EPA/600/R-02/077
November 2003
Research Summary
CSO Disinfection Pilot Study:
Spring Creek CSO Storage Facility Upgrade
by
Izabela Wojtenko and Mary K. Stinson
U.S. Environmental Protection Agency
Urban Watershed Management Branch
Edison, New Jersey, 08837
Contract Number: 7C-R394-NTLX
Project Officer
Mary K. Stinson
Water Supply and Water Resources Division
National Risk Management Research Laboratory
Cincinnati, Ohio 45268
National Risk Management Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
Notice
The U.S. Environmental Protection Agency (EPA) through the Office of Research and
Development partially funded and collaborated in the research described here under contract
No.7C-R394-NTLX to Camp Dresser & McKee of Woodbury, New York. It has been subjected
to the Agency's peer and administrative review and has been approved for publication as an EPA
document.
-11-
-------
Foreword
The U.S. Environmental Protection Agency (EPA) is charged by Congress with
protecting the Nation's land, air, and water resources. Under a mandate of national
environmental laws, the Agency strives to formulate and implement actions leading to a
compatible balance between human activities and the ability of natural systems to support and
nurture life. To meet this mandate, EPA's research program is providing data and technical
support for solving environmental problems today and building a science knowledge base
necessary to manage our ecological resources wisely, understand how pollutants affect our
health, and prevent or reduce environmental risks in the future.
The National Risk Management Research Laboratory (NRMRL) is the Agency's center
for investigation of technological and management approaches for preventing and reducing risks
from pollution that threaten human health and the environment. The focus of the Laboratory's
research program is on methods and their cost-effectiveness for prevention and control of
pollution to air, land, water, and subsurface resources; protection of water quality in public water
systems; remediation of contaminated sites, sediments and ground water; prevention and control
of indoor air pollution; and restoration of ecosystems. NRMRL collaborates with both public
and private sector partners to foster technologies that reduce the cost of compliance and to
anticipate emerging problems. NRMRL's research provides solutions to environmental problems
by: developing and promoting technologies that protect and improve the environment; advancing
scientific and engineering information to support regulatory and policy decisions; and providing
the technical support and information transfer to ensure implementation of environmental
regulations and strategies at the national, state, and community levels.
This publication has been produced as part of the Laboratory's strategic long-term
research plan. It is published and made available by EPA's Office of Research and Development
to assist the user community and to link researchers with their clients.
Lee A. Mulkey, Acting Director
National Risk Management Research Laboratory
-111-
-------
Abstract
This Research Summary presents the results of a pilot-scale disinfection study performed
for the New York City Department of Environmental Protection and the U.S. Environmental
Protection Agency (US EPA) under a contract No. 7C-R394-NTLX to Camp Dresser & McKee
of Woodbury, New York. The main objective of the pilot study was to demonstrate alternatives
to hypochlorite disinfection for application to the Spring Creek facility and potentially to other
combined sewer overflow (CSO) facilities. The pilot testing was divided into two phases. Phase
I was performed from December 1996 through March 1997, and Phase II was performed from
August through November 1999. US EPA provided technical assistance to the entire study.
Phase I evaluated treatment performance of five technologies: ultraviolet (UV) irradiation;
ozonation (O3); chlorine dioxide (C1O2) disinfection; chlorination/dechlorination (Cl2/deCl2); and
electron beam irradiation (E-Beam). The fifth technology, E-Beam was evaluated during
supplemental Phase I pilot testing sponsored by the New York Power Authority and the Electric
Power Research Institute. Based on the results from Phase I, Phase II provided additional
evaluation of technologies that had shown potential for CSO applications. These were UV,
C1O2, and Cl2/deCl2. This Research Summary concentrates on these three technologies, but the
overall results of both phases for each technology along with the cost are also discussed.
Generally, all tested disinfection technologies, with the exception of E-beam, were able
to achieve targeted bacterial reductions of 3 to 4 logs. Chlorination/dechlorination, C1O2, and O3
at the doses tested were able to provide these levels of disinfection over the full range of
wastewater quality tested. To date, ozonation was not found to be cost effective for CSO
applications. While C1O2 was superior in effectiveness and similar in cost to Cl2/deCl2, the
generation technology for C1O2 which avoids the need for gaseous C12 needs further
development. Because an effective Cl2-gas-free process of C1O2 generation has not been proven
to be reliable and, because C12 gas cannot be transported within New York City, disinfection
with C1O2 cannot be recommended for use within New York City at this time.
The Spring Creek facility upgrade construction was scheduled for the Fall of 2002. The
upgraded Spring Creek facility will continue to use sodium hypochlorite for disinfection, with
provisions to add dechlorination at a later date. Improvements will be made to increase
disinfectant flash mixing and to automate hypochlorite feed and residual control. While the
decision to continue the use of hypochlorination for the upgraded facility was based upon the
Phase I study, it is recognized that the overall pilot results (Phases I and II) will be used to guide
decisions at other CSO facilities.
-IV-
-------
CONTENTS
Notice ii
Foreward iii
Abstract iv
List of Acronyms vii
Acknowledgments viii
1. INTRODUCTION
2. DESIGN AND OPERATION OF THE PILOT STUDY
3. PILOT INVESTIGATIONS
4. PROCESSES INVESTIGATED .......................................... ^
4.1 UV Light Disinfection ............................................ -_4^
UV Pilot Equipment .............................................. -7-
UV Pilot Operation .............................................. ^
4.2 Chlorine Dioxide ................................................ ^
Chlorine Dioxide Pilot Equipment .................................. -9-
Chlorine Dioxide Pilot Operation .................................. -13-
4.3 Chlorination/Dechlorination ...................................... -14-
Chlorination/Dechlorination Pilot Equipment ....................... -14-
Chlorination/Dechlorination Pilot Operations ......................... -14-
4.4 Ozone [[[ -15-
Ozone Equipment ............................................... -15-
Ozone Operations .............................................. -15-
5. RESULTS [[[ -17-
5.1 Ultraviolet Irradiation ........................................... -17-
Dose-Response Relationships -17-
Water Quality Relationships -18-
Viral Reductions -20-
Disinfection Byproducts -20-
5.2 Chlorine Dioxide -20-
Dose-Response Relationships -21-
Water Quality Relationships -23-
Viral Reductions -23-
Disinfection Byproducts -23-
Chlorine Dioxide Residuals Vs Dose -26-
5.3 Chlorination/Dechlorination -27-
Dose-Response Relationships -27-
-------
6. COST COMPARISON -34-
7. CONCLUSIONS -38-
7.1 Wastewater Quality -38-
7.2 Treatment Performance -40-
7.3 Disinfection Residuals and Toxicity -41-
7.4 Chlorine Dioxide Generation -42-
8. SUMMARY -42-
REFERENCES -44-
-VI-
-------
List of Acronyms
AOC - assimilable organic compounds
APWA - American Public Works Association
BMP - best management practice
BOD - biochemical oxygen demand
BOD5 - five-day biochemical oxygen demand
Cl! - chloride
C12 - chlorine
C1O2 - chlorine dioxide
C1O2! - chlorite
C1O3! - chlorate
COD - chemical oxygen demand
deC!2 - dechlorination
DNA - deoxyribonucleic acid
DO - dissolved oxygen
NaClO2" - sodium chlorite
NPDES - National Pollutant Discharge Elimination System
NRMRL - National Risk Management Research Laboratory
NYCDEP - New York City Department of Environmental Protection
O2 - oxygen
O3- ozone
ORD - Office of Research and Development
PAH - polycyclic aromatic hydrocarbon
POTW - Publically Owned Treatment Works
SS - suspended solids
SSO - sanitary sewer overflow
SW - stormwater
TCOD - total COD
THM - trihalomethane
US EPA - United States Environmental Protection Agency
UV - ultraviolet
UWMB - Urban Watershed Management Branch
WSWRD - Water Supply and Water Resources Division
WWF - wet weather flow
WWTP - wastewater treatment plant
-Vll-
-------
Acknowledgments
Camp Dresser & McKee (CDM) of Woodbury, NY, was contracted by the New York
City Department of Environmental Protection (NYCDEP) to provide engineering design services
for the upgrade to the Spring Creek CSO Storage Facility. This Research Summary is based on
the original study reports prepared by CDM and CDM's subcontractor, Moffa & Associates
(currently Brown & Caldwell) of Syracuse, NY (CDM, 1997 and 1999).
The authors are with the Urban Watershed Management Branch (UWMB) of the Water
Supply and Water Resources Division (WSWRD) in Edison, NJ at the National Risk
Management Research Laboratory (NRMRL) of the U.S. Environmental Protection Agency's
(US EPA's) Office of Research and Development (ORD).
Technical review was provided by Shirley E. Clark, Ph.D., P.E. Assistant Professor,
School of Science, Engineering and Technology, Penn State, Harrisburg, PA.
Editorial review was provided by Laura Panos of the Environmental Careers
Organization.
-Vlll-
-------
1. INTRODUCTION
Camp Dresser & McKee (CDM) of Woodbury, NY, was contracted by the New York
City Department of Environmental Protection (NYCDEP) to provide engineering design services
for the upgrade to the Spring Creek CSO Storage Facility. These services included a pilot study
to evaluate the following disinfection technologies for treatment of combined sewer overflows
(CSOs): ultraviolet (UV) irradiation; ozonation (O3); chlorine dioxide (C1O2) disinfection;
chlorination/dechlorination (Cl2/deCl2); and electron beam irradiation (E-Beam). The pilot
testing was divided into two phases. Phase I, from December 1996 through March 1997,
evaluated performance of all five technologies listed above. The E-Beam was evaluated during
supplemental Phase I pilot testing sponsored by the New York Power Authority (NYPA) and the
Electric Power Research Institute. The E-Beam technology did not appear to be feasible for
CSO disinfection, thus, the study results for this technology are not being discussed in this
Summary. Phase II of the pilot testing, from August through November 1999, evaluated
technologies which had been found promising for CSO disinfection during the first phase. Thus,
Phase II further investigated UV, C1O2, and Cl2/deCl2 under additional wet-weather conditions.
This Research Summary is based on the original study reports prepared by CDM and CDM's
subcontractor, Moffa & Associates (currently Brown & Caldwell) of Syracuse, NY (CDM, 1997
and 1999). Because the original reports were prepared for internal use and were made not
readily available to the Public, the authors believe that this summary will provide means to
distribute the information to the Public.
Chlorine has traditionally been used to provide disinfection due to its low cost. Since the
1970s, growing awareness of the adverse environmental impacts associated with the byproducts
of chlorination has led to increasingly restrictive residual C12 requirements. While the current
State Pollution Discharge Elimination System permit for Spring Creek allows a maximum total
residual C12 effluent limit of 2.0 mg/L, it is expected that more restrictive effluent limits,
consistent with the water quality standard, will be required in the future. As a result of the
impending restriction on the Total Residual Chlorine (TRC) in the effluent, dechlorination
(deC!2) was considered for disinfection of CSO at the Spring Creek and at other CSO facilities.
In addition, a need to minimize environmental risks, chemical demands, and contact times that
are required using conventional Cl2/deCl2 has fostered a strong interest in alternative disinfection
technologies for CSO. However, there is minimal data available on the effectiveness of
alternative disinfectants for the treatment of CSO. This pilot study was conducted to provide the
needed performance data on the alternative disinfection technologies for their possible future use
at the Spring Creek or elsewhere.
2. DESIGN AND OPERATION OF THE PILOT STUDY
As mentioned previously, the objectives of the pilot study were to provide: (1) a basis for
selection of a disinfection technology for use at the Spring Creek CSO Storage Facility, (2) full-
scale design criteria for application at the Spring Creek CSO Storage Facility; and (3) treatment
performance data to guide selection of disinfection technologies for other CSO facilities. The
location for both phases of the pilot study was a designated area at the 26th Ward Wastewater
-1-
-------
Pollution Control Plant (WPCP) rather than the CSO Storage Facility due to space and logistical
constraints. In addition, the 26th Ward site was able to provide a continuous supply of
wastewater regardless of weather conditions and the 26th Ward primary effluent was
representative of the range of wastewater quality found in the effluent from the Spring Creek
CSO Storage Facility (CDM, 1997). However, it is important to indicate that the characteristics
of wastewater in Phase I and Phase II were not the same. Figure 1 presents the schematic of the
CSO Pilot location.
Operation of the Phase I pilot units occurred over the period of December 17, 1996 to
March 26, 1997. The 16 main test runs included a combination of 4 actual wet-weather and 12
simulated wet-weather events. During Phase I, the UV, O3, C1O2, and Cl2/deCl2 pilots were
operated in parallel for a total of 16 test runs. Testing of the units in parallel allowed comparison
of disinfection efficiency of each technology on wastewater of identical quality. During testing,
the operating conditions of each pilot unit were held constant over the duration of the test run.
Flowrates, detention times, disinfectant dose, and mixing conditions were varied between test
runs. The purpose of these test runs was to evaluate the performance of each technology over a
range of operating conditions and over a range of wastewater quality typical of CSOs.
Wastewater flow to the pilot facility was supplied from the primary settling tank effluent
channel. Each pilot unit received flow from the common header. The flowrate to each pilot unit
was manually controlled using in-line magnetic flowmeters and throttling valves. Treated pilot
effluent was discharged directly into the 26th Ward plant recycle line through an existing
manhole adjacent to the pilot facility. Figure 2 presents the pilot system flow schematic. Each
pilot system was subjected to the same wastewater, in order to compare the performance of each
pilot system directly against the other.
Operation of the Phase II pilot units occurred over the period of August 25, 1999 to
October 21,1999. The same pilot location and source wastewater were used as in Phase I.
During Phase II, the UV, C1O2, and Cl2/deCl2 pilots were operated in parallel over eight test
events. Although these technologies had been investigated in Phases I and II, the tested units
were from different manufacturers. The operating conditions of each pilot unit were held
constant over the duration of the test run. Flowrates, detention times, disinfectant dose, and
mixing conditions were varied between test runs. Phase II pilot study supplemented the Phase I
results and performed additional research on the selected three technologies. The Phase II data
was compiled with the Phase I data to evaluate the treatment performance over the full range of
wastewater quality experienced. Ultimately, the Phase I and Phase II results were used to make
recommendations and develop design criteria for full-scale CSO disinfection in New York City.
This Research Summary concentrates on discussion of these results.
In Phase II, the three pilot units were located side-by-side for concurrent operation
alongside primary settling tank no. 4. This location provided a constant supply of wastewater
and ready access to the primary settling tank effluent channel. The three pilot technologies were
tested at various dosages, contact times and flowrates over a large range of wastewater
-2-
-------
Figure 1. CSO pilot location at the 26th Ward WPCP.
C i
' i
I «
I I \
-A \
*- \ \
A ^?
\ \,»
K
^A -, r^
C-' ' >> • ^^~-
V;^;,
N
f '
i'
ii MI ;n
4 ^W=
\JP
l.. .i, L
TT-V
a I
:=.'t"^=±nT"L«"T
.-=_!_ i w-^ a^*" __-__ __- r
i;
f'l
I"
i-.nj ........ -I] ,
i
V _ a...
i f*«i
.
rS" " r*tv
§H» o .-5\
bs F" X.
.ir==d ' %v
-3-
-------
conditions. Eight test runs were performed; four during wet-weather conditions and four during
dry-weather conditions. The target dosages were designed to develop an appropriate dose-
response relationship for each technology and to supplement the Phase I results. As the
analytical results from prior test runs were received, the dosages and operating conditions for
subsequent test runs were adjusted and modified to achieve the desired range of bacteria kills.
The actual chemical dosages and pilot unit operating conditions used during Phase II are
provided in Table 1. This table illustrates the flowrate, disinfection dose and detention times
which were provided for each event and for each individual pilot unit.
The objectives of the entire study were limited to the relative comparison of alternative
disinfection technologies and the evaluation of basic design parameters (e.g., dose, mixing
configuration, contact time) for application to CSO. Optimization of other detailed design
parameters and simulation of full-scale process configurations for use at the Spring Creek
facility were beyond the scope of this effort. Additionally, the small size of the pilot units as
compared to actual CSO flowrates, did not lend the pilot results to the determination of factors
such as full-scale operational complexity and safety concerns, system constructability and size/
structural constraints, and technology cost effectiveness. This information can be determined
from available full-scale facilities, but it was not within the scope of this study to do so.
3. PILOT INVESTIGATIONS
During the test runs, the pilot facilities were allowed to reach steady state conditions for a
minimum of 30-min prior to initiating sampling. The UV unit was first started up on potable
water in order to get an initial relative UV intensity reading on clean water. After the
establishment of wastewater flow to the pilot facility, typical startup would proceed as follows:
(1) balance wastewater flow to the individual pilot units, (2) restart the UV unit on wastewater,
(3) startup of the Cl2/deCl2 and C1O2 mixers, (4) startup and set the hypochlorite and bisulfite
feed pumps to the desired dose, and finally (5) startup and set either the C1O2 gas or liquid feed
to the desired dose. Following this last step, at least 30 min were allowed to pass before
beginning sampling. During each test run, samples of the common pilot influent and the treated
pilot effluents were collected at various frequencies over the 2-h sampling period. In addition to
offsite laboratory analyses, field monitoring of pH, temperature, dissolved oxygen (DO),
oxidation/reduction potential probe (ORP), C12 and C1O2 residual were performed. Process
monitoring and control parameters are summarized in Table 2.
4. PROCESSES INVESTIGATED
4.1 UV Light Disinfection
UV disinfection is accomplished by electromagnetic radiation at specific wavelengths
ranging from 100 to 400 nanometers (nm) with optimum disinfection at 253.7 nm. UV radiation
-4-
-------
Figure 2. The pilot system flow schematic.
aS
®-n
j n
UJ
aj
p f
6 -I
-5-
-------
Table 1. Summary of pilot unit operating conditions during Phase II.
Alternative Date
UV Flowrate
Lamp Power
UV Loading
Average UV
Absorbed
C1O2 Flowrate
Concentration
Generator
Mixing
Cl2/deCl2 Flowrate
Chlorine
Mixing
DRY-WEATHER EVENTS
Run#l
Viral # /
140
1
3.6
32.3
46.2
32
8
CDG
1 stage
32
18
1 stage
Run #4
Viral #2
113
3
7.1
36.8
88.6
32
6
UVD
2 stage
32
20
2 stage
Run #7
Viral #3
140
1
3.6
35.8
50.1
32
10
UVD,
1 stage
32
24
1 stage
Run #8
Viral #4
58
3
13.8
37.0
172.6
32
8
CDG
1 stage
32
28
2 stage
WET-WEATHER EVENTS
Run #2
113
3
7.1
33.9
83.5
32
6
UVD
1 stage
32
20
1 stage
Run #3
77
3
10.4
27.4
105.0
32
8
UVD
2 stage
32
24
2 stage
Run #5
58
3
13.8
36.1
169.3
32
6.5
UVD
2 stage
32
28
1 stage
Run #6
77
3
10.4
25.5
100.0
32
10
UVD
2 stage
32
18
2 stage
Note:
1. Power levels 1, 2 and 3 provide lamp power outputs of 125, 160, and 200 watts UV-C per lamp, respectively.
2. Absorbed UV dose calculated using UV manufacturer's empirical equation.
3. UVD Generator used during the first hour of sampling, while the CDG Generator was used during second hour.
-6-
-------
Table 2. Process Monitoring and Control Parameters.
Process Monitoring
Influent Wastewater
UV Pilot Disinfection System
Chlorine Dioxide Pilot Disinfection System
Chlorination/deChlorination Pilot
Ozone Pilot Disinfection System
E-Beam Pilot Disinfection System
Control Parameters
Temperature
pH
Available Plant Wastewater Qualitv/Flow Data
Wastewater Flowrate Through Pilot
UV Dose
Wastewater Flowrate Through Pilot
Chlorine Dioxide Purity
Chlorine Dioxide Dose
Wastewater Flowrate Through Pilot
Sodium Hypochlorite Feed Concentration
Sodium Hypochlorite Dose
Sodium Bisulfite Dose
Ozone Dose
Wastewater Flowrate Through Pilot
Beam Current
Wastewater Flowrate Through Pilot
Temperature Changes
Absorbed Dose
is the most common electromagnetic radiation technique used for disinfection. It is a physical
process offering short detention times (5 to 7 sec). UV disinfection does not produce residuals
or byproducts that are known to produce risk to humans or aquatic systems. Some concerns have
been raised regarding the development of organism mutations, but no conclusive data exists. UV
technology works on the principle that all microorganisms that contain nucleic acids are
susceptible to damage through the absorption of radiation in the UV energy range. The extent of
damage, mutation, or death will depend upon the organism's resistance to radiation penetration.
UV disinfection is well demonstrated for water and wastewater treatment.
UV Pilot Equipment The UV pilot disinfection system was provided by Aquionics, Inc., of
Erlanger, Kentucky, the same unit used during both phases. This skid mounted unit was
manufactured by Berson Milieutechniek of Nuenen, Holland. The unit was a medium pressure,
high intensity UV unit. The unit was housed inside a trailer to provide a barrier to climatic
condition. A process flow diagram of the UV pilot unit is provided in Figure 3.
The stainless steel reaction chamber was fitted with four high intensity mercury vapor
lamps which were mounted for protection inside four quartz sleeves. Wastewater flowed
through the unit in a horizontal, parallel-flow configuration. A quartz window was provided on
the
chamber wall through which UV intensity was monitored by a UV sensor. The bulbs were fitted
-7-
-------
with a mechanical wiper type automatic cleaning mechanism, mounted on a worm gear drive.
The wiper removed deposits of materials from both the quartz sleeves and quartz window. The
wiping frequency was set to once every 10 min throughout the pilot testing. A panel display
indicated the wipe count. The effluent pipe from the reaction chamber was installed in such a
way that it could be removed for inspection of the UV lamps.
UV Pilot Operation The UV pilot unit was operated at a constant flowrate and lamp intensity
throughout a sampling event. The flowrate and lamp power levels were varied between tests.
Flowrate to the unit varied from 58 gal/min to 140 gal/min and was controlled manually using a
flowmeter and throttling valve located at the influent side of the UV pilot. Lamp intensity was
also controlled manually by adjusting the power level of the UV lamps to one of three preset
power levels. Power levels 1, 2 and 3 provided lamp power outputs of 125, 160 and 200 w UV-
C per lamp, respectively. Treated UV effluent samples were collected from a sample port
installed on the effluent discharge pipe and were analyzed for indicator bacteria (total coliform,
fecal coliform, Escherichia coli, and enterococcus), UV transmissivity, particle size, volatile
organic compounds (VOCs), semivolatile organic compounds (SVOCs), halogenated aromatic
amines (HAAs,) Whole Effluent Toxicity (WET), and Microtox. All samples, except for the
particle size samples, were discrete (grab) samples. The particle size sample was a composite
sample made up of discrete samples taken at 15-min intervals over the duration of the test run.
Absorbed dosage was also measured once per test event via a collimated beam test. The
analytical results of the samples are summarized in Table 3.
The UV pilot unit was found to be the simplest unit to operate, requiring minimal
operator attention. As was the case with Phase I, the wiper mechanism periodically became
jammed by rags or other bulk material in the wastewater. The jammed wipers were fixed easily
by resetting the wiper fault condition and reversing wiper direction. At the end of the pilot
operations, the quartz sleeves showed signs of slight reddish discoloration, possibly due to iron
deposits, at the ends of the sleeves. In addition, a slight transparent film was present over the
entire length of sleeve. In a full scale system this would likely necessitate periodic chemical
cleaning of the quartz sleeves.
4.2 Chlorine Dioxide
Chlorine dioxide has proven its capabilities as an outstanding bactericide and viricide. It
is ten times more soluble in water than C12. In contrast to C12, C1O2 does not react with ammonia
and other nitrogenous compounds to form chlorinated organics, and its disinfection efficiency is
high over a wide pH range. Due to its instability, C1O2 must be generated on-site on an as
needed basis. It may be generated on-site by: acid/sodium chlorate generation; acid/sodium
chlorite generation; chlorine/sodium chlorite generation (solution generators, gas-solid
generators); and UV radiation/sodium chlorite generation. A recent advance involving UV
radiation of sodium chlorite (NaClO2) has emerged as a new and innovative technology for C1O2
generation.
Chlorine dioxide is produced by this method through the disassociation of chlorite, a process that
requires very little energy in the generation process. The primary benefit of this generation
method compared to classical C1O2 generation methods is that C12 gas is not used in the
-8-
-------
generation process. In Phase I, the C12 (gas) - NaClO2 (solid) generation process was used.
During Phase II, the UV-chlorite generation process was intended to be used throughout the
testing. However, due to operational problems with the UV-chlorite pilot generator, the backup
gas-solid type generator was used.
Typically for wastewater treatment, aqueous disinfectants such as sodium hypochlorite
and C1O2 are injected into wastewater using a diffuser arrangement, without providing additional
mixing. However, studies on CSO disinfection performed for the US EPA in Philadelphia,
Syracuse, and Rochester demonstrated that disinfection performance could be significantly
increased and required contact times reduced by providing mixing at high velocity gradients (G)
(e.g., high-rate mixing) (US EPA 1978a and 1978b). These studies showed that values of GT (G
x contact time [T]) between 10,000 and 100,000 provided effective disinfection at contact times
of less than 5 min. Additionally, recent studies and operations experience with disinfection of
secondary effluent have shown that high-energy mechanical mixing can improve chemical
performance and minimize effluent residuals. Based upon this information, high-rate mixing
was selected for use in the pilot testing.
Chlorine Dioxide Pilot Equipment This pilot system consisted of two main components: the
contact tank system and the C1O2 generator. The contact tank system included separate flash mix
tank, contact tank and weir tank. The contact tank system was identical for the C1O2 and
Cl2/deCl2 pilots. Two separate C1O2 generators were used as part of the pilot study. Ultra Violet
Dioxide, Inc., (UVD) of Syracuse, New York, manufactured the first generator, which produced
a C1O2 liquid using an innovative UV-chlorite process. The second generator, which produced a
C1O2 gas using the C12 (gas)/chlorite (solid) process, was manufactured by CDG Technology,
Inc., (CDG) of Bethlehem, Pennsylvania. Because the UVD system is a new technology, the
system manufactured by CDG, used during Phase I was available as a backup system and was
used during Test Runs No. 1, 7 and 8.
UVD Pilot Equipment The UVD C1O2 generator generates C1O2 by passing an aqueous NaClO2
solution through a UV reactor. The C1O2 gas is then stripped out of the NaClO2/ClO2 solution
and sparked into water producing an aqueous C1O2 feedstock. Liquid C1O2 product, strength
between 864 mg/L and 2,300 mg/L C1O2 was generated up to a week prior to testing. The C1O2
product was then stored in a product tank, which was cooled to retard degradation. The product
strength was then measured prior to testing. Based on the measured product strength and the
desired dose, the C1O2 product was metered into the wastewater flash mix tank.
Several difficulties were encountered in the operation of the UVD generator. Most of the
difficulties can be attributed to the fact that the unit was still an early prototype. Chlorine
dioxide could only be prepared on a batch basis and operator attention was required throughout
the generation process. During test runs 2 through 6, volumes of gas in the product pump
suction line would periodically air bind the product metering pump, potentially decreasing the
-9-
-------
Figure 3. A process flow diagram of the UV pilot unit.
-10-
-------
Table 3. UV pilot sample parameter table.
Conventional Parameters
SS
vss
Settleable Solids
BOD
Soluble BOD
TKN
Ammonia
COD
TOC
Iron (total)
Bacterial Parameters
T. Coliform
F. Coliform
Escherichia coli
Enterococcus
Target Organic DBFs
VOCs
SVOCs
HAAs
Other
Total
Number of
Samples
76
26
26
26
26
76
76
26
26
26
152
656
488
152
80
72
68
Analytical Reference
Method
EPA- 160 .2
EPA- 160. 4
SM 20-2540 F
EPA-405.1
EPA-405.1
EPA-351.2
EPA-350.1
EPA-410. 1/410.2
EPA-415.1
EPA-200.7
SM 20-9222 B (3)
SM 20-9222 D (3)
SM 20-9222 G (3)
SM 20-9230 C
EPA-624 / NYSDEC
ASP
EPA-625 / NYSDEC
ASP
EPA-552.2
Sample Preservation ("
cool to 4° C
cool to 4° C
cool to 4° C
cool to 4° C
cool to 4° C
cool to 4° C, H2SO4 to
pH<2
cool to 4° C, H2SO4 to
pH<2
cool to 4° C, H2SO4 to
pH<2
cool to 4° C, lOmg
Na2S2O3
cool to 4° C, lOmg
Na2S2O3
cool to 4° C
cool to 4° C
cool to 4° C, ammonium
chloride
Maximum
Holding Times m
7d
7d
48 h
48 h
48 h
28 d
28 d
28 d
28 d
6 months
6h(4)
6h(4)
6h(4)
7d
5 d extraction, 40
d analysis
7d
Container
2-L poly (A)
2-L poly (A)
2-L poly (A)
l-Lpoly(B)
l-Lpoly(B)
500-mL poly (C)
500-mL poly (C)
500-mL poly (C)
500-mL poly (C)
250-mL poly
250-mL Whirl Pack (D)
250-mL Whirl Pack (D)
250-mL Whirl Pack (D)
250-mL Whirl Pack (D)
2 x 40-mL vials
4 x 1-L amber
250-mL amber
-11-
-------
UV Transmissivity (unfiltered)
UV Transmissivity (filtered) (5)
UV Collimated Beam
Particle Size
Chlorate
Chlorite
Whole Effluent Toxicity
Viral
Chlorine Residual (total and free)
Chlorine Residual (amperometric
titration)
Chlorine Dioxide Residual
Chlorine Dioxide Residual
Microtox
Total
Number of
Samples
76
76
9
9
93
93
34
101
NA
NA
NA
NA
NA
Analytical Reference
Method
NA
NA
EPA-625. 1/86.021
SM 20-2560
EPA-300 D
EPA-300 D
EPA 600/4-90/027F
NA
HACH DPD Method
SM 20-4500-CID
gas diffusion FIA and
HACH Chlorophenol
Red Method
direct measurement
SM 20-8050, ASTM
D5660-96
Sample Preservation ("
cool to 4° C
cool to 4° C
cool to 4° C
cool to 4° C, 0.25 mL 10%
NaOCl
purge w/N, gas (for
samples with C1O2), 50
mg/L EDA, cool to 4° C,
protect from light
keep cool, protect from
light
cool to 4° C, 10 mg
Na2S2O3
NA
NA
NA
NA
NA
Maximum
Holding Times m
NA
NA
14 d
14 d
28 d
14 d
48 h
1 month
NA
NA
30min
NA
NA
Container
250-mL poly (E)
250-mL poly (E)
3 x 1-L poly
500-mL poly
250-mL poly (E)
250-mL poly (F)
5 x 1-gal poly
4 x 1-L poly
NA
NA
NA
NA
NA
Note: 1. Samples were preserved immediately upon sample collection; 2. Unless otherwise noted, all holding times are from the time of sample collection; 3.
For all 9222 methods, mastication of samples must be completed prior to analysis; 4. A maximum holding time of 6 hrs for all bacterial samples was a goal. At
no time did holding times exceed 10 h; 5. Filtered UV transmissivity samples were filtered at the laboratory using a 0.45 un filter. A. These parameters were
combined in the same 2-L polyethylene container; B. These parameters were combined in the same 1-L polyethylene container; C. These parameters were
combined in the same 500-mL polyethylene container; D. These parameters were combined in the same sterilized 250-mL swirl container; E. These parameters
were combined in the same 250-mL polyethylene container; F. These parameters were combined in the same 250-mL polyethylene container.
-12-
-------
disinfectant feed rate. Also, due to leaking fittings/valves, the NaClO2 solution leaked into the
C1O2 product tank, resulting in a feedstock contaminated with up to 7,500 mg/L of chlorite. The
maximum strength of C1O2 produced was 2,300 mg/L.
CDG Pilot Equipment Because the UVD system was a new technology, the CDG C1O2
generation system used during Phase I had been used as a backup system. In this system, C1O2
gas was generated by passing a humidified 4% C12 gas in nitrogen blend through a packed
column of solid NaClO2. A rapid reaction between the C12 gas and the NaClO2 results in a C1O2
gas. The CDG unit was used to produce C1O2 during pilot test runs no. 1 and 8, and for half of
pilot test run no. 7. In the contact tank system, wastewater flowed into the flash mix tank where
C1O2 was mixed with the wastewater using a 1/2-hp, Series 32F, Gas Mastrrr, high-rate, chemical
induction mixer. After the flash mix tank, the wastewater was directed into the contact tank.
The contact tank was subdivided into 2 sections, one with corrugated longitudinal baffles and
one with flat longitudinal baffles. Each section received equal flow, which was controlled by the
adjustable v-notch weir plate located in the weir tank. The corrugated baffles were designed to
provide a headloss of 1-in. H2O. The actual headloss was found to be approximately 0.75-in.
H2O. This provided a gradient G for the corrugated baffled section of 303 s"1. In addition the
contact tanks also had a second mixer located at a contact time of 0.5 min to evaluate the
performance of single-stage versus two-stage mixing. For the mixing configuration used during
each test event and the design conditions of the contact tank system please refer to the original
CDM report (CDM, 1999).
Chlorine Dioxide Pilot Operation The C1O2 pilot system was operated at a constant flowrate
and chemical dose throughout each sampling event. Flash mixing and the addition of
disinfectant into the wastewater stream was performed in all test runs using the high-rate
chemical induction mixer in the flash mix tank. The corrugated and flat baffled tank sections
were operated for all 8 test runs, and samples were collected at three different contact times (2.7,
5 and 8 min) in each section to measure treatment performance. The majority of the sampling
was performed at the 5-min sample port in the flat baffled section, to be consistent with the
Phase I test conditions. Samples for bacterial parameters, target organic DBFs, chlorate,
chlorite, WET and microtox were collected at the 5-min sampling location. Process monitoring
for TRC and ORP was performed at a detention time of approximately nine minutes on the flat
baffled section only. Monitoring for pH and C1O2 was performed at the effluent of the weir tank.
Bacterial samples for fecal coliform and Escherichia coll were also collected at the 2.7-min and
8-min contact times in both the baffled and unbaffled sections to correlate inactivation against
contact time. The analytical results are summarized in the Results section.
Because of questions on the accuracy of the UVD C1O2 product strength measurements,
pilot test runs no. 2 through 7 may have been underdosed with C1O2. Without accurately
knowing the C1O2 product strength, a reliable relationship between applied dose and toxicity
could not be established. In order to make this correlation an additional test was performed.
Using the C1O2 product from the CDG system, wastewater was dosed with five different known
dosages of C1O2 (4, 6, 8, 10, and 12 mg/L C1O2).
-13-
-------
4.3 Chlorination/Dechlorination
Chlorine has been the most widely used disinfectant for wastewater and potable water in
the United States due to its low cost, reliable disinfection capabilities, and adequate supply.
Generally, bacteria are more susceptible to C12 than viruses. The disinfection effectiveness of
C12 is largely a function of the chemical form of the disinfecting species. Chlorine is applied to
the waste stream in molecular (C12) or hypochlorite (OC1~) form. Chlorine is available in many
forms including C12 gas and C12 products such as sodium and calcium hypochlorite. Liquid
sodium hypochlorite has become widely used for wastewater disinfection due to its reliability
and ease of handling. Sodium hypochlorite can be purchased in bulk forms of 10 to 15% of
available C12 or can be manufactured on site. Sodium hypochlorite has limited shelf-life and is
subject to loss of available C12 content by decay to C12 gas. Sufficient mixing, contact time, and
dosages are necessary to maximize the use of C12 disinfection.
Dechlorination may be accomplished through injection of a solution of sodium bisulfite
(NaHSO3) or sulfur dioxide (SO2) gas into the process flow, following the chlorination process.
Figure 4 presents Cl2/deCl2 pilot unit flow schematic. The deC!2 process is nearly an
instantaneous process. A potential problem with deC!2 is the possible depletion of dissolved
oxygen by excess sulfite ion.
Chlorination/Dechlorination Pilot Equipment The Cl2/deCl2 pilot unit mixing configuration
and sampling program was generally identical to that of the C1O2 system, with the exception of
deC!2. In this pilot system an 8 to 15% solution of sodium hypochlorite was introduced, using a
chemical metering pump, through the vacuum port on the 1/2-hp Gas Mastrrr chemical induction
mixer. Dechlorination was performed in the weir tank downstream of the flow control weirs by
injecting a 38% sodium bisulfite to the waste stream near the impeller of a small propeller mixer
to provide for deC!2 of both free and combined C12 residual. Dosages of OC1" and NaHSO3 were
controlled manually by adjusting the pump stroke/frequency. The pilot system included residual
instrumentation for continuous monitoring of DO, total residual C12, deC!2 residual and ORP.
The C12 and C1O2 contact tanks were both identical in design.
Chlorination/Dechlorination Pilot Operations Operations of the Cl2/de C12 pilot were similar
to the C1O2 pilot. The Cl2/de C12 pilot system was operated at a constant flowrate and chemical
dose throughout each sampling event. Flash mixing and the addition of disinfectant into the
wastewater stream were performed in all test runs using the high-rate chemical induction mixer
in the flash mix tank. The corrugated and flat baffled tank section were operated for all 8 test
runs, and samples were collected from three different contact times (2.7, 5, and 8 min) in each
section to measure treatment performance. The majority of the sampling was performed at the 5-
min sample port in the flat baffled section, to be consistent with the Phase I test conditions.
Samples for bacterial parameters, target organic disinfection byproducts (DBFs), chlorate,
chlorite, WET, and Microtox were collected at the 5-min sampling location. Process monitoring
for TRC and ORP was performed at approximately nine minutes on the flat baffled section only.
Monitoring of the deC!2 effluent was performed by measuring the DO and the free residual C12 at
the effluent of the deC!2 tank. Bacterial samples for fecal coliform and Escherichia coli were
also collected at the 2.7-min and 8-min contact times in both the baffled and unbaffled sections
-14-
-------
to correlate inactivation against contact time. The analytical results can be found in the Results
section.
4.4 Ozone
Ozone is a chemical oxidizing agent that has been widely used for disinfection of
drinking water systems and bleaching in the pulp and paper industry. It is an extremely strong
oxidant and is well established for its powerful antibacterial and antiviral properties (Wojtenko
et al. 2001). Ozone is a rapid disinfectant, requiring substantially less contact time than
conventional chlorination disinfection systems. Based upon research performed by the US EPA
in the 1970s and early 1980s, O3 was considered to be one of the most feasible disinfection
alternatives to C12; the other technology being UV radiation. However, there presently are few
operating facilities using O3 for disinfection of municipal wastewater. This may be attributable
to the relatively high initial capital costs associated with O3 generation equipment and the poor
operating records of previous O3 generators.
Ozone Equipment The ozone unit was a trailer mounted system manufactured by Aquifine
Wedeco Environmental Systems, Inc., (AWES) of Valencia, California. Ozone was generated
on-site and on-demand using 90% pure oxygen and a corona discharge type O3 generator.
Oxygen was supplied using an air compressor and a pressure swing adsorption oxygen
generation unit. Ozone was transferred to the wastewater using a pressure booster pump and a
gas eductor. Following the eductor, contact time for disinfection was provided by a baffled
contact tank. The tank was sized to provide a minimum detention time of 10 min at a flow of 10
gal/min. Flow out of the tank was controlled by level sensors and an automatic valve.
Ozone Operations The O3 pilot was operated at an approximately constant wastewater flowrate
and O3 feed gas concentration throughout each sampling event. Some variation in wastewater
flowrate occurred as a result of clogging of the strainer baskets. However, this variation was
generally only significant when primary influent wastewater was used as the source water. This
wastewater contained large particles of solids and waste which rapidly clogged the basket
screens. Several difficulties were encountered in the operation of the O3 pilot. Some of these
can be attributed to the use of the eductor mass transfer system rather than the technology itself.
The O3 pilot could not be operated during test runs no. 1 and 2 due to a problem with a
low differential pressure alarm. This may have been related to potential clogging of the booster
pump impellers. The pressure transmitter alarm circuit was disconnected and the unit was
successfully operated from runs no. 3 through 16. For runs no. 3 through 6, the booster pump
discharge pressure was observed to be lower than required for normal operation. Lower
pressures may have resulted in lower mass transfer efficiency in the eductor. During run no. 6,
-15-
-------
Figure 4. Chlorination/dechlorination pilot unit flow schematic.
Hcu
55
-16-
-------
the unit was subject to numerous shutdowns during the test period due to clogging of the basket
strainer and the pump impellers. Following cleaning of the booster pump, the pump discharge
pressure in runs no. 7 through 16 was much improved. In these runs, the pump discharge
pressure was generally in the range of 58 to 65 lb/in2. In addition, this particular unit was subject
to O3 gas leaks. Because of the difficulties observed with the system during Phase I, the
technology was found not to be feasible for CSO applications at this time. As a result, this
technology was not investigated during Phase II, and this Research Summary does not discuss
the ozonation study results.
5. RESULTS
5.1 Ultraviolet Irradiation
Dose-Response Relationships Scatter plots were developed for dose versus log reduction and
effluent bacterial concentrations for each bacteria group. These relationships were developed to
identify the dose required to achieve a range of bacterial log reductions and effluent
concentrations. These dose-response relationships were generated for total coliform, fecal
coliform, Escherichia coli, and enterococcus data. Blending was performed on the samples in
order to release potentially entrained bacteria by shearing the solid particles without causing
significant kills to the bacteria. Blending requirements for CSO wastewater samples were
initially developed by US EPA (US EPA 1975). An example of the dose-response relationship
generated for fecal coliform is presented in Figure 5. The figure shows the Phase I dose-
response curve along with the Phase II scatter plot. In general, these graphs demonstrate that as
UV dose increases, log reduction of bacteria also increases while the effluent mean concentration
decreases.
The dose-response relationships provide two obvious trends: a tailing-off effect for UV
effectiveness, and higher variability of wet weather data vs. dry weather data. Based on the
dose-response relationship, it appears that as dose increases in the range of 10 to 75 mW/cm2, log
reduction of bacteria also increases. However, as dose increases above the 75 mW/cm2UV,
effectiveness tails-off. The tailing effect in the dose-response data is more clearly shown in the
collimated beam data in Figure 6. Above approximately 100 mW/cm2, the lab collimated beam
dose curve becomes asymptotic to 100 cfu/100 mL of fecal coliform. The pilot unit data follows
the same trend, although more variable and with a higher asymptote.
The wet-weather runs captured during the Phase II pilot study were generally the result of
large storms. Consequently, the dose-response data show much more variability for the wet-
weather data than for the dry-weather data. The bacteria log reduction data for dry-weather tend
to be grouped together; whereas the wet-weather data are more widespread. The variability in
the wet-weather dose-response data can be attributed to the variability in the pilot influent
wastewater during wet-weather events. This is likely the result of variable solids concentration
and particle size, both known to significantly affect UV disinfection performance.
-17-
-------
Table 4 summarizes the estimated range of UV doses required to achieve corresponding
bacterial concentrations of 1,000 cfu/100 mL, 3-log and 4-log bacterial reductions. These
targeted doses were based on the Phase II and adjusted Phase I results. The dose required to
achieve the fecal coliform effluent target compares favorably with the results of Phase I, after the
UV doses were recalculated. Additionally, data from both Phase I and Phase II show that the
lowest effluent fecal coliform concentration achieved by UV disinfection was approximately 100
cfu/1 OOmL.
Table 4. Targeted UV dose.
Parameter
Total Coliform
Fecal Coliform
Escherichia coll
Enterococcus
UV Dose (mWs/cm2)
1,000 cfu/1 00ml
N/A
60-120
40-80
75
3-log
reduction
40
30
30-40
40-80
4-log
reduction
120-180
60-120
75-160
130-160
Water Quality Relationships Constituents such as suspended solids (SS) and iron absorb UV
light, thus decreasing the available light intensity within the reactor, which in turn reduces the
UV dose. Additionally, constituents such as SS limit the exposure of bacteria to UV radiation by
shielding or harboring the bacteria from exposure to the UV light.
Plots of UV transmittance versus SS and iron concentration are presented in Figure 7.
Plots of SS concentration versus log reduction and effluent fecal coliform concentration by dose
are shown in Figure 8. This relationship indicates a slight trend of reduced disinfection
effectiveness with increasing SS concentrations. This trend is likely the result of harbored
bacteria within the solids, a phenomenon that is not accounted for when measuring UV
transmittance. The phenomenon of harbored bacteria is also likely the cause of the "tailing-off'
effect. Because the bacteria are entrenched in the solids they are not exposed to the UV light,
and therefore are not destroyed. The relatively constant bacteria concentrations at the higher UV
doses could reflect a relatively constant concentration of harbored bacteria. Besides SS, no other
water quality parameters showed a positive correction to UV disinfection effectiveness. These
water quality results, as they relate to UV disinfection effectiveness, were similar to the trends
observed during Phase I.
-18-
-------
Figure 5. UV dose response relationship for fecal coliform.
Effluent ftea Coltomn Reduction v*. uv Bo*»
7.00
D.OO
0.00 20,00 40.00 60.00 BO. CO 1DO.QO 120.DO 140.00 1SO.OO 1BO.OO 200, CO
UV Dost jmWstem2)
EMusnt Fecal Ccitorm CofKanttatton w, MV Dost
1.ME+0?
1.WEH40
0,00 2©',IW 40,00 60.00 80.00 100.00 f30.00 140.00 180.00 110.00 200.00
0¥ Do» JraWafctis*)
D W*t Events -«-pliai«l.
-19-
-------
Viral Reductions All four viral runs clearly demonstrated that UV disinfection produced
nearly complete reductions of bacteriophage. As a result, no apparent relationship between dose
and phage reductions could be drawn from this data set. However, the reductions provided by
UV disinfection would likely inactivate most wastewater enteroviruses at concentrations in
CSOs. Table 5 presents the results of the viral disinfection by UV dose and phage reductions.
Because of the small number of positive observations of naturally occurring enteroviruses in the
pilot influent, the UV treatment could not be evaluated satisfactorily on the basis of the tissue
cultural infectivity assays.
Disinfection Byproducts No disinfection byproducts were detected in the UV effluent with
greater concentrations than in the pilot influent. UV has the distinct advantage of producing
little or no byproducts that may cause a concern for toxicity. Haloacetic acids and semi-volatile
and volatile organic compounds, as well as acute whole effluent toxicity were measured in the
UV
effluent. These samples were used as being representative of the pilot influent because there was
essentially no difference between the UV effluent and the influent.
Table 5. UV viral reductions.
Parameter Sample location
UV Dose (mWs/cm2)
T4&f2
Seeded influent concentration (PFU/mL)
Effluent concentration (PFU/mL)
Log reduction
S2 & X174
Seeded influent concentration (PFU/mL)
Effluent concentration (PFU/mL)
Log reduction
Run#l
43
IxlO3
15
1.9
IxlO4
150
2.9
Run #4
75
500
0
2.7
2x1 04
0
4.3
Run #7
43
3.7xl05
0
5.6
7xl05*
0*
5.8
Run #8
145
3.5xl05
60
3.9
l.lxlO6*
0*
6.0
* -XI 74 not included in these samples
5.2 Chlorine Dioxide
The C1O2 disinfection pilot was operated for eight runs at a controlled flow of 32 gal/min.
The flowrate was held relatively constant during each run. Difficulties were encountered with
the C1O2 injection system, which are believed to have resulted in lower actual applied dosages
than calculated. This became apparent when the results using the UVD system were compared
against those for the CDG system at the same calculated dose. Due to these mechanical
difficulties, the CDG C1O2 generator system was used for runs 1 and 8, and part of run number 7.
The above mentioned mechanical difficulties excluded the use of data collected during runs
number 2, 3, 4, and part of 7, while laboratory difficulties excluded the use of the data collected
during the other part of run number 7. Only data from runs number 1,5,6, and 8 were
considered valid and were used in the data analyses. This limited the amount of data and
-20-
-------
restricted the C1O2 dose range tested to 6.5 mg/L to 10 mg/L.
Dose-Response Relationships Scatter plots were developed for dose versus log reduction and
effluent bacteria concentrations for each bacteria group. These relationships were developed (for
total coliform, fecal coliform, Escherichia coli, and enterococcus data) to identify the dose
required to achieve a range of bacterial reductions and effluent concentrations. The dose versus
log reduction plot for fecal coliform is shown in Figure 7. The Phase II data exhibited a high
degree of variability; more so than the Phase I data. This variability is similar to the UV and
Cl2/deCl2 results and is at least partially attributable to more variable influent wastewater quality
resulting from the greater number of and more intense storm events in Phase II. The Phase II
data does show an increase in bacterial log reduction between 6.5 and 8 mg/L, as would be
expected, and a slight decrease between 8 and 10 mg/L. It is possible that the actual dose was
less than 10 mg/L due to the malfunctioning C1O2 injection system. Except for the total coliform
data, Phase I data shows a higher degree of inactivation at a given dose than Phase II. This could
be due to differences in wastewater quality between the two Phases.
It was difficult to discern the difference between wet-weather versus dry-weather
disinfection efficacies because of the way that the pilot units were operated. For instance, the
CDG C1O2 product was used during runs number 1 and 8, both dry-weather runs, and the UVD
C1O2 product was used during runs number 5 and 6, both wet-weather runs. This limited data set
and its variability made it difficult to differentiate between performance during wet-weather and
dry-weather events. Table 6 summarizes the estimated range of C1O2 doses required to achieve
corresponding bacterial concentrations of 1,000 cfu/100 mL, 3-log and4-log bacterial
reductions. These target doses were based on both Phase I and Phase II results. In general, the
performance of the C1O2 system was less effective during Phase II than during Phase I. For
example, the Phase II results for fecal coliform reduction were approximately 0.5 log less than
Phase I results at equivalent C1O2 dose (Figure 9).
Comparisons of these pilot results with full-scale facilities cannot be done since there are
presently no known facilities that use C1O2 to disinfect CSO or municipal wastewater. However,
these results are consistent with bench-scale and field pilot work previously done under US EPA
funded demonstrations (US EPA 1975; US EPA 1979).
Table 6. Targeted C1O7 doses.
Parameter
Total Coliform
Fecal Coliform
Escherichia coli
Enterococcus
C1O2 Dose (mg/L)
l,OOOcfu/100mL
N/A
7
6-8
6-9
3-log reduction
9
5-7
4-8
6-9
4-log reduction
N/A
8-10
6-9
>10
-21-
-------
Figure 6. UV collimated beam results.
Cmk Phase II
EtOacnt Focal Culiform fofiOTtnitien »s UV CoffinwtcU Sraw Oo« By Ev«nt
I»«nt7 —o—EvntS
Spifai Cr«k Phase II
ItlusBt Fwri CaWbm CsflMntritioo w C»fcia«eil ieam tnd PBel Dose
1.SE-HI7
1JI*W
2 I OTT+OS
'] § UiEHM i -i'
?but 11 Flint Dili Trend
E_^ i,nr*™<* | ~* s^~^
3 t "*• a
£ u 4 t J'- C *
! I !*i^
Z 1 6EfA3 * "S3 ^«
1
1JE+02
™ rontantfd
OG
ISO
s« itt
UV OIIK (toWttetam')
O BUSK II Hint Data * CoUtfluud a*aai D*t*
100
2S»
-22-
-------
Water Quality Relationships Disinfection performance is dependent upon wastewater
characteristics such as SS and BOD. Many constituents found in wastewater limit disinfection
by either exerting a disinfectant demand or shielding bacteria from contact with the disinfectant.
Suspended solids limit the exposure of the disinfectant by shielding or harboring bacteria from
contact with the disinfectant.
Similar to Phase I, trends of reduced disinfection effectiveness as a result of increased
concentrations of TKN, BOD, COD and TOC were not apparent. In general, the concentrations
of these parameters measured during Phase I and Phase II were comparable, albeit Phase II was
characterized by greater variability. Based on both Phase I and Phase II results, it does not
appear that these parameters have a significant affect on C1O2 disinfection efficiency at the
concentrations measured.
In contrast to Phase I, SS concentrations appeared to affect disinfection effectiveness; as
SS concentration increased disinfection effectiveness decreased. This relationship is depicted in
Figure 10. During Phase II, a higher percentage of the samples had SS concentrations greater
than 120 mg/L, thus offering more information regarding the effects of higher SS concentrations.
These data indicate that as SS concentration increase, disinfection effectiveness decreases.
Viral Reductions For test runs no. 1 and 7, C1O2 disinfection produced nearly complete
reductions of bacteriophage. Data from the second viral run (run no. 4) was not included in this
analysis because of the operational difficulties experienced with the C1O2 disinfection system.
Data from the last viral run (run number 8) showed lower phage reductions than the other runs.
It is not clear why the viral reduction is lower for this test run. By comparison the mean fecal
reduction during test run no. 8 was 3.0 log. Because of the relatively limited data set, no
apparent relationship between dose and phage reductions could be drawn from this data set.
However the reductions provided by C1O2 disinfection would be expected to inactivate most
wastewater enteroviruses at the concentrations in CSOs. Table 7 presents the results of the C1O2
disinfection by dose and phage reductions. Because of the small number of positive observations
of naturally occurring enteroviruses in the pilot influent, the C1O2 treatment could not be
evaluated satisfactorily on the basis of the tissue cultural infectivity assays.
Disinfection Byproducts The generation of toxic byproducts and disinfectant residuals has
become a concern for chemical disinfectants. Byproducts from the reaction of C1O2 with
wastewater, depending upon the generation process, include chlorate ion, chlorite ion, and C12.
However, C1O2 produces far fewer byproducts than C12 and is a more effective disinfectant
because of its superior penetration characteristics and bactericidal properties. The main
byproducts of C1O2 disinfection are chlorite ion and chlorate ion. The presence of these ions can
be the result of both the C1O2 generation process and reactions in the wastewater.
There is little data available on the toxicity of chlorite and chlorate. Only two toxicity
studies relating to lethal concentrations of chlorite and chlorate were found after a search of the
-23-
-------
Figure 7. UV transmittance versus SS and iron concentration.
Effluert Urtfiltered UV Tnuismfttanca v» SS Cancmtratton
SOD
GOO
EByunt Unfllwwl U¥ TfammlttwiM v» tan Concinbathsn
EOSO
-24-
-------
Figure 8. SS concentration versus log reduction and effluent fecal coliform concentration by
dose.
Effluent Fiieal Coilteim Log Iteteion v* sispendod Solids ( »8J By HV
Post
0,0
Eflhwnt Fecal Codfonn Conwufraion ¥* Sjispendwl Bonds IT 3>SJ By yV
Dw*
1.0E+OS
-25-
-------
US EPA's ECOTOX database. The first was a chlorate toxicity study predominately related to
fresh and salt-water algae species. The second was a chlorite toxicity study by the US EPA's
Office of Pesticide Programs on many fresh water species including opossum shrimp. The
opossum shrimp resulted in LC50 values of approximately 0.6 mg/L. It is important to note that
the toxicity data provided in the US EPA's ECOTOX database are from single chemical
exposures, and therefore, do not reflect synergistic affects with other chemicals. Additionally,
they may be the result of different experimental designs.
Residual chlorate ion and chlorite ion concentrations were evaluated to determine if there
was a correlation between these residuals and ClO2dose, and if the concentrations of these
byproducts exceed regulatory standards. The data does not show a trend of increasing chlorate
or chlorite ion concentrations with increasing C1O2 doses as was observed in Phase I. However,
the Phase II data set was very limited due to the operational problems experienced with the UVD
Inc. generator, and therefore, the trends observed during Phase I should be considered more
reliable.
The chlorite and chlorate data from Phase I demonstrate that a relatively small amounts
of chlorite and chlorate were found in the pilot effluent when using the CDG C1O2 product.
During Phase I, effluent chlorite generally ranged from 0.5 to 4 mg/L for applied CLO2 doses of
4 to 10 mg/L. Conversely, relatively high concentrations of chlorite and chlorate were found in
the pilot effluent when using the UVD C1O2 product. This is the result of mechanical problems
encountered with the UVD prototype C1O2 generator. These mechanical problems were
discussed in greater detail in the previous section. It is important to note that UVD Inc.
corrected these mechanical problems after the testing was complete, and produced a C1O2
feedstock with chlorite and chlorate concentrations of 44 and 210 mg/L, respectively. At the
highest C1O2 dose of 10 mg/L, these concentrations of feedstock would contribute approximately
0.22 and 1.1 mg/L, to the effluent chlorite and chlorate concentrations, respectively.
Total residual oxidant (as C12) was also evaluated to determine if there was a correlation
between TRC and C1O2 dose, and if the concentration of TRC exceeds regulatory standards.
Total residual oxidant concentrations were low; the highest observed concentration of TRC was
1.5 mg/L. It should be noted that the DPD TRC method used did not differentiate between the
various oxidizing forms of chlorine and included C12, C1O2, C1O"2 and C1O3.
Chlorine dioxide residual concentrations were measured to identify the lowest C1O2 dose
at which a residual would be detected. However, this relationship could not be developed due to
the limited data set. Measurable residual C1O2 concentrations were found at a dose of 8 mg/L,
which corresponded to runs when the CDG C1O2 unit was used. However, no C1O2 residual was
apparent when the UVD C1O2 unit was used, even at a dose of 10 mg/L. Based on water quality
data such as organics, it is unlikely that the C1O2 demand explains the lack of residual when
using the UVD C1O2 unit. It is more likely that the actual C1O2 dose using the UVD C1O2 unit
was less than 10 mg/L due to the operational problems, discussed previously.
Chlorine Dioxide Residuals Vs Dose Similar to Phase I, ORP sensors and data loggers were
-26-
-------
installed on the effluent side of the C1O2 contact tank. No relationship between ORP and C1O2
residual concentrations was apparent. Conversely, a strong correlation between ORP and TRC
was apparent, thus producing a n44early linear trend throughout the range of ORP and TRC
values that were measured. The ORP values quickly rose and stabilized at a relatively high ORP
reading (> 650 mV) and consequently the Phase I data was not suitable for developing a
relationship as a process control technique. The Phase II ORP versus TRC data shows more
potential for developing a relationship as a process control technique than the Phase I data.
More testing is needed to determine the discrepancy between Phase I and Phase II data.
Table 7. Chlorine dioxide viral reductions.
Parameter Sample Location
C1O2 Dose (mg/L)
T4&f2
Seeded influent concentration (PFU/mL)
Effluent concentration (PFU/mL)
Log reduction
MS2 & X174
Seeded influent concentration (PFU/mL)
Effluent concentration (PFU/mL)
Log reduction
Run#l
8
IxlO3
0
3.0
IxlO4
0
4.0
Run #7
10
3.7xl05
0
5.6
7xl05
0
5.8
Run #8
8
3.5xl05
5x1 03
1.9
l.lxlO6
3xl03
2.9
5.3 Chlorination/Dechlorination
The Cl2/deCl2 disinfection pilot was operated for 8 runs during the pilot study at a
controlled flow of 32 gal/min. The pilot flowrate was held relatively constant during each run.
Dose-Response Relationships Scatter plots were developed for dose versus log reduction and
effluent bacteria concentrations. These were developed for each bacterial group to identify the
dose required to achieve a range of bacterial log reductions and effluent concentrations. The
dose-response relationships for total coliform, fecal coliform, Escherichia coli, and enterococcus
data had been generated. The example of one for fecal coliform is presented in Figure 11.
Similar to the UV and C1O2 results, the Phase II Cl2/deCl2 data exhibited a high degree of
variability. This is at least partially attributable to the more variable influent wastewater quality
resulting from the greater number and more intense storm events during Phase II. The Phase II
effluent bacteria concentrations for doses of 18, 20, and 24 mg/L are generally consistent with
the results from Phase I, although in most cases the Phase II data exhibits a trend of poorer
disinfection performance than Phase I. It is interesting to note that between doses of 24 and 28
mg/L, no further decrease in effluent bacteria concentrations is observed. In some cases an
increase in effluent bacteria is apparent at the higher dose.
-27-
-------
Figure 9. Chlorine dioxide dose versus log reduction plot for fecal coliform.
EWiHWt Feeal Cot;, :-m H«4ucttpn «, GIQj Dos«
Wit & Dry Events at i-mlnute Conaei Time
7.00
O.QO
) Z
4
u a
Q
t S 1fl
1
12
CIO, DOS* {Ǥ/)}
a Witiwntsi -w-'Pttisei
EfRuent Fecal ColHtaim Concentration vs. C'.O-, Coie
Wit $ Dry Efitnte at S-m!nwte Contact Tlm»
1 OE+OS •
1 OP+9I «
t
o f iFMiJ,
;
n
• i. 1 ' °
t /v 1 °
: / X^0 I f
1 • - / ~ - ^^--^^t _^— — "i
i P 4
; Q
1 .240 6 10 1
CIO, Octi* (mg i1!)
* DtyEvwnte a Wetiwnts -w-Phswl
-28-
-------
Figure 10. Chlorine dioxide fecal coliform dose response relationship for SS.
Effluent Fecal Conform Log Rrfyctioi* v*. influent Suspended Solid*
By CJOj DOM
At 5-W.inuts Contacl Time
140 tiO
Suipendld Soldi . SomeOTiratlon
Bast; »CS jngfl A « mg/i u 10 rag/1
110
Effluent Ftcal CollfWNi Oofmndnflpn vs. faftjestt Sisp«id«l
At MHInule ConUtct TJm«
Suspended Solids
Dswe: *BJ>mg/l
lOO 111 140 1M 110
doficentradon (rtig.'ii
a 10 mglt
-29-
-------
The factors which may contribute to these observations include variations in wastewater
quality, pH, and temperature. Wastewater temperature, for instance, is known to significantly
affect the rate of reaction of hypochlorous acid (free C12) with ammonia to form
monochloramine. During Phase I, the mean wastewater temperature was 11.6 °C, versus a
temperature of 20.9 °C for Phase II. The colder winter temperatures would impede the formation
of monochloramine, which has approximately 25 time less germicidal efficiency as free C12.
This effect is corroborated to some degree by the C12 residual measurements during Phases I and
II. As seen in Table 8, the lower doses during Phase I, when the mean water temperature was 9.3
°C cooler, produced significantly higher free residuals than were observed for the higher applied
doses of Phase II.
The decrease in treatment performance at a dose of 28 mg/L corresponds with a change
in the chlorine-ammonia chemistry. The poor performance at this dose occurred during test run
no. 5, a wet weather event when the mean ammonia concentration dropped to 5.6 mg/L versus an
average dry weather concentration of 13.6 mg/L. The resulting C1:N weight ratio of 5.0
corresponds with the typical inflection in the chlorine-ammonia breakpoint curve. Beyond the
break, further C12 addition will actually decrease the measured residual as dichloramine and
organochloramines are formed. Although dichloramine is believed to be approximately twice as
germicidal as monochloramine, organochloramines are generally nongermicidal. This and other
factors may partially explain the tailing effect observed at these high doses.
Table 8. Mean residual chlorine vs. applied dose for Phases I and II.
Applied C12 Dose
(mg/L)
12
16
20-21
24-25
18
20
24
28
Mean Total Residual Chlorine mg/L (DPD Method)
Phase I
9.0
13.5
17.0
20.0
-
-
-
-
Phase II
-
-
-
-
5.5
7.0
10.0
13.0
The Phase II log reduction plots also depart from the Phase I data at the applied doses of
24 and 28 mg/L. This can be attributed to the tailing effect observed in the effluent bacteria
concentration plots and to the dilution of the untreated bacteria concentrations during wet
weather events. The Phase II data for doses of 24 and 28 mg/L include data from wet weather
events no. 3 and 5. In these events, the mean influent fecal coliform, for example, was only
1.06E+05 and 8.32E+05 respectively versus an average dry weather concentration of 4.3E+06.
-30-
-------
So for similar treated effluent concentrations, the log reduction may be lower due to dilution of
the influent.
Table 9 summarizes the estimated C12 dose required to achieve corresponding bacterial
concentrations of 1,000 cfu/100 mL and 3- and 4-log bacterial reductions. These doses were
based on the results of Phase I and Phase II. The Phase II dose required to achieve the fecal
coliform effluent target did not exactly match with the results of Phase I. The Phase II results for
fecal coliform reduction were approximately 0.5 log less than the Phase I results. Similarly, the
effluent fecal coliform concentration results were approximately 0.5 log greater than the Phase I
results. These results are consistent with work previously done under US EPA supported efforts
(US EPA 1975, US EPA 1979). Similar log reductions of bacteria resulted for the range of doses
of C12 tested under this pilot study.
Table 9. Targeted C12 doses.
Parameter
Total Coliform
Fecal Coliform
Escherichia coli
Enterococcus
C12 Dose (mg/L)
1,000 cfu/lOOmL
N/A
20
17-22
>22-28
3-log reduction
25
12
12
>22-28
4-log reduction
N/A
20-28
20-28
N/A
Water Quality Relationships Generally, disinfection performance is dependent upon
wastewater characteristics such as SS, ammonia, and BOD. Many constituents found in
wastewater limit disinfection by either exerting a C12 demand or shielding bacteria from contact
with C12. Suspended solids limit the exposure of embedded bacteria by shielding them from
contact with the disinfectant. Plots of SS concentration versus log reduction and log
concentration of fecal coliform for the different doses are shown in Figure 12. These
relationships show that there is no apparent trend between disinfection effectiveness and SS.
This is particularly evident at a dose of 24 mg/L for SS concentrations ranging from 200 to 500
mg/L.
Residual chlorate and chlorite concentrations were evaluated to determine if there was a
trend between these residuals to sodium hypochlorite dose and the relative concentrations of
these byproducts. Only a small percentage of the sodium hypochlorite remained in the form of
chlorite, and essentially all of the sodium hypochlorite was converted into chlorate. Chlorate
concentration increased with increasing dose hypochlorite concentrations. These chlorate
concentrations are much higher than the chlorate concentrations measured during Phase I. In
some cases the chlorate concentrations represent more than 80% of the sodium hypochlorite dose
concentrations.
-31-
-------
Figure 11. Chlorine dose-response relationships for fecal coliform data.
Effluent Fwsai Conform Reduction «• C:|j
Wit & Dry Ewifcs srt 5-Mlnutet Cantaci Time
18
* Dry E¥tnts
1S
a, Daw fmtffj
O WW Ewnts
1.0E*tt!
Efflu«it F«aJ Coltonm Canomtratiorl Vs< Cls Dos*
Wet ( Dry Events a! 5-Mlnu?e5 Contact Tim«
+ OryEvttria
Ol, Dose (ngl)
o WetEwtnfss
-*- Phase 1
-32-
-------
Oxidation-reduction potential versus C12 dose and TRC relationships were developed to
determine the practicality of using ORP technology to control disinfection dosing processes.
ORP is not very sensitive to changes in chlorine residual at the levels required for CSO
disinfection. For the range and magnitude of the TRC values measured during Phase I and Phase
II for Cl2/deCl2, ORP probes did not appear to work effectively because of the lack of sensitivity
at higher TRC values.
In general, the concentration of the haloacetic acid compounds increased after
disinfection with C12; however, the magnitude of this increase was very small (on the order of 10
to 50 ug/L). Bromochloroacetic acid and dichloroacetic acid were the haloacetic acid
compounds with the greatest increase in concentration after disinfection. However, the absolute
magnitude of these concentrations is low. The US EPA's ECOTOX database did not contain
information regarding bromochloroacetic acid but did include LC50 results for dichloroacetic
acid. The LC50 results for dichloroacetic acid based on freshwater fish test species was on the
order of 5 mg/L. It is important to note that the toxicity data provided in the US EPA's
ECOTOX database are from single chemical exposures, and therefore, do not reflect synergistic
affects with other chemicals. Additionally, they may be the result of different experimental
designs. A few semi-volatile and volatile organic compounds appeared to be produced during
the C12 disinfection process. These compounds are listed in Table 10. Again, the absolute
magnitude of these concentrations is low, and it is unlikely that these compounds will exceed
discharge criteria or play a role in effluent toxicity because of the very low concentrations. Over
78 haloacetic acids, and semi volatile and volatile organic compounds were measured during the
Phase II pilot study. Table 12 presents only the compounds that increased after disinfection with
C12 for one or more runs.
5.4 Toxicity
Due to operational problems with the UV-chlorite C1O2 generator, the C1O2 effluent
toxicity data for test runs no. 2 through 6 are not qualified. During these test events, the C1O2
feedstock from the UV-chlorite generator was high in chlorite and the C1O2 strength could not be
determined accurately (Santos et al., 2000). This resulted in effluent wastewater chlorite
concentrations as high as 55 mg/L and high chlorite likely was a significant contributor to
effluent toxicity. As reported in the US EPA ECOTOX database, chlorite exhibits chronic
toxicity to opossum shrimp at an LC50 of 0.6 mg/L. Since the high chlorite levels were due to
equipment operational problems, the toxicity observed during these runs may not be indicative of
C1O2 itself or of a properly operating C1O2 generator. To account for this, an additional C1O2 test
event (test event no. 9) was performed using the CDG Inc., C12 (gas)/sodium chlorite (solid)
generator.
Toxicity to opossum shrimp or sheepshead minnow in the untreated influent wastewater
was observed in runs no. 1, 4, and 5 with LC50 values ranging from 66 to 76% (Santos et al.,
2000). Somewhat higher toxicity effects were observed in the UV and Cl2/deCl2 effluent for
runs no. 1 and 4 (LC50 of 39 to 52% effluent). However, when compared to the variability
observed in the WET results, these values are generally within the observed influent toxicity. By
comparison, results of WET analyses on field duplicate samples showed a mean relative percent
-33-
-------
difference
(RPD) between duplicates of 24%. The WET data for runs no. 2 and 7 show a slight toxic effect
for the UV and Cl2/deCl2 effluents compared to no toxicity on the influent wastewater. However
again, the 95% confidence intervals for the data show that there may be no difference between
the influent and effluent data for these samples.
In contrast to the UV and Cl2/deCl2 effluents, the C1O2 effluent did show significant
toxicity to opossum shrimp in runs no. 1, 2, 6, and 7 as compared to influent toxicity. LC50s in
these test events ranged from less than 6% effluent up to 26% effluent. This could not be
attributed to influent toxicity, which had an LC50 of 82% effluent for run no. 1 and greater than
100% effluent for runs no. 2, 6, and 7. It is believed that the toxicity of the C1O2 effluent
observed during this test runs was at least partially due to the high chlorite levels mentioned
previously. This was assessed by plotting relative toxicity (i.e., influent LC50 minus effluent
LC50) of the C1O2 effluent versus effluent chlorite concentration. A strong correlation between
chlorite concentration and effluent toxicity was observed. One anomaly is apparent in this trend;
no toxicity was observed in the effluent sample with the highest chlorite concentration (55
mg/L). This data point is probably erroneous. The C12 (gas)/sodium chlorite (solid) generator
was used in the toxicity analyses for test event no. 9. In this test event, the effluent chlorite
concentration varied from less than 0.2 mg/L up to a maximum of 5.8 mg/L, and no toxicity was
observed in the WET tests up to a C1O2 dose of 10 mg/L. Therefore, it appears the effluent
toxicity observed in prior test runs was related to high chlorite concentrations resulting from an
improperly operating generator.
The field microtox data may still be used to identify relative trends in toxicity within a
CSO event. One can see two trends in this data. First, the influent wastewater generally shows
higher mortality than the UV effluent and similar, though in some cases higher, mortality as the
Cl2/deCl2 effluent. This indicates that the toxicity observed in these effluents is most likely
associated with that of the untreated wastewater. Second, the C1O2 effluent overall shows higher
mortalities than the other two pilot units or the untreated influent wastewater.
6. COST COMPARISON
During the Phase I pilot study, conceptual level cost projections were prepared for each
disinfection technology for comparison purposes, with the goal of recommending a technology
for implementation at the Spring Creek CSO Storage Facility. The Phase II pilot study results
served to verify the Phase I result; as such, the assumptions and approach used for the original
cost comparison were applicable. Costs for each disinfection technology were prepared on a
common flow basis and were prepared for a range of flow rates experienced at Spring Creek
CSO Storage Facility. This approach shows the sensitivity of cost to flow rate, and allows
independent comparison of technology costs at similar flow rates.
Equipment capital costs were developed for peak design flow conditions of 1,250 cfs
(800 mgd), 2,500 cfs (1,600 mgd), and 5,000 cfs (3,200 mgd) for a duration of 4 hours. The
5,000 cfs flow rate represents approximately the maximum facility inflow for the 5-year storm.
-34-
-------
Table 10. Organic compounds concentration, chlorine effluent vs. pilot influent.
Parameter
VOCs:
Bromodichloromethane
Dibromochloromethane
Bromoform
Chloroform
SVOCs:
4-Methylphenol
Benzoic Acid
HAAs:
Bromochloroacetic acid
Dibromoacetic acid
Dichloroacetic acid
Monobromo acetic acid
Monochloroacetic acid
Trichloroacetic acid
Detection
Limit d-ig/1)
10
10
10
10
10
10
1
1
1
1
1
1
Concentration (|Jg/l)
Run#l
Infl.
ND
ND
ND
ND
ND
ND
ND
1
10
1
3
17
Eff.
ND
ND
ND
ND
22
69
ND
1
35
1
1
19
Run #2
Infl.
ND
ND
ND
ND
ND
ND
1
1
4
1
1
7
Eff.
3.5
1.5
0.9
10
ND
ND
6
1
27
1
1
14
Run #3
Infl.
ND
ND
ND
ND
ND
ND
1
1
1
24
1
2
Eff.
ND
ND
ND
ND
ND
ND
10
11
22
2
1
6
Run #4
Infl.
ND
ND
ND
ND
ND
3
1
1
5
1
3
9
Eff.
10
14
12
ND
4
23
7.4
1
64
2.6
2
10.4
Run #5
Infl.
ND
ND
ND
ND
ND
ND
1
1
1
1
1
4
Eff.
22
18
ND
18
ND
ND
20
4
42
2
1
13
Run #6
Infl.
ND
ND
ND
ND
ND
ND
1
1
1
1
1
4
Eff.
ND
ND
ND
ND
ND
ND
12
3
38
1
5
8
Run #7
Infl.
ND
ND
ND
ND
ND
ND
1
1
9
1
1
12
Eff.
11
18
14
ND
ND
ND
23
9
52
1
1
13
Run #8
Infl.
ND
ND
ND
ND
ND
ND
1
1
7
1
1
9
Eff.
ND
ND
ND
ND
10
71
17
6
54
1
1
13
Notes:
1. "Infl" and "Eff" denote pilot influent and C1O2 pilot effluent, respectively.
2. ND denotes non-detected.
-35-
-------
Figurel2. Plots of SS concentration versus log reduction and log concentration of fecal
coliform for the different doses.
Effluent Fecal Coliform Concentration yg. Suspended Solids
By PI Dose at 5-m Inutes Contact Time
100
200 300 400 $00
Suspended Sotlds , Concentration JmgJIJ
Lftlucnt Fe&al Coliform Laf Reduction vx. SU£p«flde<) Solids
By Cl] Dose at 5*mlnutM Contact Time
200 300 400
Suspended Solids Concentration (mgj'li
o2iǤfl
800
-36-
-------
The original design report for Spring Creek CSO Storage Facility (Greeley and Hansen, 1962)
calculated a peak 5-year storm inflow of 3,750 cfs. However, this flow rate was based upon a
runoff coefficient of 0.65. As identified by CDM in the Spring Creek Stabilization Study (1990),
a more realistic runoff coefficient would be 0.90 - 0.95.
This would result in a peak inflow of approximately 5,000 cfs. This flow rate and
selected duration are also consistent with inflows observed during the Stabilization Study. The
lower flow conditions were selected at reasonable fractions of the 5-year condition. Operating
costs were developed based on an estimate of approximately 40 events/year producing inflow to
the Spring Creek CSO Storage Facility, at a volume of 15 million gallons (MG) treated per
event. This condition was selected based upon a review of the design inflow volumes (Greeley
and Hansen, 1964).
Costs were developed for UV, ozone, chlorine dioxide, and chlorination/dechlorination.
Due to the limited effectiveness and high power consumption of the E-Beam pilot unit, the E-
beam technology as tested was not considered feasible for CSO disinfection. By comparison, the
power usage for E-beam based on the pilot unit was approximately 3.5 kW/gpm (or 2,430
kW/mgd) while the power usage for UV at 4-log reduction was 0.0325 kW/gpm (or 22.6
kW/mgd). Therefore, since the E-beam technology was not considered feasible for CSO, costs
were not developed for the technology. The cost projections were developed for a 4-log
reduction of fecal coliform and included the following process options:
• UV - Medium-pressure, high-intensity lamps
• Ozone - Oxygen feed ozonation with eductor or side stream venturi type mass transfer
configuration
• Chlorine Dioxide - High-rate mixing, generation of C1O2 using the chlorine (gas)/sodium
chlorite (solid) process, with onsite generation of chlorine gas via the acidification of
NaOCl with HC1, use of emergency gas scrubber for potential chlorine gas, and 5 minute
contact time (provided by existing basins at approximately 5,000 cfs)
• Chlorination/Dechlorination - High-rate mixing, use of 15% sodium hypochlorite and
38% sodium bisulfite, and 5 minute contact time (provided by existing basins at
approximately 5,000 cfs)
The cost projections are shown in Table 11. The table presents estimated capital, annual
O&M, and total annualized costs. Annualized costs were prepared on the basis of a 20-year
period at an 8 percent discount rate. The capital costs only include the costs for the basic process
equipment associated with each technology and do not include:
The construction of additional basin tankage or structures for contact/disinfection,
• Modifications to the existing basins,
• Building expansion to house disinfection equipment,
Support equipment or facilities, for example: additional power supply equipment, HVAC
equipment, and plumbing equipment
As shown in Table 11, chlorination/dechlorination and chlorine dioxide are significantly
less costly than either UV or ozone. Due to the intermittent nature of CSOs, disinfection
-37-
-------
technologies like chlorination and chlorine dioxide, which are less capital intensive with higher
O&M costs are favored over high capital cost technologies with lower O&M costs.
It is important to note that for other CSO facilities, the cost for construction of
disinfection contact tanks for the chlorination/dechlorination and chlorine dioxide alternatives
would need to be considered and may make UV somewhat more attractive. It is also important
to note that the cost of contact tankage for chlorine dioxide could be almost 40% less than
chlorination/ dechlorination. This difference is attributed to chlorine dioxide's greater
bactericidal properties and solids penetration characteristics than those of chlorination, as
demonstrated during the contact time analysis performed for the Phase II contact time.
Chlorine dioxide cost projections were developed using the demonstrated chlorine
(gas)/sodium chlorite (solid) process. This system was chosen as more reliable for cost
projections because it is in use at a number of full-scale installations. Because the UV-chlorite
chlorine dioxide generator is currently in the prototype status, costs using this technology were
not developed.
7. CONCLUSIONS
7.1 Wastewater Quality
During the Phase I and Phase II pilot studies five disinfection technologies, UV, C1O2,
C12, O3, and E-Beam were piloted to determine their effectiveness in reducing bacteria levels in
water representative of the CSO at the Spring Creek CSO Storage Facility. These pilots were
tested during wet and dry events. In general, the pilot influent water quality was variable but
representative of CSO water quality from the Spring Creek CSO Storage Facility. The variation
in wastewater temperature between Phase I (mean of 11.6 °C) and Phase II (mean of 20.9 °C) is
believed to have had a significant impact on the performance of C12 disinfection. While the
majority of CSO discharges from Spring Creek are likely to occur during the late summer and
early fall months, discharges have also occurred during the winter and early spring months.
Therefore, it was an added benefit to characterize the difference in performance resulting from
temperature effects.
To achieve a four-log reduction of fecal coliform and fecal coliform effluent
concentrations less than 1,000 colony forming units/100 mL (cfu/100 mL) required doses for
UV, O3, C1O2, and C12 of 60-80 mWs/cm2, 24 mg/L, 8-10 mg/L, and 20-28 mg/L, respectively.
The spread in disinfectant doses for each technology reflects the variation in performance
between Phase I and Phase II.
-38-
-------
Table 6-1. Cost projections.
Technology
Peak Design
Flow (cfs)
Capital Costs
Annualized
Capital Costs
Annual O&M
Cost
Total
Annualized
Costs
Conceptual Level Disinfection Costs ($)
Chlorination/ Dechlorination
1,250 2,500 5,000
912,000 1,045,000 1,219,000
93,000 107,000 124,000
255,000 255,000 255,000
348,000 362,000 379,000
Chlorine Dioxide
1,250 2,500 5,000
695,000 1,159,000 1,932,000
70,000 119,000 196,000
294,000 294,000 294,000
364,000 413,000 490,000
Ozone
1,250 2,500 5,000
19,221,000 24,560,000 30,539,000
1,957,000 2,502,000 3,111,000
534,000 587,000 657,000
2,491,000 3,089,000 3,768,000
UV
1,250 2,500 5,000
48,052,000 67,272,000
87,774,000
4,894,000 6,852,000 9,592,000
248,000 497,000 992,000
5,142,000 7,349,000 10,584,000
Notes:
1. Costs are present worth in 2000 dollars.
2. Capital costs are based upon sizing to meet peak design flow and a 4-log reduction in fecal coliform.
3. Capital costs are for installation of Spring Creek and are for process equipment only. Costs do not
Include additional contact tankage (if required) or support facilities
4. Annual operating costs are based upon an assumed typical 40 CSO events/year at a volume treated
of 15 million gallons per event.
5. Annualized costs are based upon a period of 20 years at an interest rate of 8%.
-39-
-------
7.2 Treatment Performance
Four bacteria indicators were used as a measure of the effectiveness of each of the
disinfection technologies; namely total coliform, fecal coliform, Escherichia coli, and
enterococcus. Kills of each of the indicators, in terms of log reduction and concentration, were
related to dose for each of the disinfection technologies. Currently, there are no effluent bacteria
criteria established for the Spring Creek CSO Storage Facility or for other CSO facilities.
However, these targeted bacterial reductions were selected as conservative estimates of levels
that could be met by the technologies and that may represent permit criteria. Generally, all the
tested disinfection technologies, with the exception of E-beam, were able to effectively provide
bacterial reductions of 3 to 4 logs. Chlorination/dechlorination, C1O2, and O3 at the doses tested
were able to provide these levels of disinfection over the full range of wastewater quality tested.
UV disinfection effectiveness tended to drop off at higher SS concentrations (e.g., SS greater
than approximately 150 mg/L). This was attributed to lower effective penetration of UV due to
harboring of bacteria in solids.
Fecal coliform and Escherichia coli exhibited similar dose-response relationships.
However, total coliform and enterococcus generally required higher doses to achieve the same
level of inactivation as that for fecal coliform and Escherichia coli. This was observed in all
technologies except for the E-beam, where the inactivation results were inconclusive. The
existing receiving water quality standards for Spring Creek and Jamaica Bay address only total
and fecal coliform. However, in the future regulators may use indicator bacteria that are more
specific to human waste such as Escherichia coli and enterococcus for water quality standards.
In the absence of specific densities indicating health risk, this information can only be preserved
for future reference at this time.
The UV and C1O2 technologies provided nearly complete reductions of bacteriophage.
However, the viral inactivation data for the C1O2 system was limited to only two out of the four
runs due to operational problems. Of the valid data considered, the effluent concentrations of
bacteriophage ranged from non-detect to 60 pfu/mL. Low influent concentrations of the seeded
phage limited the maximum log reduction that could be observed. The log reduction of
bacteriaphage ranged from 1.9 to 6.0. Because of the low concentrations of naturally occurring
enteroviruses in the pilot influent, the UV could not be evaluated satisfactorily on the basis of the
tissue culture infectivity assays. However, based upon the reductions of the marginal
concentrations found and upon the bacteriophage results, it is likely that theses technologies
would inactivate most natural enteroviruses found in wastewater at concentrations on the order
of 106pfu/mL.
UV disinfection achieved 4-log bacteria reduction but at extremely high dosage levels
owing to the impediments of poor water quality. UV effectiveness tended to be reduced by high
SS concentrations (e.g., greater than 150 mg/L). Additionally, UV effectiveness tended not to
increase at doses greater than 75 mWs/cm2; a phenomena known as "tailing-off."
Ozone disinfection can be accomplished but at dosage levels more than one and one-half
times that of C12. However, the O3 pilot unit did not include a contactor design appropriate for
-40-
-------
the wastewater conditions tested. Thus, the required O3 dosages may have been less if a more
applicable O3 dissolution/contactor system were provided. An O3 disinfection system would
require contact chambers other than the tankage that presently exists at Spring Creek.
Chlorine disinfection included dechlorination to eliminate residual C12. Both Cl2/deCl2
can be accomplished using the existing tanks at the Spring Creek CSO Storage Facility. High-
rate mixing can be added to the head end of the tanks. Chlorine dioxide disinfection can be
accomplished at doses on the order of 30% of the required C12 dose.
Chlorination/dechlorination and C1O2 were determined to be the most cost effective
technologies for application to Spring Creek. However, neither C1O2 generation method tested is
currently feasible within New York City; the C12 gas solid sodium chlorite generation method
because of its use of C12 gas, and the UV-sodium chlorite generation method because of its
developmental status as a prototype. The capital costs for UV and O3 were significantly more
expensive than Cl2/deCl2 or C1O2. For other CSO facilities that do not have existing tanks for
contact time, UV could be somewhat more cost competitive.
In the case of C1O2, there is no significant increase in disinfection performance beyond a
contact time of 3 min. This is in contrast to the chlorination results, which show a greater
dependence on contact time and required five minutes for comparable kills. The difference is
attributed to ClO2's greater bactericidal properties and solids penetration characteristics than
those of chlorination. The results of this study confirm the optimum contact times for C1O2 and
Cl2/deCl2 of 3 and 5 min respectively, originally determined in the Syracuse and Rochester
studies (US EPA, 1979a and 1979b). Chlorination/dechlorination and C1O2 were determined to
be the most cost effective technologies for application at this facility. Further development of
the UV-chlorite C1O2 generator is required before reliable costs for this technology can be
developed.
Comparison of the dry weather performance data for single and two-stage mechanical
mixing configurations for chlorine disinfection implied a slight increase in disinfection
effectiveness for two-stage mixing. The wet weather data, with its higher variability, was
excluded from this comparison as it appears to have obscured the effects of 2nd stage mixing.
The evaluation of single versus 2-stage mixing could not be performed for the C1O2 system due
to the limited data from the field operational problems.
7.3 Disinfection Residuals and Toxicity
Selected disinfection effluent residuals and byproducts, namely C1O2, chlorate, chlorite,
TRC, volatile and semivolatile organics, haloacetic acids, were monitored to relate these
residuals to disinfectant dose. UV disinfection had the distinct advantage of producing no
byproducts. This is in contrast to C12 and C1O2, which produced increased levels of TRC,
chlorate, chlorite and haloacetic acids in the effluent. The slightly increased haloacetic acid
concentrations were considered insignificant. The increased TRC, chlorate and chlorite
concentrations were directly related to increased C12 and C1O2 dose.
-41-
-------
No additional toxicity was observed in the UV effluent as compared to the UV pilot
influent. In contrast, there were occurrences where the C1O2 effluent was considerably more
toxic than the pilot influent. An attempt was made to correlate this toxicity with the specific
disinfection byproducts, in particular TRC, chlorate and chlorite, but no correlation could be
made. It is likely that the increased effluent toxicity is directly related to influent toxicity (i.e.,
influent water quality) or a synergistic effect of the disinfectant residuals, which could not be
measured. Although the concentrations of TRC, chlorate and chlorite did not cause a concern
for effluent toxicity, this relationship should be revisited when establishing C1O2 dose for
specific sites.
Effluent TRC was generally below 0.1 mg/L following deC!2, as compared to a receiving
water quality standard of 0.0075 mg/L. This value of deC!2 effluent TRC reflects the practical
quantitation limit of the process instrumentation used. Lower TRC values could not be
quantified. Often, the deC!2 effluent TRC instrumentation displayed a negative value indicating
the presence of excess bisulfite. Residual C12 was also monitored in the C1O2 effluent. However,
these TRC values include all oxidizing species of C12 and the possible presence of free and
combined C12 could not be differentiated from C1O2, C1O2", and C1O3".
7.4 Chlorine Dioxide Generation
The method of generating C1O2 must be considered when selecting the appropriate
disinfection process. The chlorine gas/solid sodium chlorite generation method was tested
during the Phase I and Phase II pilot studies. Although this pilot unit was reliable, the use of
chlorine gas (either with chlorine cylinders or with on-site C12 gas generation) in this process
may limit its application in residential and urban areas, including New York City. The UV-
sodium chlorite solution generation method was also tested during the Phase II pilot study. This
method had the distinct advantage of not using or generating chlorine gas in the generation
process. However, this technology is currently in the prototype stages of development and
would need to be developed as a full-scale unit to be considered further. The UV-chlorite
generator from the UVD Inc., was a prototype unit.
8. SUMMARY
Pilot testing of disinfection technologies on CSO wastewater, as part of the upgrade to
the Spring Creek CSO Storage Facility, was performed in two phases. Phase I was performed
over 16 pilot test events from December 1996 through March 1997. This testing evaluated the
performance of five high-rate disinfection technologies: UV, O3, C1O2, Cl2/deCl2, and E-Beam.
The results from Phase I were presented in a final report dated November 1997. The purpose of
the Phase II pilot testing was to address data gaps identified in the Phase I study, provide
additional wet-weather data, and perform additional research that was beyond the scope of the
original study. The Phase II pilot program was performed from August through November 1999
and evaluated the performance of disinfection with UV, C1O2, and Cl2/deCl2.
Influent and effluent from each pilot unit were analyzed by a certified laboratory for
-42-
-------
bacterial and conventional wastewater quality parameters: VOC's, SVOCs, HAAs, and toxicity.
Four indicator bacteria were used as a measure of the effectiveness of each of the four
disinfectant technologies: total coliform, fecal coliform, Escherichia coli, and enterococcus.
Bacteria kills for each of the indicator bacteria, in terms of 3 to 4 log reduction, were related to
dose for each of the four technologies.
Based upon the technologies evaluated in these pilot studies, only Cl2/deCl2 was
recommended for CSO disinfection at the Spring Creek CSO Storage Facility and other New
York City CSO facilities. While C1O2 was superior in effectiveness and similar in cost to
Cl2/deCl2, the generation technology for C1O2 which avoids the need for gaseous C12 needs
further development. Because an effective Cl2-gas-free process of C1O2 generation has not been
proven to be reliable and, because C12 gas cannot be transported within New York City,
disinfection with C1O2 cannot be recommended for use within New York City at this time.
However, the City and its engineers should remain apprized of advances in alternative
disinfection technologies, such as UV-chlorite C1O2 generation. These advances may make these
technologies more effective, both in terms of cost and disinfection effectiveness.
While UV and O3 treatment were technically viable, the study showed that, given the
intermittent nature of CSO treatment and the high peak flows involved, the high capital cost of
these technologies makes them cost prohibitive. Electron beam disinfection did not meet the
treatment goals.
The original, detailed reports from the two studies, Phase I and Phase II, were produced
in 1997 and 1999 by CDM of Woodbury, New York, and CDM's Subcontractor Moffa &
Associates, a unit of Brown & Caldwell, of Syracuse, New York, for the NYCDEP and the US
EPA (Urban Watershed Management Branch, Edison, New Jersey). Funding for the studies was
provided by NYCDEP (NYCDEP Capital Project No. WP-225) and partial funding for Phase II
only was provided by US EPA (US EPA Purchase Order No. 7C-R394-NTLX). Partial funding
for Phase I only was provided by New York Power Authority (NYPA) and the Electric Power
Research Institute, both of New York City, New York.
-43-
-------
REFERENCES
Camp Dresser & McKee. 1997. "Spring Creek AWPCP Upgrade. CSO Disinfection Pilot Study-
Part I" - Final Report.
Camp Dresser & McKee. 1999. "Spring Creek AWPCP Upgrade. CSO Disinfection Pilot Study-
Part II" - Final Report.
Greeley and Hansen, Engineers. 1964. Report on Use of Hypochlorite for Treatment of Storm
Overflow, Spring Creek Marginal Pollution Control Project.
Santos, L. N., K. J. Smith, D. Davis, and E. Delva. 2000. Acute Toxicity Testing of Alternative
High-Rate Disinfection Technologies for CSO Treatment in New York City. Presented at the
WEF Disinfection 2000 Specialty Conference in New Orleans.
Stinson, M. K., Field, R., Moffa, P.E., Goldstein, S. L., Smith,K. J., Delva, E. 1998. High-Rate
Disinfection Technologies for Wet-Weather Flow (WWF), in Proceedings of Water Environment
Federation's Specialty Conference entitled "Advances in Urban Wet Weather Pollution
Reduction," Cleveland, Ohio, June 28-July 1, 1998.
U. S. Environmental Protection Agency. 1975. Bench Scale High-Rate Disinfection of
Combined Sewer Overflows with Chlorine and Chlorine Dioxide. EPA/670/2-75-021,
Cincinnati, OH.
U. S. Environmental Protection Agency. 1979a. Combined Sewer Overflow Abatement Program
Rochester, NY, Vol. II Pilot Plant Evaluations; EPA/600/2-79-03 Ib, Cincinnati, OH.
U. S. Environmental Protection Agency. 1979b. Disinfection/Treatment of Combined Sewer
Overflows, Syracuse, New York; EPA/600/2-79-134, Cincinnati, OH.
Wojtenko, L, M. K. Stinson, and R. Field. 2001 a. Performance of Ozone as a Disinfectant for
Combined Sewer Overflow. Manuscript to be published in the Critical Reviews in
Environmental Science and Technology, 31(4): 295-309.
Wojtenko, L, M. K. Stinson, and R. Field. 2001b. Challenges of Combined Sewer Overflow
Disinfection with Ultraviolet Light Irradiation, Critical Reviews in Environmental Science and
Technology, 31(3): 223-239.
-44-
------- |