NH3-N Applied: 21.8 mg
DEPOSIT
LOCATION
EEC VALUE
PRETKEATMENT
BEFORE
AFTER
Malheur
Owyhee
Hector
Maricopa
Washoe
0.390
0.343
0.311
0.162
0.146
0.475
0.330
0.273
0.309
0.125
PHOSPHORUS REMOVAL
Successful phosphorus removal in the single unit process design was
dependent on the combination of the effects of both direct and indirect
factors. The direct factors are considered those associated with the
precipitation and complexation of phosphorus and its eventual incorporation
into an aggregated alum floe, whereas, the indirect factors are associated
with the effective destabilization and subsequent removal of the floe.
Included in the direct factors are alum dosage, mixing and floccupation
energies, pH, temperature and the phosphorus concentrations associated
with the influent wastewater. Indirect factors include parameters such
as media size, bed depth, rate of filtration and influent suspended
solids concentration.
Influent phosphorus concentrations of 4.3 mg/1 were effectively and
consistently reduced to levels below 0.10 mg/1 with the addition of
either 100 mg/1 or 200 mg/1 of alum. This is not surprising since the
resultant aluminum to phosphorus (A1:P) molar ratios of 2.1 and 4.3,
respectively, are equal to or greater than the A1:P molar ratio of 2.0 that
most investigators have found to be required for optimum phosphorus removal.
An alum dosage of 50 mg/1 was just as effective as the 100 mg/1 and 200
mg/1 dosages in reducing phosphorus levels below 0.10 mg/1. Based on
influent concentrations of 2.6 mg/1 and 2.0 mg/1, the resultant A1:P
molar ratio is computed to range between 1.7 and 2.3, respectively.
Alum dosages of less than 50 mg/1, which would have resulted in even
lower A1:P ratios, were not evaluated.
Mixing and flocculation energies are known to influence the
effectiveness of alum in removing phosphorus. Since the system under
investigation relied entirely on in-line addition and mixing of alum
with subsequent in-line flocculation, the energies of mixing and
flocculation were particularly important and dependent on flow rate.
While the system operated in a decreasing flow mode of operation, the
effect of these energies could be readily evaluated by monitoring the
effluent phosphorus concentrations at the commencement and termination
14
-------
of the runs. Effluent phosphorus concentrations at terminal flow rates
were usually lower than the effluent phosphorus concentrations at the
initial flow rates in spite of a decrease in the mixing and flocculation
energies induced by decreased flow rates.
AMMONIA REMOVAL
By analyzing the ammonia removal efficiency under conditions of the
various runs, it was possible to verify the effect of system variables
such as polymer dosage, alum dosage, primary medium depth and clinoptilolite
depth on the ammonia removal process. In addition, three tests were
developed to specifically determine the effect of particle size on the
ammonia removal capacity of the clinoptilolite.
In determining the ammonia removal capacity, clinoptilolite depths of
.38 m (1.25 ft) and .76 m (2.5 ft) were compared. Still another test
compared the ammonia removal capacities of 2 different particle sizes.
The smaller mesh clinoptilolite used throughout the entire study, had an
effective size of .615 mm and a uniformity coefficient of 1.76, while
the larger mesh had an effective size of 1.47 mm and a uniformity coefficient
of 1.56. These tests were run with no additions of either alum or
polymer, in order to preclude the chances of the column plugging up with
floe and thereby interrupting the test.
Analysis of the data revealed that the ammonia removal mechanism is
unaffected by either alum concentration, polymer concentration or the
primary medium depth. Similarly, the saturation of the clinoptilolite
layer with floe particles, as is the case during a solids breakthrough
condition, does not appear to inhibit the effectiveness of the clinop-
tilolite as a result of the blinding of surface or pore structure
active sites.
The clinoptilolite performed reasonably well under the conditions
of a high application rate and a low contact time. In spite of these
unusual opeating conditions, initial effluent ammonia concentrations
below 0.2 mg/1 were consistently recorded. Typical flow rates for
exchange-adsorption mechanisms are in the order of 10 bed volumes/hour;
however, in the course of this study, depending on the clinoptilolite
bed depth, typical flow rates ranged from 30 to 60 bed volumes/hour.
As is evidenced in Figure 1 the ammonia removal efficiency at the
termination of a typical test run was BQ% or greater in the majority of
tests. This indicates that the rate of increase of headless far exceeds
the exhaustion rate of the clinoptilolite in the course of a normal run.
This suggests that several backwashings of the primary filter medium may
be in order prior to clinoptilolite regeneration. No conclusions could
be drawn regarding the effect of clinoptilolite depth on ammonia removal
efficiency, because the removal efficiencies for all depths were quite
similar, indicating minimal exhaustion as witnessed in Figure 1.
15
-------
o
Oh.
'r*> °-
o
r
E «-
o>
o:
ief
Q>
fcj
£
o..
cvj
1.2m (4ft.) clmo
• 6m(2ft)clino
Constants;
Application Rate'- 529.1 i/min/sq. m.
(13 gpm/ft)
Alum Dosage - 25 mg/1
Polymer Dosage - 2 mg/1
Qk-
TIME (HRS)
1234
EIGURF 1. AMMONIA REMOVAL EFFICIENCY _
To overcome this problem a numoer of tests were specifically ran to
determine the effective exchange capacity of the clinoptilolite. These
tests indicated that the removal efficiency varied directly with the
clinoptilolite bed depth as tabulated in Table 6.
TABLE 6. EFFECT OF CLINOPTILOLITE DEPTH ON AMMONIA REMOVAL EFFICIENCY
Constants: Primary Media Depth
Application Rate
Alum Dosage
Polymer Dosage
2.44 m (8.0 ft.)
530 l/min/m2 (13 gpm/ft2)
0.0 mg/1
0.0
4
Run
Time
(Hr.)
1.25 ft Clinopti
Throughput
lolite
Volume NHs-N
liters (Gal.) (mg/l)
% NH3-N
Removed
0
.5
1.0
1.5
2.5
3.5
4.5
5.0
5.5
0(0)
65.9 ( 17.4)
130.2 ( 34.4)
196.4 ( 51.9)
317.6 ( 83.9)
442.1 (116.8)
567.0 (149.8)
629.1 (166.2)
691.9 (182.8)
9.00
0.76
2.45
3.96
5.71
6.72
7.23
7.38
7.54
0.0
91.6
72.9
56.1
36.7
24.4
19.8
18.0
16.2
2.5 ft Clinoptilol
Throughput
, . Vol ume , .
liters (Gal.)
0(0)
67.4 ( 17.8)
135.5 ( 35.8)
204.4 ( 54.0)
351.6 ( 92.9)
497.0 (131.3)
635.9 (168.0)
702.5 (185.6)
766.1 (202.4)
NH3-N
(mg/1
9.00
0.19
0.19
0.28
1.27
1.81
4.13
4.64
4.94
ite
RpmnypH
0.0
97.9
97.8
96.9
85.9
68.8
54.1
48.4
45.1
16
-------
The advantages of this greater clinoptilolite bed depth are readily
apparent in Table 6 which summarizes the data. It is evident that at
the onset of the first run the effluent NHs-N concentration is far
lower in that for the run using the greater clinoptilolite bed depth.
Secondly, a lower effluent NHs-N concentration is estimated for a
greater period of time. For example the volume of wastewater processed,
before an effluent NH3-N concentration of 1.0 mg/1 is obtained, is
approximately 322 1 (85 gal) and 76 1 (20 gal) for the .76 and .38 m
(2.5 and 1.25 ft) depths of clinoptilolite, respectively. This is more
than a fourfold increase in volume throughput for a twofold increase
in bed depth. As the tests proceed, however, and both samples
become more exhausted, this difference in removal efficiency between the
two bed depths narrows. This can be better visualized in Figure 2, which
shows the percent ammonia removal for both bed depths converging as the
tests proceed.
o ••
'+
§•
T° «
X 10
o ? +
c 3
I
fi a-
o ,..
.76m (2.5ft.) clino
.38m < 1.25 ft.) clino
Constants:
Application Rate - 529.1 1/min/sq.m (13 gpm/ft )
Media Depth - 2.4 m (8 ft)
Alum Dosage - 0 mg/1
Polymer posage - 0 rng/1 , ,
I
2 3
TIME (HRS.)
FIGURE 2. EFFECT OF CLINOPTILOLITE DEPTH ON AMMONIA REMOVAL EFFICIENCY
17
-------
In addition these two tests were used to determine the effective
exchange capacity (EEC) of the clinoptilolite, under the conditions of
an application rate of 530 1/min/m (13 gpm/ft ), an empty bed contact
time of approximately 2.5 minutes and an influent NHs-N concentration
of 9.0 mg/1. The EEC was based on a 50 percent reduction of the influent
NHs-N concentration rather than after complete exhaustion of the
clinoptilolite. The results of these calculations are presented in
Table 7.
TABLE 7. EFFECTIVE EXCHANGE CAPACITY OF CLINOPTILOLITE (EEC) UNDER
CONSTANT FLOW CONDITIONS
(Based on 50% NH3-N Reduction in Influent)
Application Rate: 530 1/min/m2 (13.6 gpm/ft2)
Clinoptilolite Size: -16+30 mesh
Effective Size: 0.615 mm
Uniformity Coefficient:1.76
Clino
Depth
meters (ft)
.38 (1.25)
.76 (2.50)
Clino Throughput
Wt. Vol. @ 50% Red
(9)
1325
2625
0)
227
681
(qal)
60
180
Flow
. (Bed Vol.
per hour)
83.6
41.8
NH3-N Effective
Removed Exchange Capacity
(mg) (meq NHvN/q Clino)
1530
4990
0.086
0.140
It is worth mentioning that typical exchange or adsorption processes
proceed at flow rates of 10 to 15 bed volumes per hour. In the above
two cases the tests were allowed to proceed at a surface application rate
of 530 1/min/m (13 gpm/ft). The flow rates (BV/hr) through the clinop-
tilolite portion of the system were substantially greater than one would
expect, as is shown in Table 7. In comparing the EEC of the two tests it
is anticipated that reducing the flow rate through the clinoptilolite media
to 10 BV/hr may further increase the EEC.
The result from tests evaluating the effect of clinoptilolite
particle size on ammonia removal efficiency showed that the two varied
inversely. A graphic representation of the data is presented in Figure 3.
18
-------
o
oo
o
r-
O
10
0>
0>
Q.
-I6t30 mesh clino
2 3
TIME (HRS.)
FIGURE 3. EFFECT OF CLINOPTILOLITE PARTICLE SIZE ON AMMONIA REMOVAL EFFICIENCY
Although the run with the larger clinoptilolite size had a one foot
depth advantage over the smaller clinoptilolite size, it nevertheless
had 25-30 percent lower ammonia removal capability throughout the course
of the run. Ficmre 4 illustrates the significance of particle size in an
exchange or removal process. In spite of the fact that the contactor
volume represented by the larger particle size clinoptilolite was almost
three times the volume of that used in testing the smaller size clinop-
tilolite., its removal efficiency was inferior, which subsequently led to its
quicker exhaustion.
19
-------
(015)
I
(A
E
s
en
« (.OIO)
13
0)
§
0)
CC
I (.005)
6.81
.4.54
2.27
.76m (2.5ft.)-16+30
.39m(l.25ft)-|6t30
*-
-------
SECTION 5
DEMONSTRATION PILOT PLANT EXPERIMENTAL DESIGN AND SYSTEM DESCRIPTION
DESIGN PHILOSOPHY
General
Analysis of combined sewer overflows and creek water indicated that
both the ammonia and phosphorus influent concentrations ranged from ap-
proximately 0.2 - 0.5 mg/1. In order to appraise the wider range of con-
centration that could exist, the wastewaters were spiked with dibasic
ammonium phosphate, (NH4)2HP04. The ammonium phosphate solution was pumped
into the influent holding tank; the turbulence in the tank was sufficient
to provide adequate mixing of the spiking solution and the wastewater. A
variable speed pump delivered the volume of solution necessary to obtain
the desired 7 to 15 mg/1 concentration of ammonia in the applied wastewater.
The phosphorus concentration was predetermined by the desired ammonia
concentration.
Four initial flow application rates were evaluated during the study:
529, 407, 326, and 65 1/min/m2 (13, 10, 8 and 1.6 gpm/ft^). The combined
sewer overflow was applied to the system at two initial rates, 529 and 407
1/min/m2 (13 and 10 gpm/ftr) whereas, the creek water was applied at all
four application rates. Application rates of 326 and 65 l/min/m2(8 and 1.6
gpm/ft2) were only used in tests evaluating ammonia nitrogen removal with
clinoptilolite. The columns were operated in the declining rate filtration
mode.
The filter was always operated in the downflow mode. On the other hand,
the clinoptilolite columns were operated in the upflow mode, except for two
creek water runs which were operated in a downflow mode.
Solids Removal
High rate filtration, 407 1/min/nr (10 gpm/f1r), was used to remove
the suspended solids from the combined sewer overflows. A previous study
reported suspended solids removals from 40 to 60 percent when using high
rate filtration ( 8). However, in this study high concentrations of sus-
pended solids in the effluent from the filter could adversely affect the
total system performance by plugging the clinoptilolite columns, particularly
if the clinoptilolite columns were operated in the downflow mode. Therefore,
high suspended solids removal efficiencies were desired. Following the
addition of alum, an anionic polymer (Magnifloc 836A) was added to increase
the strength of the alum floe so that a high degree of removal could be
obtained. The filter media was an anthracite material with an effective
21
-------
size of 6.1 mm and a uniformity coefficient of 1.3.
Both the alum and polymer were stored as a 25 percent and 0.04 percent
by weight solutions, respectively, and then pumped into the treatment
process influent stream at a rate to yield the desired dosages. Alum
dosages used during the study ranged from 50-220 mg/1; the polymer dosages
ranged from 1 - 4 mg/1. The effect of various alum and polymer dosages on
suspended solids removal, as well as, the wastewater application rate were
evaluated.
Phosphorus Removal
Phosphorus removal from the combined sewer overflow was achieved
in a two-step process; first, precipitation of the phosphorus, and second,
removal of the phosphate floe particles from the waste stream by high rate
filtration. Alum was added to the wastewater to precipitate the phosphorus.
The chemistry of phosphorus removal with alum has been presented in detail
elsewhere ( 9 ). Two equations used to illustrate the reactions which
occur when alum is added to phosphorus-containing wastewater are as follows:
(1) A12(S04)3. 14 H20 + 2 P0~3 = A1P04 + 3 S04"2 + 14 H20
(2) A12(S04)3. 14 H20 + 6 HCO~ = 2 A1(OH)3 + 6 C02+ 14 HgO + 3 SO"2
In the first equation an aluminum phosphate precipitate is formed;
actually the precipitate is probably of an intermediate composition between
the crystalline solid AlPO^ A1(OH)3 (second reaction) and an amorphous
structure. From the first reaction the aluminum to phosphorus (A1:P) molar
ratio appears to be 1:1; however, the second reaction competes with the
phosphorus for the aluminum, and is thus at least partially responsible for
the excess alum which is required. With the addition of alum the second
reaction results in the loss of alkalinity, with the subsequent pH reduction
dependent on the initial alkalinity concentration.
The solubility of the A1P04 is pH dependent. Theoretical solubility
concentrations have been calculated at various pHs. The calculations have
shown that the minimum solubility of A1P04 occurs in the pH range of 5.5 -
•6.5.
Since the combined sewer overflows were spiked with dibasic ammonium
phosphate, (NH/^ HP04, essentially all the phosphorus in the phosphorus in
the wastewater was in the soluble form, as an orthophosphate P04~3
The process variables normally optimized in phosphorus removal evalu-
ations include the applied alum concentration, and the resultant A1:P molar
ratio. Phosphorus removal with alum addition has been quite extensively
studied; however, in most of those studies the alum and phosphorus-containing
wastewater were contacted in rapid-mix tanks for several minutes and the
aluminum phosphate precipitate subsequently removed by sedimentation in a
settling tank. Since the system evaluated in the course of this study con-
sisted of in-line static mixers and filtration columns, the physical processes
were substantially different from those evaluated by other investigators, and
22
-------
the process determining variables, stated above, normally associated with
phosphorus removal may not be determining in this study.
The contact time in the in-line static mixer ranged from 0.4 to 1.2
seconds compared to several minutes in the rapid-mix tank, and the time in
the filter ranged from 2-4 minutes compared to 60 to 120 minutes in the
settling tank. Also, as the particular run proceeded, the filtration
column simulated a packed bed reactor, containing alum which was still
available to react with the phosphorus. Jenkins, et al. (9) indicated that
the initially formed precipitate tended to contain an excess of hydroxide
over the phosphate solid, thereby increasing the possibility that the
recycle and reuse of precipitated solids would be effective in increasing
phosphate removal. Therefore, the presence of an alum sludge in the filter
may remove .additional phosphorus. If this proves to be the case, then once
alum sludge is accumulated, the required A1:P molar ratio may be decreased
due to the additional available alum contained in the sludge.
As previously mentioned, the calculated velocity gradients in the
static mixers, 6085 fps per foot at 529 1/min/m2 (13 gpm/ft2) application
rate, and 1390 fps per foot at 203 1/min/m2 (5 gpm/ft2) application rate,
were much higher than the velocity gradients normally associated with
conventional rapid-mix tanks. The data was analyzed to determine if the
decreasing flow rate had any effect on the phosphorus removal, possibly
related either to decreasing velocity gradients or increasing contact time
in both the mixer and filtration column.
The alum dosages evaluated included 50, 100 and 200 mg/1. Operating
under a declining rate filtration necessitated constant modification of the
alum flow rate. This caused minor variations in the alum dosage during a
particular run. Since the alum flow rates were monitored, the actual alum
dosages were recorded and A1:P molar ratios were established.
Ammonia Removal
The natural zeolite, clinoptilolite, was used as an ion exchange agent
to remove ammonia from the combined sewer overflows. Previous studies by
others using clinoptilolite with municipal, industrial, and synthetic
wastewaters have revealed many of the variables which effect the adsorption
capacity of the clinoptilolite (6,10). Variables such as wastewater pH,
cationic strength, nature of the competing cations, ionic form of the
zeolite (H+, Na+, or Ca"1"1"), grain size of the clinoptilolite, influent
ammonia concentration, and contact time, have all been shown to effect the
adsorption capacity of the clinoptilolite.
Only the influent ammonia nitrogen concentration and contact time were
purposely varied to determine their impacts on the EEC of clinoptilolite
and the effluent quality. It was realized that the competing ion con-
centrations would vary somewhat throughout the testing; however, this
variation was not expected to be significant-enough to warrant the monitoring
of competing cations. Other factors such as the variation in the influent
ammonia levels were felt to be much more significant. In fact Koon, et al,
(6) have stated that the impact of the cationic strength will vary for
23
-------
different combinations of cations.
The clinoptilolite was received in the form of a -16 +50 mesh size (<1.19
mm(0.047 in.); > 0.3 mm (0.012 in)), with about 20% finer material. The clinop-
tilolite columns were backwashed at 582 1/min/m2 (14.3 gpm/ft2) to remove the
finer material, before each series of tests. Samples were collected from the
surface, 0.30 m (1 ft) and 0.91 m (3 ft) depths and a sieve analysis was
conducted to determine the size distribution of the clinoptilolite after
backwashing. Table 8 illustrates the particle size distribution pattern
after backwashing. Prior to backwashing the clinoptilolite had an effective
size of 0.42 mm and a uniformity coefficient of 1.40.
TABLE 8. CLINOPTILOLITE PARTICLE SIZE DISTRIBUTION AFTER BACKWASHING
DEPTH (ft.) ' EFFECTIVE SIZE (mm)
Surface
1
3
0.35
0.42
0.59
Several influent NH3(N) concentrations were used in the various testing
series. Although throughout a particular testing run an attempt was made
to keep the influent NHs(N) concentration fairly constant, certain variations
in the influent did occur; therefore, the weighted average influent con-
centration was used for comparison between test series.
With the exception of two cases, all the runs proceeded with the
clinoptilolite columns operating in the upflow mode at flow rates already
mentioned. Although suspended solids may not have interfered with the
actual adsorption capacity of the clinoptilolite, the possibility that
breakthrough of the polymer-flocculated solids would plug the finer mesh
clinoptilolite columns necessitated operating those columns in the upflow
mode.
Tests were conducted to determine the effect of particle contact time
on the adsorption capacity of the clinoptilolite. The particle contact
time was assumed to be different from the empty bed residence time. The
empty bed residence time can be altered by increasing the bed depth. The
particle contact time is defined as the time that the dissolved NH3(N)
molecule is in contact with an individual clinoptilolite particle and
therefore is dependent on the wastewater velocity through the bed and
independent of bed depth.
Although the actual particle contact time would have been extremely
difficult to calculate, it was thought that a five-fold difference in the
velocity of the wastewater passing through two columns of the same
clinoptilolite depth could illustrate the dependence of the contact time on
the EEC of the clinoptilolite.
The first 13 individual runs were operated at an influx of 407
24
-------
1/min/m2 (10 gpm/ft2) and following alum and polymer addition the waste-
waters were applied in an upflow manner to a column containing 1.52 m (5
ft) of clinoptilolite. The second series of tests consisted of 12 individual
runs operating in an upflow mode through the clinoptilolite and at an in-
flux of 529 1/min/m2 (13 gpm/ft2). However, the second series operated
with two clinoptilolite columns each filled with 1.37 m (4.5 ft) of clinop-
tilolite and connected in series. A media height of 1.37 m (.4.5 ft) in
the clinoptilolite columns was limiting, since the columns were operated in
the upflow mode with the bed expanding to approximately 2.44 m (8 ft) at
the 529 l/min/m2 (13 gpm/ft2) influx. Operating two clinoptilolite
columns in series allowed a determination of the effective ammonia exchange
capacity from the data derived from the first column as well as the effective-
ness of the series configuration.
The third series of tests involved the application of creek water to
the clinoptilolite column in the downflow mode. Since the creek water
contained less than 40 mg/1 of suspended solids, no alum or polymer were
added, thereby eliminating the plugging of the clinoptilolite column. The
application influxes were 326 1/min/m2 (8 gpm/ft2) and 65 1/min/m2 (1.6
gpm/ft2). Since these tests were conducted for the purpose of evaluating
the effect of contact time on the effective ammonia exchange capacity, only
one clinoptilolite column was used in each test.
Sampling and Analyses
Sampling of the system was conducted both manually and automatically.
The filter influent wastewater was sampled in the Zurn screening unit effluent
tank using an automatic sampler.
The effluent was sampled with an automatic sampler at the effluent
pipe from the clinoptilolite column. A third sampling site, between the
two clinoptilolite columns, was introduced when the series configuration
utilized two clinoptilolite columns.
Analysis of the influent and effluent wastewaters included: suspended
solids.(SS), ortho and hydrolyzable phosphate (TIP), ammonia nitrogen (NH,-N)
total alkalinity (TALK), and pH. 6
The TIP, NH3-N, and TALK analyses were done utilizing a Technicon
Auto Analyzer, while the SS and pH analyses were conducted in accordance
with Standard Methods for Examination of Water and Wastewater, 13th Edition.
EQUIPMENT DESCRIPTION ,
The nutrient removal demonstration facilities were housed in an
existing combined sewer overflow treatment facility located at Maltbie
Street on the west side of Syracuse, N.Y. Combined sewer overflows from a
drainage area of 115 acres zoned for commercial and light industrial use
were treated at the Maltbie Street Facility.
The fundamental purpose of the Maltbie Street treatment facility was
25
-------
to evaluate high rate solids removal and high rate disinfection of combined
sewer overflows. The facility was equipped with a pumping station that
collected, metered and pumped the wastewater to the treatment section which
consisted of three screening units each with a different size screen aperture
and followed by a disinfection unit. The screening units were connected in
parallel and could be operated independently. Following the screening
operation, the wastewater proceeded to the disinfection chambers where it
was treated with either chlorine (012) or chlorine dioxide (CIC^) and then
discharged into the receiving stream, Onondaga Creek. A diagram of the
Maltbie Street-site is shown in Figure 5.
Nutrient
Removal
Facilities
-Interceptor Connector
to Main Interceptor
ONONDAGA
CREEK
Bypass
Overflow Regulator
Leaping Weir Type
Pump Discharge Line
Solids Concentrate Line
Siphon to Main
Interceptor
FIGURE 5. MALTBIE STREET SITE PLAN
The combined sewer overflow used in the nutrient removal system was
passed through a Zurn MicroMatic horizontal drum screen which had a screen
aperture size of 71 microns. After screening, a portion of the flow was
pumped to the nutrient removal system. The remainder of the flow proceeded
to the disinfection chambers and was subsequently discharged to the creek.
During periods of dry weather, creek water was used as the influent
wastewater to the nutrient removal system. Creek water was pumped up to the
treatment system by a Barnes submersible pump, Model No. 12 ASE 1, which was
capable of pumping 6.31 I/sec (100 gpm) at 15.2 m (50 ft) head. Although the
creek water typically had a suspended solids concentration of less than 100
mg/1, it was'nevertheless screened through the Zurn unit. After screening
it was applied to the nutrient removal system in the same manner as the
combined sewer overflow.
.<
The nutrient removal system process flow diagram is presented in
Figure 6.
26
-------
Static
Mixer
iL.
o
"5
0
c
o
3
.
*
T
)
CM
fe
^
O
c
o
0
JK
-1
M|
ro
2
0
o
c
c
>
o
i ^
FIGURE 6, PROCESS FLOW DIAGRAM
Contactor .Effluents
and
Backwash Influents
The screened effluent of the Zurn unit was pumped to- the nutrient removal
facilities via a centrifugal pump manufactured by Weinman, Model No.
6AE20S having a capacity of 4.4 I/sec (70 .gpm) at 19.8 m (65 ft) of
head The end of the intake line of the pump was fitted with a check
valve and was placed about 0.61 m (2 ft) from the bottom of the Zurn
tank. The discharge line was equipped with a bypass valve that allowed
any portion of the pumped CSO to be returned to the Zurn unit. The
bypass valve not only allowed for control over the treatment rate but it
also provided additional turbulence in the Zurn unit which, to some
extent, prevented the solids from settling.
Once the treatment rate was established for a particular run, no
further adjustments were made. The system was operated in a declining
rate filtration mode'.
Flow rate .was monitored by a Wallace and Tiernan Rotameter,' tube
#WT2-40-610 equipped with a stainless steel float #26-12rVI. The effective
range of the rotameter was between 0.32 I/sec (5 gpm) and 4.23 I/sec (67
gpm). .
Chemical addition of both alum and polymer was accomplished with a
duplex head chemical feed pump, manufactured by Madden, Model-JR-4.
Each side of the feed pump could be adjusted independently so that the
proper dosages of each chemical could be added to solution. Alum was
added first to the CSO. High energy mixing was achieved with a 5'. 1 -,cm
(2 in. PVC) in-line static mixer manufactured by Kenics, Model .No. 2-50- .
541-7. The mixer was 0.53 m (1.7 ft) long and contained three standard
helixes. Following the high-energy mix, a low energy mix, to facilitate
flocculation. was provided by installing a 30.5 m (100 ft) coil of 5.1
cm (2 in.) black polyethylene pipe. Polymer addition followed the alum
27
-------
low energy mixing coil and its mixing sequence was identical to that of
alum.
The velocity gradients for the static mixer and the mixing coil
were determined at several flux rates for both the high arid low energy
mixers and are presented in the table below.
TABLE 9. VELOCITY GRADIENTS FOR HIGH AND LOU ENERGY MIXERS
Velocity Gradient Velocity Gradient
Static Mixer Mixing Coil
How Rate (fps/ft) (fps/ft)
2..52 I/sec (40 gpm) 3960 572
0..94 I/sec (15 gpm) 1390 : 140
The above data indicate that velocity gradients of the static mixer,
even at terminal flux rates, are considerably higher than those normally
associated with rapid mix tanks, 700 fps per foot. With respect to the
.mixing coils, the data suggested that the coils provided a fluid regime
somewhere between mixing and flocculation.
Upon leaving the polymer mixing coil, the wastewater was applied to
a filter and subsequently to two ion exchange columns containing
clinoptilolite. The filter contained 1.52.to 2.13 m (5 to 7 ft) of No.
3 Anthrafilt. The particle sizes of No. 3 Anthrafilt which is manufactured
by Blue Coal Corporation, ranged from 4.76 to 7.92 mm (0.188 to 0.313 in.)
and the effective size was determined to be 6.1 mm. Clinoptilolite was
purchased from Bariod, Division of National Lead, which owns the Hector
clinoptilolite deposits in San Bernardino County, California. The
clinoptilolite was processed by the Colorado School of Mines Research
Institute at Golden, Colorado, to yield a -16 + 50 mesh material'. A
sieve analysis of the clinoptilolite as received indicated an effective
size of 0.42 mm and a uniformity coefficient of 1.40.
The filter, and ion exchange columns were constructed of 7.11 cm
(0.280 in.) plate steel and were 3.05 m (10 ft) by 0.61 m (2 ft) in
djameter. Each of the three columns were equipped with 76.2 mm ( 3 in.)
influent and effluent ports. Additionally, the columns were fitted with
a 127 mm (5 in.) flanged openinq near the top of the column for media
loading and .a 0.305 m (12 in.) flanged manway near the bottom. E«ch
column contained a perforated steel plate which was used as a media
support. Perforations were twelve 38.10 mm (1.5 in..) holes, eight equally
spaced on an 20.32 cm (8 in.) radius and four equally spaced on a 10.16
cm (4 in.) radius. The holes in the base plate were fitted with a AISI Type
304 stainless steel Microwedge strainer, manufactured by Roberts Filter
Manufacturing Company. The strainers, shown in Figure 7, which had an
opening of 0.38 mm (0.015 in.), were approximately 30.2 mm (1.188 in.) in
height and 53.98 mm (2.215 in.) diameter and were held in.place by a
toggle bolt assembly. Since the total area of the base plate was 0.29 m2
28
-------
(3.14 ft ) there was one strainer per 0.024 m2 (0.26 ft2).
I 1/4" Hole
1/16 "x 1/8" D.P.
Screwdriver Slot
5/16" x 2 3/4" 304S.S.
Hex Head Machine
Bolt and Nut
304 SS. Cap
5/16" Infernal Tooth
Lockwasher 3O4 SS.
304 S.S- Micro Wedge
Strainer X5I5" Opening
304 S.S. Strainer Base
Filter Plate "l/2*
1 1/4" Hole Drill or
Punch Only
304 S.S. Toggle
FIGURE 7. MICROWEDGE STRAINER ASSEMBLY FOR FILTERS
(Courtesy of Roberts Filter Co.)
The interiors of the columns were sandblasted and coated with two coats 0.25
mm (0.01 in.) of an epoxy enamel (Mobil Hi-Build 78 Series). The eppxy coat-
ing was specified since it would withstand the abrasion of the Anthrafilt and
clinoptilelite during backwashing. Once the exterior surfaces of the columns
were coated with two coats of high gloss enamel, they were placed into
position. The columns were connected such that they could be operated in
series or parallel. The effluent line of each column was equipped with quick
disconnect coupler which allowed for rapid conversion to an upflow mode by
connecting two columns with a section of 76.2 mm (3 in.) flexible hose.
Backwashing was achieved by connecting a flexible hose from a metered tap
water outlet to the quick disconnect coupler at the bottom of each column.
Only one column could be backwashed at a time. Similarly^each column was
equipped with a quick disconnect coupler for air scouring.
29
-------
SECTION 6
DEMONSTRATION PILOT PLANT STUDY
RESULTS AND DISCUSSION
COMBINED SEWER OVERFLOW APPLICATION
Suspended Solids Removal
The results of the suspended solids removal data are shown in Table
10 and Figures 8 through 12.
Gross suspended solids removal was accomplished as the wastewater
passed through a 73 micron Zurn Micro-Matic screening unit. However,
during the first flush, that is the first half hour to hour of a combined
sewer overflow occurrence, the effluent from that screening unit contained
SS concentrations as high as 250-400 mg/1. Since the system was not
automated to start whenever an overflow occurred, many of the runs did not
start in time to remove first flush SS. Influent SS concentrations during
the first hour of filtration varied from 90 to 250 mg/1.
Since alum addition generates suspended matter and the filter run
lengths are dependent on solids loading to the filter, the alum dosages
would affect the filter run length. However, the variability in the SS
concentration in the applied wastewater prior to chemical addition makes
the determination of the effect of alum dosages on a filter run length
difficult.
o
o
FIGURE 8.
400r
LEGEND
APPLICATION RATE-IOGPM/FT5
ANTHRAFILT DEPTH - 5 FEET
ALUM DOSAGE -25 MG/L
POLYMER DOSAGE- 4 MG/L
1.0
3.0
3.5
15 2O 2.5
RUN LENGTH (HRS.)
SUSPENDED SOLIDS REMOVAL RUN #2 (COMBINED SEWER OVERFLOW)
30
-------
TABLE 10. SELECTED DATA FOR COMBINED SEWER OVERFLOW RUNS
Run No.
2a
2
5
6
a
9
10
11
'5
22
23*
2U
35
36
Influx
1/Min/iaZ (gpui/ftZ)
407 (10)
407 (10)
407 (10)
407 (10)
407 (10)
407 (10)
407 (10)
407 (10)
407 (10)
529.1 (13)
529.1 (13)
529.1 (13)
407 (10)
407 (10)
Final Flux
1/mln/m2 (gp«i/ft2)
260.5 (6.4)
358.1 (8.8)
203.5 (5.0)
264.5. (6.5)
2U4.9 (7.0)
329.6 (8.1)
215.7 (5.3)
382.6 (9.4)
370.3 (9.1)
293 (7.2)
362.2 (8.9)
337.0 (8.3)
284.9 (7.0)
272.7 (6.7)
Run Length
(llrs.T
3.0
3.5
3.0
3.0
3.0
3.6
3,5
3.0
2.0
3.5
2.25
3.5
3.0
3.0
Aliin Dosage
, ("9/1 )
SO
SO
100
100
100
so
100
50
50
215
215
70
215
215
Polymer
Dosage
(«w/1)
4
4
2
2
2
1.5
2
2
2
1
1
2
1
1
Inf. SS
Ranne (mq/l)
10- 62
100-435
40-260
20- BO
100-288
43-156
22- 63
13- 69
28- 68
26- 64
44- 52
40-100
100-300
40-150
Eff. SS
Range (inq/1)
4- 60
8- 40
6- 16
15- 25
62-182
11- 68
2- 23
18- 30
7- 47
0- 10
0- 10
7- 30
0-1 SO
2- 60
*Run Terminated Due to pump
X Removal
Range (mq/l
67-97
0-99
74-95
0-75
34-69
56-81
60-85
0-82
0-75
81-93
85-99
38-92
20-99
50-96
Failure
Applied SS
.kn (Ib)
4.7 (10.36)
1.91 ( 4.21)
.83 ( 1.84)
2.98 ( 6.56)
1.49 ( 3.29)
.62 ( 1.37)
1.0 ( 2.20)
.54 ( 1.19)
1.18 ( 2.61)
1.03 ( 2.27)
1.55 ( 3.41)
3.22 ( 7.10)
l.SO ( 3.32)
Residual SS
Kg Ob)
.52 (1.61 )
.17 (0.39 )
.35 (0.77 )
1.52 (3.35 )
.5 (1.10 )
.19 (0.42 )
.57 (1.27 )
.39 (0.87 )
.13 (0.29 )
.10 (0.228)
. 4 (0.88 )
.78 (1.72 )
.3 (0.66 )
t
Mass Removed
89
91
58
49
66
69
43
23
89
90
74
76
BO
-------
LEGEND
APPLICATION RATE- I0gpm/ft2
ANTHRAFILT DEPTH - 5 feet
ALUM DOSAGE. - 100 mg/1
POLYMER DOSAGE- 2 mg/l
Influent
I " <\~~~ \
-5 1 1.5
RUN
— — o
2
LENGTH
2.5
(hours)
3
uem
35
FIGURE 9. SUSPENDED SOLIDS REMOVAL RUN # 5 (COMBINED SEWER OVERFLOW)
•x
at
p 200
cc
UI
u
o
o
100+
o
in
LEGEND
APPLICATION RATE'- 13 gpm/ft2
ANTHRAFILT DEPTH - 7 feet
ALUM DOSAGE - 215 mg/l
POLYMER DOSAGE- | mg/l
1.5 2 2.5
RUN LENGTH (hours)
Influent
Effluent
\
35
FIGURE 10. SUSPENDED SOLIDS REMOVAL RUN #22 (COMBINED SEWER OVERFLOW)
32
-------
o>
Q
£200
cc
t-
z
UJ
o
o
o
100- •
LEGEND
APPLICATION RATE- 13 gpm/ft2
ANTHRAFILT DEPTH- 7 feet
ALUM D.OSAGE - 70 mg/l
POLYMER DOSAGE- 2 mg/l
Influent
1.5 2 2.5 3 *5 '
RUN LENGTH (hours)
FIGURE 11. SUSPENDED SOLIDS REMOVAL RUN #28 (COMBINED SEWER OVERFLOW)
LEGEND
2
O
200-•
UJ
o
O
o
IOO--
o
UJ
o
UJ
Q.
V)
25--
APPUCATION RATE- I0gpm/ft2
ANTHRAFILT DEPTH - 7 feet
ALUM DOSAGE -215 mg/l
POLYMER DOSAGE- I mg/l
Influent
.5
I
1.5
RUN LENGTH (hours)
FIGURE 12. SUSPENDED SOLIDS REMOVAL RUN #36 (COMBINED SEWER OVERFLOW)
33
-------
The run lengths were normally 3 to 3.5 hours, and were terminated
when either the total head loss reached 10-13 psi (0.70-0.91 kg/cm2) or
there was a continuous condition of SS breakthrough in the effluent. In
cases where the termination was caused by excessive head loss, the
terminal flow rates were approximately one to two thirds of the initial
flow rates.
Effluent quality from the high rate filter was, for the most part,
below 40 mg/1 TSS. Since the SS concentration in the CSO varied widely, the
removal efficiencies fluctuated with the influent SS concentration. It
was especially difficult to obtain high SS removal efficiencies when the
influent SS were very low, 20-50 mg/1. Generally, the effluent SS
ranged from 20-40 mg/1 over the entire filtration run.
It can be seen in Table 10 under the column heading SS Specific
Removal, that there was no apparent benefit in operating the filtration
column at a depth of 2.14 m (7 ft) as compared to 1.52 m (5 ft).
During most of the runs approximately 15 to 19 m3 (4000-5000 gal) of CSO
were filtered in run lengths of 3-3.5 hours. The total mass of SS removed
ranged from 27 to 91% of the total mass of SS applied.
Combined Effect of Alum Dosage and Influent SS--
Figures 13 and 14 best illustrate the effect of the alum and polymer
dosages on filtration. Because of data scatter, no well defined optimum
dosage could be determined. However, there was a discernible trend
o
tu
o
IU
ID
O
cc
tu
Q.
100-
90.
8O.
70-
60.
50.
40.
30.
2O.
10.
•X
X,
X
•A
ALUM
X-IOO mg/1
0_ 5O mg/1
A- 50 mg/1
• X
POLYMER
2 mg/1
1.5 mg/1
4 mg/1
—i—
100
300
FIGURE 13.
200
INFLUENT SS (mg/1)
PERCENT SUSPENDED SOLIDS REMOVAL VS INFLUENT SUSPENDED SOLIDS
34
-------
100-
90.
% 80.
a 70-
9 60.
8 50.
a
til
| 40.
Ul
§> 30.
3
t; 20.
Bio.
Ul
a.
A AA A A A
, ^L ^ •*
XX AJf." A A
X * m » A
A X%
+ A A A ALUM
• X- 100 mg/l
0- 50 mg/l
A-2l5mg/l
• - 70 mg/l
A
POLYMER
2 mg/l
2 mg/l
1 mg/l
2 mg/l
100
300
200
INFLUENT SS (mg/l)
FIGURE 14. PERCENT SUSPENDED SOLIDS REMOVAL VS INFLUENT SUSPENDED SOLIDS
indicating that the lower alum dosages, 50, 70, and 100 mg/l yielded
lower quality effluent, ranging from 45 to 70% SS removal. On the other
hand, an alum dosage of 215 mg/l appeared to be very effective in removing
solids over the entire range of influent SS, 25-400 mg/l, and achieved SS
removals between 85 to 99%. This is in agreement with Packham (11) who
reported that the amount of alum required for coagulation increased as
the turbidity and suspended solids concentration in the influent decreased.
This reflects the viewpoint.that coagulation improves as the number of
suspended particles, which act as nuclei, is increased.
Figure 15 shows the effect of the flow, on the SS removal under an
applied alum dosage of 215 mg/l, which obtained the best SS removal
efficiency. The flow rates are the actual flowrates at the instant a
sample was obtained during the runs. This data would tend to indicate
that the solids removal system was least efficient at the lower flow
rates. However, since the system was operated in the declining rate
filtration mode, the lower flow rates are realized at the end of the
run when solids breakthrough was most probably occurring. Therefore,
would appear that the flow rate is a parameter which must be closely
scrutinized when evaluating the declining rate filtration system.
Phosphorus Removal
The first series of tests were operated under declining rate,
starting at 407 l/min/m^ (10 gpm/ft2) and decreasing to 203 l/min/m2
it
35
-------
V.
O)
1
i
0
CO
Q
Ul
Q
Z
Ul
0.
r>
CO
1-
Ul
u.
u.
Ul
100-
90-
80-
70-
60-
50-
40-
30-
20-
10-
LEGEND
ANTHRAFILT DEPTH - 7 feet
ALUM DOSAGE -215 mg/l
X POLYMER DOSAGE- 1 mg/l
X
X
X X
xx
X xx \**f "x X¥XX
4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0 13.0
APPLICATION RATE (gpm/ft2)
FIGURE 15. EFFLUENT SS Ms APPLICATION RATE
(5 gpm/ft?) at the termination of the run. The influent phosphorus
concentrations ranged from 2.5 to 22 mg/l TIP during this test series.
Since a synthetic phosphorus solution was added to the combined sewer
overflow, essentially all the phosphorus was in the form of ortho phosphate.
Alum dosages of 50 and 100 mg/l used during this series, gave an A1:P
molar ratio ranging from 0.3 to 3.9.
Although the data were scattered, the general trend indicated that
95% TIP removal could be obtained with an A1:P molar ratio equal to or
greater than 2.0 (Figure 16). With an A1:P molar ratio of 1.0 or greater,
the system was assured of at least a 80% TIP reduction. The data showed
that alum dosages resulting in A1:P molar ratios less than 1.0 caused
the phosphorus removal efficiency to decrease rapidly. It can be seen
in.Figure 16 that at an A1:P molar ratio of 0.7 only about 75% TIP
removal can be expected. A1:P molar ratios lower than 0.7 are impractical
for TIP removal considerations.
The plots of the individual runs indicated that very little additional
TIP removal was obtained in using A1:P molar ratios above 2.0. For
example, in Figures 17 and 18 the effluent TIP values were approximately
the same, 0.3 mg/l TIP, although the A1:P molar ratio ranged from 1.05
to 3.87. (Figure 19). The effluent phosphorus quality ranged from 0.5
to 0.9 mg/l TIP above the 1.4 A1:P molar ratio (Figure 20).
36
-------
A
701
60
50
40
30
20
15
10
9
8
7
6
5
4
3
2
1
• AA
. A °-
'D
Aa
tr LEGFMD
A RUN 5- x
RUN 6- •
A a RUN 8- A
0 RUNS- •
RUN 10- a
°*o o RUN1I- £
D° °0 RUN IS"0
° ^PRLJJAJION=IOGPM/
m
m
•
A
•
A • A
• A
A
X
f
X
x A
.5 l& 1.5 2.0 ZS 30 3,5
'FT.2
A
4.0
AKP MOLAR RATIO
FIGURE 16. PERCENTAGE OF PHOSPHORUS REMAINING (ALL RUNS COMBINED SEWER OVER-
FLOW)
LEGEND
APPLICATION RATE .lOgpm/ft2
AI/P MOLAR RATIO -
O Influent
L5 2 2.5
RUN LENGTH (HRS.)
3.5
FIGURE 17. PHOSPHORUS REMOVAL (COMBINED SEWER OVERFLOW)
37
-------
I4+
8
a.
P I.
LEGEND
APPLICATION RATE * 10 gpm/ft2
AI/P MOLAR RATIO -/X
i Influent
1.33) (1.52
(1.05'
•v_
-0-
1.5 2 2.5
RUN LENGTH (HRS.)
-©Effluent
3.5
FIGURE 18. PHOSPHORUS REMOVAL RUN # 9 (COMBINED SEWER OVERFLOW)
(9
Z
REMAIN
a.
P
*
SO
40
30
20
15-
ID-
S'
8
r
5
4-
3
1
A
A A ° D
A
RU
a ~ ~
A n 2
u APPUW
A RAT
X
X
X
X
x x
.3 1.0 15 20 23 3.0 33 4.0
1 i i i i ii i i
WG. INK
TJP CONC.
8MG/L
_ 10
28-A 4
At:P MOLAR RATIO
FIGURE 19. PERCENTAGE OF PHOSPHORUS REMAINING (COMBINED SEWER OVERFLOWS)
38
-------
5
44
3
2
.9.
S 3
.2
LEGEND
RUN
aa-x e MG/-L
23-D 10
28 -A 4
RATt
A*
1.0 LS ZQ 2.5 30 35 4.0
AKP MOLAR RATIO
FIGURE 20 RESIDUAL PHOSPHORUS CONCENTRATION (COMBINED SEWER OVERFLOW)
Minton and Carlson (12) showed the results of various researchers
in a plot of residual phosphorus versus A1:P molar ratio. The plot
indicated that the results were quite variable in the A1:P molar ratio
range of 1.0-4.0. In their presentation, an A1:P molar ratio of 2.0 or
more generally produced an effluent with a filtrate phosphorus residual of
0.5 mg/1 or less.
In Figure 20, the data indicated that the residual TIP concentration
was unaffected by the A1:P molar ratio, when that ratio was above 1.4. With
the exception of three data points all the residual TIP concentrations
ranged from 0.3 - 0.9 mg/1. This suggested that there would be no advantage
to using very high A1:P molar ratios to minimize the effluent phosphorus
concentration.
Low effluent phosphorus concentrations encountered at minimum flux at
the termination of runs reinforced the view that the in-line static mixers
provided sufficient velocity gradients tor the effective mixing of chemicals.
Ammonia Nitrogen Removal
The results of the ammonia evaluations are shown in Figures 22 through
25. These figures show the influent and effluent NHo-N concentrations
versus throughput volume for the various testing
The effective ammonia exchange capacity was based on a dry weight bulk
density that was determined to be 762 kg/m3 (47 Ibs/cf) Hector clinoptilolite.
39
-------
A preliminary set of 6 runs (runs No. 64 through 67 and No. 1 and No.
2) was made prior to the addition of any synthetic ammonia solution to the
wastewater. The depth of the clinoptilolite was 1.52 m (5 feet) having a
bed volume of 0.44 m3 (15.7 feet3). The initial application-rate was 407
l/min/m2 (10 gpm/ft2) equivalent to an application rate of 15.9 BV/hr.
In these preliminary runs the influent NH3-N concentration ranged
from 0.20 - 0.97 mg/1 NH3-N, with a weighted average influent concentration
of 0.37 mg/1 Nt-L-N as presented in Figure 21. The effluent NH3-N con-
centration, initially below detectable levels, gradually rose to 0.2 mg/1
NH3-N. The total NH3-N removed by the clinoptilolite was only 37.2
grams, which represented a partial EEC of 0.04 meq NHs-N/g clinqptilolite.
Although the column was not operated to exhaustion, this preliminary test
did show that even at very low influent NHs-N concentrations the clinop-
tilolite was very effective in the removal of NH3-N.
O.6T
LEGEND
• -INFLUENT
• -EFFLUENT
20
30 40 50
THROUGHPUT VOLUME, GALIDNS x 10 3
60
FIGURE 21. AMMONIA-NITROGEN REMOVAL BREAKTHROUGH CURVE PRELIMINARY TESTING
CURVE
40
-------
First Testing Series--
The first testing series consisted of 13 individual runs applied to a
single column of clinoptilolite. The depth of the clinoptilolite was 1.52
m (5 feet) giving a bed volume 0.44 m3 (15.7 ft3) and the initial column
application rate was 407 1/min/m 2(10 gal/ft2/min) equivalent to an applica-
tion rate of 15.9 BV/hr.
As can be seen in Figure 22, the influent NHs-N concentration was
quite variable ranging from 2.0 mg/1 to 17.8 mg/1 with a weighted average
of 7.5 mg/1 NHs-N. The wide variation in influent NHo-N concentration
reflected operational problems experienced with the synthetic nutrients
feed pump.
20
30 40 SO
THROUGHPUT VOLUME ,
60
Gallons x 10
FIGURE 22. AMMONIA-NITROGEN REMOVAL BREAKTHROUGH CURVE FIRST TESTING SERIES
A total of 159 m3 (42,000 gallons) had been applied to the column when
the effective ammonia exchange capacity was reached, which is that point in
the course of the run when the effluent and influent NHs-N concentrations were
nearly equal. At that point both influent and effluent concentrations were
nearly 10 mg/1 NHs-N (the weighted average influent NH3-N concentration
was slightly lower).
It would appear that the breakthrough to 1 mg/1 NH3-N had occurred at
38 m3 (10,000 gallons), 85 BV; however, the effluent ammonia concentrations
then decreased to as low as 0.5 mg/1 NH3-N at 76 m3 (20,000 gallons),
BV. From the 76 m3 (20,000 gallon) mark to the termination of the run"at 185
m3 (49,000 gallons) the effluent concentration gradually increased to
41
-------
approximately 10 mg/1 NH3-N. Unfortunately, the influent NH3-N concentra-
tion was too variable to yield a smooth breakthrough curve; however, the
test series demonstrated the ability of clinoptilolite to effectively
remove theNHa-N. After 185 m3 (49,000 gallons) had been applied, 964 g of
NHs-N had been subsequently removed. The effective ammonia exchange capacity
of the clinoptilolite in this testing series was 2.88 mg NHs-N/g clinop-
tilolite, which is equivalent to 0.20 meq NHs-N/g clinoptiTolite.
When the column of clinoptilolite reached its effective ammonia exchange
capacity after the application 159 m3 ( 42,000 gal Ions) of wastewater,an
additional 113 m3 (40,000 gallons) 340 BV were applied to the column. As
can be seen in Figure 22, there was no additional observed NH3-N removals,
on the contrary there appeared to be some desorption occurring.
Second Testing Series—
The second testing series consisted of 10 separate runs using two
columns of clinoptilolite in series. Both columns of clinoptilolite were
kept on-line until the second column had reached its effective ammonia
exchange capacity. The depth of clinoptilolite in both columns was .the
same, 1.22 m (4 feet) giving a media bed volume of 0.35 m3(12.4 ft3). The
initial influx rate in this testing series was 529 1/min/m2 (13 gpm/ft2)
equivalent to 25.5 BV/hr.
As can be seen in Figure 23, the influent concentration was somewhat
variable, although less than observed in the previous testing series. The
influent spiked NHo-N concentration applied to the first clinoptilolite
column ranged from'b-17 mg/1, with a weighted average influent of 9.4 mg/1
NH3-N. The influent ammonia concentration applied to the second clinop-
tilolite column was the effluent concentration from the first contactor.
A- 2nd CLINO COLUMN
•-LEAD CLINO COLUMN
•- INFLUENT
2030 4o 50 60 70 80 90 100 HO 120 130 1*40 \ko 160
THRUPUT VOLUME gd x K)3
FIGURE 23. AMMONIA-NITROGEN REMOVAL BREAKTHROUGH CURVE SECOND TESTING SERIES
42
-------
A total of 530 n^ (140,000 gallons) of wastewater was applied to
the contactors before the effluent from the second clinoptilolite column
was approximately equal to the weighted average influent concentration
applied to the same column.
The benefit of using two columns in series became readily apparent in
this testing series. In the first column breakthrough to 1 mg/1 NH3-N
occurred after 42 m3 (11,000 gallons), 117 BV had been applied to the column.
However, with the use of the second column, the volume at breakthrough from
the entire two column series was increased almost fivefold to approximately
200 m3 (53,000 gallons) or 281 BV based on the combined volumes of both
columns. While this increase was not unexpected, it illustrated the value
of using columns of clinoptilolite in a series mode of application.
When breakthrough to 1 mg/1 NHs-N occurred in the first column, 42 m3
(11,000 gallons) of wastewater or approximately 355 g (0.8 Ib) of NH3-N had
been removed by the clinoptilolite. When breakthrough at 1 mg/1 NH3-N
occurred in the second column, 200 m3 (53,000 gallons), an additional 895 g
(1.9 Ib) of NH3-N was removed by clinoptilolite in the first column, bringing
the total amount of Nl^-N removed to T250 g. This quantity of ammonia re-
moved was equivalent to 0.33 meq NHs-N/g clinoptilolite.
In operating the two columns until the effluent NH3-N concentration
from the second column was approximately equal to the influent weighted
average NH3-N concentration, the first column had 5927 g (13.1 Ib) NH3-N
applied and (3.0 Ib) NH3-N removed. Therefore, under, the conditions of this-*
testing series, the effective ammonia exchange capacity of the clinoptilolite
in the first column was 5.12 mg HH3-N/g clinoptilolite, equivalent to 0.37
meq NHs-N/g clinoptilolite, considerably exceeding that realized with the
second, column.
In comparing the quantity of NHs-N removed by the first clinoptilolite
column upon breakthrough to 1 mg/1 NHs-N to that removed by the second
column (0.09 vs 0.33 meq NHs-N/g clinoptilolite), it is readily apparent
that a much larger portion of the clinoptilolite in the first column had
reached its effective exchange capacity before the breakthrough was ex-
perienced in the second column.
This follows the pattern of any typical ion exchange system where the
breakthrough concentration front moves down through the media until the
effluent concentration then begins to approach the influent concentration.
It was apparent that the first column was at nearly 91% of its effective
ammonia exchange capacity when breakthrough occurred in the second clinop-
tilolite column. During the entire testing series the second clinootilolite
column had 4554 g (10.0 Ib)' NH3-N applied of which 1430 g (3.1 Ib) NH3-N was
removed. The effective ammonia exchange capacity of the clinoptilolite in
the second column was 5.33 mg NH3-N/g clinoptilolite, equivalent to 0.38
meq NH3-N/g clinoptilolite. Within the experimental error the two columns
had the same effective adsorption capacity.
Even though the NHs-N concentration being applied the second clinop-
tilolite column was quite low (less than 1 mg/1) during the first third of
43
-------
the testing series, as the influent NH3-N concentration to that column
increased (simulating moving to the lead column in the series), the driving
gradient became large enough to achieve the effective exchange capacity.
After the first clinoptilolite column had been exhausted, it still had
the capacity to reduce any peak influent NH3-N concentrations that were
higher than the weighted average influent of 9.4 mg/1. However, any time
the influent NH3-N concentration dropped much below 7-8 mg/1 NHs-N,
desorption was experienced.
The effluent from the second column exhibited a lower rate of NH3-N
increase breakthrough than the first clinoptilolite column. The reasan for
the rate of increase was due to low Initial influent concentrations to the
second column which gradually increased to being uniform throughout the
testing series.
Third and Fourth Testing Series —
The third and fourth testing series were evaluated together to determine
the effect of contact time on NH3-N removal capacity of the clinoptilolite.
A comparison of the results is shown in Figure 24. Both series were operated
in the downflow mode through single columns of clinoptilolite. Creek water,
spiked with synthetic nutrients, was used directly because the low suspended
solids concentration in the creek water eliminated possible column plugging.
The only parameter which was varied during the two testing series was the
application rate. The depth of the clinoptilolite bed in both testing
series was 1.37 m (4.5 ft), yielding a bed volume of 0.39 m3 (14.1 ft3).
The third and fourth testing series were operated at application rates
of 326 1/min/m2 (8 gpm/ft2), .and 65 l/min/m2 (1.6 gpm/ft2), respectively.
Although specific influent NH3-N concentrations to both columns
varied somewhat, the weighted average influent NHs-N concentrations in both
testing series was quite similar. Although the effluent NHs-N concentration
was not quite equal to the weighted average influent NHs-N concentration
at the time of termination, Koon (6) has shown that the clinoptilolite
breakthrough curve approached equilibrium (effective exchange capacity)
relatively slowly, thus the effective adsorption capacity of the clinop-
tilolite at the end of this test series was likely to be very close to
the capacity that would have been obtained had the effluent reached the
16 mg/1 NHS-N concentration. A total of 4140 (9.1 Ib) g NH3-N had been
applied to the clinoptilolite column with 2712 g (5.9 1b) NH3-N being removed.
The effective ammonia removal capacity was 9.0 mg NHs-N/g clinoptilolite,
equivalent to 0.62 meq NHs-N/g clinoptilolite. At the conclusion of the
test, the clinoptilolite was not fully exhausted. After 246 m3 (65,000 gal)
had.passed through the clinoptilolite, 3324 g (7.3 Ib) NH3-N, of 3994 a (8.8 Ib)
NHs-N applied, were removed by the clinoptilolite resulting in an EEC; of
0.79 meq NHs-N/g clinoptilolite. Since the effluent NHs-N concentration at
tne end of this test was significantly lower than the influent weighted
average of 16 mg/1 NHs-N, a significant quantity of ammonia would
probably continue to be removed by the clinoptilolite bed (assuming
that breakthrough to 1 mg/1 NH3-N represented approximately 65% of the
removal capacity), as in the third test. The effective ammonia removal
44
-------
capacity in the fourth test could then be expected to approach 13.3 mg
NH3-N/g clinoptilolite, equivalent to 0.95 meq NHs-N/g clinoptilolite.
25
^20
o»
•i
o
o
15
'10
,/
LEGEND
THIRD TEST1NG_.SERIES_
o- INFLUENT
• - EFFLUENT
EQURDi lESJlNG.SERJES.
• - INFLUENT
A- EFFLUENT
AXT
10 20 30 40 50
THRUPUT VOLUME gol.xlO3
60
FIGURE 24. AMMONIA-NITROGEN REMOVAL BREAKTHROUGH CURVE
THIRD AND FOURTH TESTING SERIES
Comparing the effect of the two application rates on the ammonia
removal capacity, particularly at breakthrough to 1 mg/1 NHs-N, it was
evident that at the lower flow rate (increased residence time) an
additional 38 m3 (10,000 gal) was passed through the column before
breakthrough occurred. Also, it would appear that additional NH3-N can
be adsorbed on the clinoptilolite at the lower flow rate. The results
show that almost a 50% increase in the adsorption capacity can be
realized with a 80% decrease in the application rate. The tests indicated
that the residence time of the wastewater in the clinoptilolite bed does
affect adsorption; however, the relative economy associated with the
higher application rate may offset the increase in-the adsorption
capacity.
The data indicated that the clinoptilolite was very effective in
removing the ammonia from the applied wastewater. The influent NH3-N
concentrations, ranging from 0.2 mg/1 to 16 mg/1, during the range of
testing series, were reduced below detectable levels in the initial
period of contact. The degradation of NH3-N effluent quality followed
the exhaustion of the available active sites.
Removal of Other Wastewater Constituents
It is common during filtration that in addition to suspended
solids, other constituents in the applied wastewater are also removed.
45
-------
If chemical addition is employed, the suspended and colloidal solids can
be swept out of solution by a "sweep floe" and the soluble constituents
can be removed from solution by coprecipitation. During one particular
run, in addition to the regularly scheduled analyses of the influent and
effluent samples, several other parameters were monitored. The parameters
included heavy metals: cadmium, chromium, copper, iron, lead, and zinc,
as well as, TOC and fecal coliform bacteria. The results of the analyses
are listed in Table 11. The influent concentration of lead and cadmium
were below the analytical detection limits, and as a result they are not
presented in Table 11.
The presented data illustrates that there is a reduction in heavy
metals concentrations, even though they are initially present in rather
low quantities. The best removal was obtained with iron, which consistently
showed removals of greater than 99%. Removal percentages for other heavy
metals ranged from 0% to > 99%, and averaged usually better than 40% removal.
The lack of correlation between the SS and the heavy metal reductions elimi-
nated coprecipitation as a possible removal mechanism for.heavy metals.
Similarly, the TOC reductions, which ranged from 28% to 75%, could not
be correlated to SS reductions. Reductions in the fecal coliform concentra-
tions in the combined sewer overflows were as high as 90%. Influent
fecal coliform levels averaged in the 500,000 colonies/100 ml range and
were reduced in many cases to below 10,000 colonies/100 ml.
The entrainment of bacteria in the alum floe and its subsequent
removal in the filter should greatly reduce the disinfectant requirements
of the CSO, should further bacterial reductions be necessary. This
would result in lower disinfectent operating costs and a reduction in
the possibility of high oxidant residual levels being discharged to the
aquatic environment.
DRY WEATHER CREEK FLOW TREATMENT
During periods of dry weather, creek water was applied to the
nutrient removal system. The original intent of this application was to
determine if the wet weather facilities could be utilized during dry
weather periods to reduce the loading of suspended solids, phosphorus
and ammonia nitrogen to Onondaga Lake. The concentrations of suspended
solids, phosphorus, and ammonia-nitrogen measured in the creek water
during the study were found to be quite low: 20-75 mg/1 SS, 0.1-0.4
mg/1 TIP, and 0.2-0.7 mg/1 NH3~N, respectively. Only the SS concentration
was found to increase to rather high levels, a maximum of 100-120 mg/1
SS following an intense rainfall. The TIP and NH3-N remained within
the previously staged range. Since the background levels of TIP and NH3-N
in the creek water were so low, dibasic ammonium phosphate was added
to increase the levels to approximately 10 mg/1 TIP and 10 mg/1 NHs-N
for the demonstration scale testing.
Of the 13 runs made using creek water only two resulted in suspended
solids removal efficiencies consistently greater than 25%. Poor SS
removal efficiencies were probably the result of low SS concentrations
in the applied creek water and the low alum dosages, 50, and 100 mg/1
46
-------
TABLE 11. ADDITIONAL ANALYSES. FOR A COMBINED SEWER OVERFLOW
Flux: 407 l/min/m2 (10 gpm/ft2) Media: #3 Anthrafilt
Alum Dosage: 215 mg/1
Run Flow Rate
Time 1pm Cr Cu Fe
Sample hr. (gpm) pH mg/1 mg/1 mg/1
Influent
Effluent
Influent
Effluent
Influent
Effluent
Influent
Effluent
Influent
Effluent
Influent
Effluent
Influent
Effluent
0.0
0.1
0.7
0.8
1.0
1.1
1.3
1.4
1.7
1.8
2.0
2.1
2.7
2.8
117.3(31) 8.1
6.9
107.9(28.5) 7.9
6.9
98.4(26) 7.7
6.8
81.4(21.5) 7.6
7.0
0.05
<0.01
0.03
<0.01
0.03
<0.01
0.03
<0.01
0.04
0.02
0.04
0.02
0.02
0.02
0.02
0.00
0.37
<0.01
0.60
<0.01
0.57
<0.01
0.57
<0.01
Media Depth: 7 ft
Polymer Dosage: 1.0 mg/1
Zn TOC TSS FColi.
mg/1 mg/1 mg/1 col /1 00 ml
0.06
0.03
0.07
0.05
0.06
0.03
0.07
0.04
32.
23.
50.
24.
55.
28.
60
14
42.
3.
50.
10.
116.
20.
50.
51.
665,000
305,000
660,000
7,400
545,000
5,850
530,000
2,553
395,000
3,750
535,000
5,850
242,700
17,850
-------
alum, used as a coagulant. Due to the dispersed and stabilized state of
the total suspended solids characteristic of the creek water large alum
dosages would be required to effect a high removal of SS. However, this
would not be a cost-effective practice. In this case, suspended solids
removal could possibly be improved by using a smaller media size in the
filters.
In both creek water runs considered to have good SS removals, the
chemical dosages were 100 mg/1 alum and 4 mg/1 polymer. It is believed
that, in these two cases, the very high polymer dosage was instrumental
in achieving the superior SS removals. Other creek water runs involved
the application of equal and even greater alum dosages but lower polymer
dosages and achieved inferior SS removals. Another factor which could
help explain the improved SS removals realized in the two runs is that,
of all the creek water applications these two had the highest influent
SS loading. In one run, the influent SS ranged from 55-115 mg/1, while
the effluent ranged from 18-30 mg/1 for a SS removal efficiency of 25-
88%. The run was 2.5 hours in length, and with the flow rate decreasing
from 1.96 I/sec (10 gpm/ft2) to 1.33 I/sec (6.8 gpm/ft2) during the run,
the total throughput volume was 14800 1 (3900 gallons). Termination of
the run after 2.5 hours was necessary because of chemical feed system
failure.
For the second run influent SS ranged from 30-70 mg/1, while the
effluent ranged from 2-6 mg/1 SS, representing a removal efficiency of
91-97%. After a 3.5 hour run length, the flow rate had decreased from
1.96 I/sec (10 gpm/ft2) to 1.14 I/sec (5.7 gpm/ft2) with a total volume
of (5200 gallons) subsequently processed through the filter.
For the remaining creek water runs SS removal efficiencies were
erratic and inconsistent. Influent suspended solids concentration in
these eight runs were lower, ranging from 5-50 mg/1 with applied alum
and polymer dosages ranging from 50-200 mg/1 and 1-2 mg/1, respectively.
Since treatment of creek water from the standpoint of SS removals
appeared to be impractical, increased alum and polymer dosages were not
applied in an attempt to increase the removal efficiency of the SS.
Phosphorus removals experienced on the creek water were more than
acceptable and were similar to the removal efficiencies experienced in
the combined sewer overflow. Analysis of the data provided no trends
unique to the creek water runs. The reason may be partially due to the
synthetic nutrient solution that was added to both the combined sewer over-
flows and the creek water. The data indicated that A1:P molar ratios
above 1.4 generally resulted in 80-95% TIP removal, equivalent to a
residual TIP concentration of 0.1 to 0.8 mg/1 TIP. As with the overflow
occurrences, an A1:P molar ratio below 1.0 yielded poor TIP removals.
The ammonia-nitrogen removal efficiencies realized in the creek water
runs were essentially the same as those experienced in the combined sewer
overflow runs. There was no indication that the ammonia-nitrogen removal
mechanisms from creek water differed in any from the ammonia removal
experienced with the overflows.
48
-------
SECTION 7
REGENERATION EVALUATIONS
CLINOPTILOLITE REGENERATION
The high cost of clinoptilolite makes the regeneration of clinoptilelite
a necessary alternative to the disposal of exhausted material. Current
regeneration technology involves the use of a caustic or caustic-brine
regenerant. However, for this process to be economically practical, the
regenerant must be stripped of ammonia and recycled. The removal of
ammonia from the regenerant is effected at an elevated pH in a stripping
tower. Because the operation of a stripping tower is limited to climatic
conditions where no freezing occurs during winter, its application is, there-
fore, not very practical in the northeastern and midwestern U.S. An alternate
method of clinoptilolite regeneration was investigated.
Thermal Regeneration
Thermal regeneration evaluations were conducted on Hector clinop-
tilolite samples at temperatures of 300°C, 400°C, 500°C, and 600°C in a
laboratory rotary tube furnace (Linberg). Figure 25 shows a schematic
diagram of the furnace apparatus.
Stoinless Steel Tube
Containing Clinoptilolite
Argon Gas Supply
Regeneration
Furnace
Rotating
Motor
H2S04
Trap
FIGURE 25. ZEOLITE REGENERATION SYSTEM
49
-------
The spent clinoptilolite samples were placed in a stainless steel tube
which could be inserted into the furnace. The furnace apparatus was con-
figured so a carrier gas could be passed through the stainless steel tube
thereby conveying the off-gases through a trap containing 500 ml of 120 g/1
HoSOrt.
The exchange capacity of the clinoptilolite was determined by passing a
concentrated NH3-N solution (3,000 mg/1 NH^-N) through a column containing
400 g (0.9.1b) of clinoptilolite until the effluent NH3-N concentration was
equal to the influent NH3-N concentration. Following exhaustion, two liters
of demineralized water were passed through the column to remove any residual
NH3-N not actually adsorbed on the clinoptilolite. Samples of both the in-
fluent and effluent were taken and analyzed for NH3-N to determine the
exchange capacity per gram of clinoptilolite.
A 15 g sample of the dried, spent clinoptilolite was placed in the
stainless steel tube and inserted into the furnace, previously heated to the
desired temperature. Air was blown through the tube at a constant rate of
one 1/min while rotating the tube at 10 rpm. Samples from the ammonia trap
were taken initially and every 15 minutes thereafter for two hours. From the
data, the amount of NH3-N desorbed could be determined.
The results of the thermal regeneration using compressed air as the
carrier gas were non-reproducible. The variation in the results were most
likely caused by the oxidation of ammonia within the temperature range of
500°C-600°C in an oxygen atmosphere. A source of the variation may stem
from the conversion of an undeterminable quantity of desorbed ammonia to N?
before being trapped in the 120 g/1 ^04 solution. Unfortunately, the
analytical control was not designed to delineate this mechanism and it was
therefore impossible to determine the exact quantity of ammonia oxidized at
the higher temperatures.
The highest percentage of recovered adsorbed ammonia was only 53% at
500°C. The results also show that only 35% of the adsorbed ammonia was
recovered at 600°C, suggesting that the difference may be due to the
oxidation of a larger fraction of the ammonia at the further elevated
temperatures.
Since significant variations were observed in the ammonia recovery in
the thermal regeneration tests using compressed air, an inert gas (argon)
was used as the carrier gas. The results obtained using argon are presented
in Figure 26. As shown previously, thermal regeneration was much more
effective at 500°C and 600°C than 300°C or 400°C. At 300°C and 400°C only
25% and 65%, respectively, of the NH3-N adsorbed onto the clinoptilolite
was desorbed and recovered.
50
-------
« «600°C
*500°C
NOTE: 276 mg NH3N
ADSORBED ON 15
GRAMS OF CLINO
.25 50 .75 1.0 1.25
REGENERATION TIME (MRS.)
FIGURE 26. AMMONIA RECOVERY FROM THERMAL REGENERATION OF CLINO
The NHs-N uptake capacities of the regenerated clinoptilolite samples
are presented in Table 12. With an influent ammonia concentration of
approximately 13 mg/1, the-effective exchange capacity of the virgin clino-
ptilolite was 0.228 meq NH3-N/g while the clinoptilolite, thermally re-
generated at 500°C and 600 C had effective exchange capacities of 0.244 and
0.203 meq NH3-N/g clinoptilolite, respectively.
TABLE 12. EFFECTIVE ADSORPTION CAPACITY OF THERMALLY REGENERATED CLINO
Sample
mg NH-y-N Adsorbed
meq
Clinoptilolite
Virgin Clinoptilolite
Hector
300°C Regenerated
Clinoptilolite
400°C Regenerated
Clinoptilolite
500°C Regenerated
Clinoptilolite
600°C Regenerated
Clinoptilolite
6.38
-24.63
- 3.37
6.86
5.71
0.228
0.245
0.204
51
-------
It should be noted, that the sample regenerated at the 500°C had a slightly
higher EEC than the virgin clinoptilolite which was used as a control. This
may be due to the removal of some impurities within the clinoptilolite, such
as carbonates or volatile adsorbates, during the thermal treatment. The
slightly lower EEC of the sample regenerated at 600°C could have been caused
by an alteration in the crystalline structure of the clinoptilolite. This
is supported by the reports of several investigators who reported that the
clinoptilolite crystalline structure begins to collapse at 750°C (13).
Samples regenerated at 200°C and 400°C showed no NH3-N uptake when contacted
with the influent NHs-N solution. Conversely, these samples released NH3-N
into the contacting solution, indicating incomplete regeneration.
As part of this study, X-ray diffraction analyses of Hector clinopti-
lolite samples were performed following the methods of Mumpton (13) and Rooney
and Kerr (14). The analysis of virgin, exhausted, and thermally regenerated
(500°C) Hector clinoptilolite samples indicated little or no structural
degradation of the thermally regenerated clinoptilolite. Several x-ray power
pattern spectra are shown in Figure 27. From Figure 27 it appears that the
effect of thermal regeneration results in the broadening and elimination of a
number of diffraction spectral lines indicating the loss of adsorbed con-
stituents, occluded water, and possibly some minimal structural alteration.
Hv
VIRGIN CLINO 450° C
r * x
Q
SPENT CLINO 450° C
REGENERATED
CLINO A500°C
SPENT CUNO 90 °C
VIRGIN CUNO 90 °C
30 28 26 24 22 20 18 16 14 12 10
DEGREES 26
FIGURE 27. X-RAY DIFFRACTION PATTERNS OF CLINOPTILOLITE
52
-------
ALUM REGENERATION
There are two methods of alum recovery, acid treatment and alkaline
treatment of the alum sludge followed by a solids separation process.
Excellent recoveries of alum were reported by Slechta and Gulp (15) when
they lowered the pH of the alum sludge to 2.5. Unfortunately, phosphorus
was also recovered, and the authors concluded that alum recovery was
feasible only when phosphorus removal was not required.
The alkaline treatment consisted of elevating the pH of the waste
alum sludge with either lime or sodium hydroxide followed by treatment
with calcium chloride. Lea, et al (16) reported that sodium hydroxide and
calcium chloride treatment to pH 11.9 resulted in a high percentage of
recovered aluminum (93%); however, the coagulation properties of the re-
claimed aluminum were inferior to fresh alum when used in dosages resulting
in an A1:P molar ratio of 2. Coagulant reclamation via lime treatment of
the alum sludge to pH 10.2 was reported by Slechta and Gulp (15) to be
ineffective as only 20% of the aluminum was recovered.
The approach taken in this study was to thermally treat the spent.
alum sludge and then dissolve it in sulfuric acid. It was hypothesized
that at elevated temperatures the phosphorus in the alum floe would
possibly convert to phosphorus pentoxide, P205 (phosphorus anhydride)
which volatilizes at about 250°C. The phosphorus pentoxide would then be
trapped utilizing a caustic scrubber, where it would form sodium phosphate
Na3(PO^)2- At temperatures approaching 900°C the hydrated aluminum
hydroxide floe may dehydrate and convert to a partially hydrated aluminum
oxide, A1203 . XH20.
Sufficient dibasic potassium phosphate was dissolved in distilled
water to yield a phosphate concentration of 1.2%. To this solution, alum
was added until a floe was formed. The floe was allowed to settle and the
clear supernatant was decanted and analyzed for phosphate. The resultant
sludge was centrifuged and dried at 103°C for an hour. A portion of the
dried sludge was placed in a furnace, preheated to 925°C, for a period of
20 minutes. The sludge samples were cooled, weighed, and digested in 10ml
of hot sulfuric acid, F^SO^, and 2 ml of concentrated nitric acid, NHO^.
After 2.5 hours of digestion, the samples were cooled, and the liquid
level was brought up to 50 ml in. preparation for phosphorus analyses.
Analyses of the resultant data indicated that thermal treatment of
the waste alum sludge at 925°C did not reduce the amount of phosphate
recovered. On the contrary, more phosphate was recovered from the alum
sludge treated at 925°C for 20 minutes, than from the alum sludge treated
at 105°C. Almost twice as much phosphate (PO*) was recovered from the
925°C treatment than from the 105°C treatment. The actual percentages of
P04 recovered were 28.3% and 14.4% for the alum sludge treated at 925°C
and 105°C, respectively.
Several other tests were conducted for the purpose of evaluating the
effect of several temperatures on PO^ recovery. The temperatures evaluated
53
-------
were 105°C, 250°C, 500°C, 800°C, and 1000°C. These tests differed from
the initial evaluations in that domestic sewage was used rather than a
phosphate solution. The samples were dosed with 150 mg/1 alum which
resulted in an aluminum to phosphorus molar ratio of 5.2. The procedure
used was essentially the same as described above, with the exception that
the sludge was filtered rather than centrifuged. The digestion step
involved the use of 5 ml of concentrated HO, and 15 ml of water.
The data from this series of tests confirmed the conclusions drawn
from the earlier tests that thermal treatment of alum sludge does not
reduce the amount of phosphate recovered during its acidification. An
interesting observation noted during these tests was that essentially all
of the phosphate was recovered from the alum sludge regardless of the
temperature of treatment. Because of the high levels of phosphate
recovered upon acidification of the waste alum sludge, the percent recovery
of aluminum was not determined and its effectiveness as a coagulant was
not further evaluated.
54
-------
SECTION 8
PROJECTED SYSTEM COSTS
GENERAL
In developing a cost estimate for a high rate combined sewer overflow
treatment system for nutrient removal, a number of assumptions must
be made in designing such a system to treat a "typical" waste. It
must be noted that the costs developed, herein, are only estimates, and
should be considered only as a guideline when contemplating the installation
of a similar system. Although major equipment items may have similar costs,
installation costs and particularly conveyance costs are site dependent.
Therefore, depending on the exact location the actual costs may be either
higher or lower.
CAPITAL COSTS
Capital costs are based on a system designed to handle a combined sewer
overflow occurrence with the following characteristics:
Number of overflows per year 100
Design capacity 10 mgd
Duration treatment facility operation 4 hours
Average treatment rate 5 mgd
Duration of overflow 8 hours
Influent quality:
Average SS cone. 200 mg/1
Average NH^-N cone. 15 mg/1
Average TIP cone. 10 mg/1
The operating characteristics of the proposed system were developed
from the demonstration scale study and are listed in Table 13.
55
-------
TABLE 13. OPERATING CHARACTERISTICS FOR A 10 MGD COMBINED SEWER OVERFLOW
NUTRIENT REMOVAL FACILITY
Horizontal drum screen size
Horizontal drum screen application rate
Filtration application rate (variable flow)
Filter solids capacity
Design for peak flow rate
Filter area (total)
Number of filters (pressure)
Filter media depth
Filter dimensions
Filter run
Alum dosage
Polymer dosage
Ion exchange/adsorption application
Clinoptilolite ammonia capacity
rate
Clinoptilolite bed volume BV
Clinoptilolite bed depth
No. of contactors
Clinoptilolite contactor dimensions
Cycle time
Thermal regeneration time
- 0.305m x 0.035m (10 ft. 0 by
10 ft.) 7
- 1 m3/min/5 gm (25 gpm/ft )
- 456 1/min/m2 (11.2 gpm/ft2)
- 5.6 kg SS/m3 (0.35 Ib SS/ft3-
includes alum SS)
- 37.85 x 103 m3/day (10 mgd)
- 65 m2 (700 ft2)
- 4
- 2.5 m (8 ft.)
- 15' dia. x 18' depth
- 6 hours
-215 mg/1
- 1 mg/1 polymer
- 15 BV/hr
-0.42 meq NHs-N/g to break-
through to 1 mg/1 NH-(N)
- 105 m3 (3750 ft3) d
- 1.83 m (6 ft)
- 12
- diameter: 4.27 m (14 ft)
height: 4.57 m (15 ft)
- breakthrough lead column:22 hrs.
exhaustion lead column:35 hrs.
- 7 days
The complete system is parcelled into several subsystems which are
listed below, and the capital costs are presented in Table 14.
1.
2.
3.
4.
5.
6.
7.
High Rate Filtration Plant Pumping Station
Screening Subsystem
Chemical Addition Subsystem
Filtration
Ion Exchange
Regeneration Facilities
Miscellaneous
TABLE 14. CAPITAL COSTS FOR 10 MGD NUTRIENT REMOVAL SYSTEM FOR CSO
Pumping Station
a) Excavation and Hauling
b) Reinforced Concrete
c) Flow Meters
d) Pumps and Control
e) Liquid Level Recorder
f) Miscellaneous
$ 41,000
40,000
20,000
54,000
5,000
10,000
$170,000
56
-------
2. Screening Facilities
3.05 m x 3.05 m (10 ft x 10 ft) Rotating
Horizontal Drum Screen $ 83,000
3. Chemical Facilities
a) Alum tanks, mixer and pump 120,000
b) Polyelectrolyte tanks, mixer and pump 65,000
c) In-line static mixers • 18,000
$203,000
4. Filter System
a) Filter with valving, etc. 250,000
b) Media (initial charge) 10,000
$ 260,000
5. Ion Exchange System
a) Contactors 400,000
b) Media (initial charge) 280,000
$680,000
6. Thermal Regeneration Facility
a) Conveyance system 50,000
b) Dewatering system 60,000
c) Multiple hearth furnace 750,000
$860,000
7. Miscellaneous
a) Building 250,000
b) Heating & ventilation 30,000
c) Electrical 30,000
d) Excavation and landfill 15,000
e) Filter pumps 60,000
f) Instrumentation 50,000
g) Piping 50,000
h) Site preparation 20,000
$505,000
Subtotal $2,761,000
Installation (40% subtotal) 1,105,000
$ 3,866,000
Construction Contingency (15%) 580,000
$ 4,446,000
Engineering & Administration (12%) 534,000
TOTAL ESTIMATED PROJECT COST $ 4,980,000
57
-------
ANNUAL COSTS
Annual operating and maintenance costs are projected in Table 15
on the basis of treating 100 CSO occurrences per year. This system is
expected to operate at a peak flow for four hours and at an average flow for
8 hours. Assuming each occurrence lasts 12 hours, this results in a total
of 1200 hours of facility operation per year. Although the CSO facility
may be automated to such a degree that requires minimum manpower involvement,
the thermal regeneration unit for spent clinoptilolite would be in operation
24 hours/day and would require moderate supervision.
The annual costs of the Nutrient Removal Facility are listed in Table
15.
TABLE 15. ANNUAL COST OF NUTRIENT REMOVAL FACILITY
Operating
a) Alum $27,000
b) Polyelectrolyte 8,500
$35,000
a) Anthrafilt $ 1,000
b) Clinoptilolite 56,000
$57,000
Labor 60>000
Maintenance
a) Mechanical (3% of Mechanical Equip. Costs) $39,000
b) Electrical (2% of Electrical Equip. Costs) 15,000
c) Structural (1% of Structural Equip. Costs) 5,000
$59,000
Utilities
a) Electricity ($.03/KWH) $12,008
b) Fuel Oil ($.40/ga1) 10,000
$22,000
Amortization
7.5% interest rate for 25 years $447,000
TOTAL ANNUAL COSTS $680,000
58
-------
As indicated in the above table the most significant cost is the
amortization of the facility, which in this case is equivalent to 70% of the
total annual costs. An accurate treatment cost per 1000 gallons treated can
not be established since the system would not be operating at full capacity
for the entire overflow occurrence. Under the conditions mentioned previously
in this section, the facility would operate at 33% capacity resulting in
a cost of $3.00 per 1000 gallons of CSO treated. It should be emphasized that
these costs are based on projections for a certain geographical area,
Syracuse, N.Y. Actual treatment costs per unit volume would vary with
location, since they are dependent on factors, such as the number of rain
events, duration and intensity of the rain events, size of drainage area, and
physical condition of the combined sewers. Also, some economy may be realized
if a centrally located thermal regeneration facility were available that would
regenerate clinoptilolite from several overflow treatment sites.
59
-------
REFERENCES
0) Field, R. Coping with Urban Runoff in the United States. Water Res.,
(9):499, 1975.
(2) Lager, J.A., and W.6. Smith. Urban Stormwater Management and Technology:
An Assessment. EPA-670/2-74-040 U.S. Environmental Protection Agency,
Cincinnati, Ohio, 1974.
(3) Colston, N. Characterization and Treatment of Urban Land Runoff.
EPA-670/2-74-096, U.S. Environmental Protection Agency, Cincinnati,
Ohio, 1974.
(4) Uttormark, P.O., et. al. Estimates Nutrient Loadings of Lakes from
Non-Point Sources. EPA-660/3-74-020 Ecological Research Series,
U.S. Environmental Protection Agency, Washington, D.C., 1974.
(5) Murphy, C.B., Jr. Effect of Restricted Use of Phosphate-Based
Detergents on Onondaga Lake. Science, 182:379, 1973.
(6) Koon, J.H. and W.J. Kaufman. Ammonia Removal from Municipal Wastewaters
by Ion Exchange. Journal Water Pollution Control Federation, 47 (3):
448-465, 1975.
(7) Murphy, C.B., and 0. Hrycyk. Ammonia Removal From Municipal Wastewater
via Zeolite Application and Subsequent Thermal Regeneration. Paper
Presented at the 45th Annual Conference of the Water Pollution Control
Federation. Atlanta, GA, 1972.
(8) Nebolsine, R., P.J. Harvey, and C.Y. Fan. High Rate Filtration of
Combined Sewer Overflows. EY1 04/72, Water Pollution Control Series
11023, U.S. Environmental Protection Agency, Washington, D.C. 340 pp.
(9) Jenkins, D., J.F. Ferguson, and A.B. Menar. Chemical Processes for
Phosphate Removal. Water Research, 5:369-389, 1971.
(10) McLaren, J.R. and G.J. Farquhar. Factors Affecting Ammonia Removal
by Clinoptilolite. J. Environmental Engineering Division ASCE, (99);
429-446, 1973.
(11) Packham, R.F. Some Studies of the Coagulation of Dispersed Clays with
hydrolyzing Salts. Journal Colloidal Science, 20: 81-92, 1965.
(12) Minton, G.R. and D.A. Carlson. Combined Biological-Chemical Phosphorus
Removal. Journal Water Pollution Control Federation, 44 (9):1736-55,
1972.
60
-------
(13) Mumpton, F.W. Clinoptilolite Redefined. American Mineralogist, (45):
351-369, 1960.
(14) Rooney, T.P. and P.P. Kerr. Clinoptilolite: A New Occurrence in N.
Carolina Phosphorite. Science, (148):1453, 1964.
(15) Slechta, A.F., and G.L. Gulp. Water Reclamation Studies at the South
Tahoe Public Utility District. Journal Water Pollution Control
Federation, 39 (5): 787 1967.
(16) Lea, W.L., 6.A. Rohlich and W.J. Katz. Removal of Phosphates from
Treated Sewage. Sewage and Industrial Wastes, 26 (3): 261, 1954.
61
-------
GLOSSARY
BED VOLUME PER HOUR (BV/hr.): A term used to simultaneously define flux
and contact time for a removal process relying on either a sorption or
exchange mechanism. Bed volume refers to the amount of exchange material
in the contactor.
CLINOPTILOLITE: A naturally occurring zeolite which has an ion exchange
specificity for the ammonium cation. It is stable upon dehydration and
readily readsorbs water. It is thermally stable up to 700° C.
EFFECTIVE EXCHANGE CAPACITY (EEC): The amount of substrate removed by an
exchange material under dynamic conditions, expressed as mi Hiequivalents
of substrate removed per gram of exchange material (meq/g). Principal
factor influencing the EEC of an exchange material is the initial substrate
concentration.
EXCHANGE CAPACITY: The amount of substrate removed per unit weight of
exchange material under equilibrium conditions, expressed as mini-
equivalents of substrate removed per gram of exchange material (meq/g).
FLUX or INFLUX: Rate of liquid flow expressed as unit volume/unit time.
ZEOLITE: Crystalline, hydrated aluminosilicates of alkaline earth
materials (sodium, potassium, calcium, magnesium, barium and strontium),
having a three dimensional silicate structure.
62
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-78-056
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
HIGH RATE NUTRIENT REMOVAL FOR COMBINED SEWER OVERFLOWS
Bench Scale and Demonstration Scale Studies
5. REPORT DATE
June 1978 (Issuing Date)
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
C. B. Murphy, Jr., Orest Hrycyk, William T. Gleason
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
O'Brien & Gere Engineers, Inc.
1304 Buckley Road
Syracuse, New York 13221
1O. PROGRAM ELEMENT NO.
1BC611
11. CONTRACT/GRANT NO.
S-802400
12. SPONSORING AGENCY NAME AND ADDRESS
Municipal Environmental Research Laboratory—Gin.,OH
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA/600/14
15. SUPPLEMENTARY NOTES
Sponsorship w1 th Division of Drainage and Sanitation, Depart
ment of Public Works, County of Onondaga, New York
EPA Pro.iect Officer; C.Y. Fan, Storm & Combined Sewer Section (201) 321-6680 _
16. ABSTRACT
A high rate physical/chemical treatment system has been evaluated for the removal
of suspended solids and the macronutrients, phosphorus and nitrogen, from combined sewc
overflow. The system utilized a single unit process concept consisting of in-line
chemical addition, coagulation, flocculation, high rate filtration and ion exchange.
The results of this program have demonstrated that the simultaneous removal of
suspended solids, phosphorus, and ammonia-nitrogen from a combined sewer overflow is
feasible using the high rate unit process concept.
Suspended solids removal ranged from 90-100% with alum and polymer dosages of 220
mg/1 and 1 mg/1, respectively.
Alum dosages resulting in Al/P molar ratios greater than 1.4 were effective in re
ducing the total inorganic phosphorus (TIP) concentrations from greater than 10 mg/1 tc
less than 0.9 mg/1.
The clinoptilolite was effective in reducing the ammonia level to below the deted
able levels of 0.02 mg/1 NH3-N. Infulent NH3-N concentrations ranging from 0.2 to 16
mg/1 were reduced below detectable levels during the initial period of contact. The
effective exchange capacity of clinoptilolite was found to range from 0.20 to 0.64 meq
NHo-N/g clinoptilolite as the influent NH«-N concentration ranged from 7.5 to 16 mg/1,
respectively.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Combined sewers*, Overflows*, Water
pollution*, Filtration*, Ion exchange*,
Coagulation*, Zeolite*
Phosphorus removal*,
Ammonia removal*,
High rate*, Single
unit process*,
Clinoptilolite*
13B
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (ThisReport)'
UNCLASSIFIED
21. NQ. OF PAGES
73
20. SECURITY CLASS (Thispage)
UNCLASSTFTFn
22. PRICE
EPA Form 2220-1 (9-73)
63
U.S. GOVERNMENT PRINTING OFFICE: 1978-757-140/1348 Region No. 5-11
------- |