EPA/600/R-16/112 I September 2016
www.epa.gov/homeland-security-research
United States
Environmental Protection
Agency
Disinfection of biological agents in the
field using a mobile advanced oxidation
process
* M
"'T»
• ^
- . /* —
Office of Research and Development
Homeland Security Research Program
-------
This page is intentionally left blank
-------
EPA/600/R-16/112
September 2016
Disinfection of biological agents in the field using a mobile advanced
oxidation process
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
This page is intentionally left blank
-------
Disclaimer
The U.S. Environmental Protection Agency (EPA), through its Office of Research and
Development's National Homeland Security Research Center (NHSRC), funded and managed
this project under Interagency Agreement (IA) DW-14-92385901-0 with the United States
Geological Service (USGS). This report has been peer and administratively reviewed and has
been approved for publication as an EPA document. It does not necessarily reflect the views of
the EPA. Mention of trade names or commercial products does not constitute endorsement or
recommendation for use of a specific product.
Questions concerning this document or its application should be addressed to:
JeffSzabo, Ph.D., P.E.
National Homeland Security Research Center
Office of Research and Development
U.S. Environmental Protection Agency
26 W. Martin Luther King Drive
Cincinnati, OH 45268
szabo.i eff@epa. gov
1
-------
Acknowledgements
Contributions of the following organizations to the development of this document are
acknowledged:
Kansas State University, Department of Biological and Agricultural Engineering
Department of the Army, Ft. Riley, Kansas
CB&I, LLC
2
-------
Table of Contents
Disclaimer 1
Acknowl edgem ents 2
Table of Contents 3
List of Tables 4
List of Figures 4
Acronyms and Abbreviations 5
Executive Summary 6
1.0 Introduction 7
1.1 Proj ect Background and Obj ectives 7
2.0 Methods & Procedures 9
2.1 AOP System Design 9
2.2 AOP Treatment Process 11
2.3 Experimental Procedure 12
2.4 Evaluation obj ectives 13
3.0 Sampling and Measurement Procedures 15
3.1 Sampling Containers, Holding Times and Preservation 15
3.2 Preservation Procedure for Microbial Samples 15
3.3 Analytical Laboratories 15
3.4 Sampling and Analytical Procedures 16
4.0 Results and Discussion 18
4.1 Water Quality and E. coli Disinfection Results 18
4.2 Comparison of data grouped by solids and flow rate 21
4.3 Statistical Analysis using ANOVA 23
5.0 Conclusions 25
3
-------
List of Tables
Table 1: Experimental Design Parameters 13
Table 2: Summary of Critical Experimental Parameters 13
Table 3: AOP System Operating Parameters 14
Table 4: Water Quality Parameters 14
Table 5: Sample Containers, Preservation Methods, and Holding Times for Grab Samples 15
Table 6: Analytical Methods Used to Analyze Grab Samples 16
Table 7: Summary of Experiments 17
Table 8: Water Quality Sample Results for Tap Water, Pond Water, and Lagoon Water 18
Table 9: Single Factor ANOVA Results for Flowrate 23
Table 10: Single Factor ANOVA Results for Total Suspended Solids 24
List of Figures
Figure 1: Schematic diagram of the pilot-scale AOP system with sampling ports 9
Figure 2: Medium pressure UV lamp system 10
Figure 3: Cone diffuser used for ozone concentration with national pipe threat taper (NPT)
stainless steel (SS) pipe 11
Figure 4: Log inactivation of E. coli at high flowrate (6 gpm) for tap water and pond water (refer
to Table 7 for definition of the samples) 20
Figure 5: Log inactivation of E. coli at low flowrate (4 gpm or less) for tap water, pond water,
and lagoon water (refer to Table 7 for definition of the samples) 20
Figure 6: Percent inactivation of E. coli at 0-60, 60-100, 100-160 and >160 mg/L total suspended
solids (TSS) 21
Figure 7: Percent inactivation of E. coli at high (6 gpm) and low (4 gpm or less) flow rates 22
4
-------
Acronyms and Abbreviations
ANOVA
Analysis of Variance
AOP
advanced oxidation process
COD
chemical oxygen demand
EPA
Environmental Protection Agency
gpm
gallons per minute
H2O2
hydrogen peroxide
H02"
hydroperoxide ion
LR
log reduction
MPN
most probable number
mS/cm
microsiemens per square cemtimeter
02
oxygen
03
ozone
•OH
hydroxyl radical
ORD
Office of Research and Development
ppm
parts per million
PSA
pressure swing adsorption
scfh
standard cubic feet per hour
TDS
total dissolved solids
TSS
total suspended solids
UV
ultraviolet
5
-------
Executive Summary
The Army's Net Zero Initiative is an energy-conservation program that focuses on energy as well
as water and waste usage procedures. All Net Zero projects are geared toward helping the
military installation or community become more sustainable and resilient, with an emphasis on
taking a systems approach. Net Zero projects must advance the state of the science and are
focused on three general topic areas: water, energy, and waste, and the nexuses among them.
This project examined the inactivation and/or removal of biological contaminants in dirty wash
water using a portable ozone-UV AOP process. The strain of E. coli used in these experiments is
not a biological warfare agent, but acts as a surrogate for certain of the vegetative biological
agents such as the enterohemorrhagic strain designated E. coli 0157:H7.
When operated at lower flow rates (4 gpm [15.1 L/min] or less), inactivation of E. coli by AOP
treatment ranged from 4 to 6 log. Increasing flow rate to 6 gpm reduced inactivation to 1 to 2
log. Decreased flow rates resulted in a longer periods of exposure to the AOP treatment, and
which may have permitted more E. coli contact with the disinfection process. The decrease in
flow rates from 6 gpm (22.7 L/min) to 4 gpm (15.1 L/min) appears small, but the results
demonstrate a significant change in the rate of E. coli inactivation. The data suggests that
increasing contact time via recirculation may be necessary for flow rates above 6 gpm (22.7
L/min).
Statistical analyses conducted using Analysis of Variance (ANOVA) suggests no significant
relationship between the level of E. coli inactivation and the amount of suspended solids in the
source water. However, the ANOVA results do suggest a significant relationship between
inactivation and flow rate, which can be influenced by disinfectant contact time. Although E. coli
inactivation was not complete at either flowrate, the once-through flow system could be adjusted
to recirculate water for additional treatment. If the AOP was utilized in the field to disinfect and
reuse wash water for vehicle washing, adequate contact time would be needed to ensure
personnel are not contaminated. Due to time constraints, additional testing was not possible, but
the benefits could be examined in future research. Finally, the equipment used in this study is
scalable. Treated volumes were 100 to 150 gallons in this study, but the UV and ozone units
could be sized to handle more flow if larger volumes were generated during washing activities.
6
-------
1.0 Introduction
1.1 Project Background and Objectives
The Army's Net Zero Initiative is an energy-conservation program that focuses on energy as well
as water and waste usage procedures. All Net Zero projects are geared toward helping the
military installation or community become more sustainable and resilient, with an emphasis on
taking a systems approach. Net Zero projects must advance the state of the science and are
focused on three general topic areas: water, energy, and waste, and the nexuses among them.
Net Zero seeks to reduce water consumption and improve water reuse on military installations
throughout the world. The U.S. Army and the U.S. Environmental Protection Agency's (EPA)
Office of Research and Development (ORD) are currently partnering to promote and
demonstrate innovative water reduction and reuse technologies on Army installations in support
of the Army's Net Zero Initiative. In 2011, Fort Riley, the source of the waste water used in this
study, was selected to participate in the Army's Net Zero Initiative as one of six Net Zero Water
Pilot Installations. A Net Zero Water Installation limits the consumption of freshwater resources
and returns water back to the same watershed so as not to deplete the groundwater and surface
water resources of that region in quantity and quality.
One area of interest is treating and potentially reusing the large volumes of wash water used to
clean military vehicles. Wash racks, or areas where military vehicles are washed with large
volumes of potable water, generate waste water contaminated with oil, grease, some metals and
mixtures of suspended solids (dirt and mud). Reusing wash rack water can be difficult due to the
contamination, but reuse would become even more difficult if disinfection of biological
contaminants were needed. Biological contamination could come from sources such as untreated
sewage if, for example, a military vehicle were near an open sewer during combat or exercises.
It could also come from deliberate contamination with a biological warfare agent while in
theater. Access to the wash rack water provides a unique opportunity to evaluate disinfection of
biological agents in the field with waste water that could hinder the disinfection process.
The use of a mobile disinfection system can support community water conservation goals and is
key to the military for locations, domestically and in theater, where copious volumes of fresh
water are not readily available. In theater, immediate access to chemical disinfectants such as
chlorine bleach or chloride dioxide may not be available. Furthermore, transport of large
amounts of such chemicals could prove impractical or hazardous. In situations where chemical
disinfection is impracticable, disinfection technologies such as ultraviolet light and ozone are
7
-------
favorable since they do not need chemicals or reagent, and power can be supplied by a mobile
generator. Ozone and UV light are powerful disinfectants on their own, but when ozone is
exposed to UV light, hydrogen peroxide and hydroxyl radicals are formed, both of which are
also potent disinfectants. The in situ production of a potent oxidizer such as the hydroxyl radical
at concentrations strong enough to disinfect water is known as an advanced oxidative process
(AOP).
This project will examine the inactivation and/or removal of biological contaminants in wash
water using a portable ozone-UV AOP process. The AOP equipment is mounted on a trailer that
can be towed with a pick-up truck. Wash water will come from the wash racks at Ft. Riley,
Kansas, and Escherichia coli will act as the biological contaminant. The strain of E. coli used in
these experiments is not a biological warfare agent, but acts as a surrogate for certain of the
vegetative biological agents such as the enterohemorrhagic strain designated E. coli 0157:H7.
Data from these experiments will help decision makers in the Army determine if vegetative
biological agents are disinfected to a degree that the wash rack water could be reused for further
vehicle washing or some other use. In addition, the wash rack water is representative of water
washed from cars or structures after an outdoor contamination event. Therefore, this data may
be applicable to a scenario where a wide area biological contamination event occurs, and dirty
water must be disinfected before being disposed of in a sewer.
8
-------
2.0 Methods & Procedures
2.1 AOP System Design
The AOP trailer system consists of a 1-inch stainless steel pipe loop system, a variable speed
recirculation pump, a medium pressure ultraviolet (UV) lamp and a low pressure UV lamp, an
oxygen (O2) concentrator, an ozone (O3) generator, an ozone injection system, and an O3
destructor (Figure 1). Influent E. coli samples were taken from the blend tank used to feed the
AOP unit. Water quality samples for determining TSS, pH, etc. were taken just after initial
entrance to the unit. Effluent samples for bacteria and water quality were taken from the
sampling port immediately in front of the location where the treated water was discharged.
Water with biological agents was exposed to the ozone, UV light and products generated from
the UV-ozone reaction as it passed through the AOP unit.
Medium
Air Discharge
Lamp
Air
u2 u3 ^3
Concentrator Generator Compressor
Low Pressure
' UV Lamp
Dechlorinated
Tap Water
Venturi
Destructor
Brewer Site #1
Groundwater
Cone
Diffuser
Contactor
Blend Tank
Treated Water
Discharge
SP
SP
SP
Iron
Eater
Carbon
Adsorber
•Blend tank sample port (E. •Influent sample port (water 0 Ozone sample port • Effluent sample port (E. coli
coli) quality samples) and water quality samples)
Figure 1: Schematic diagram of the pilot-scale AOP system with sampling ports.
UV radiation was provided by the medium pressure UV reactor (Aquionics InLine 20 UV
System, Aquionics, Inc., Erlanger, KY) in Figure 2 (the low pressure lamp in Figure 1 was not
used in this study). Ozone was generated using an O2 concentrator and an O3 generator. The O2
concentrator separates O2 from compressed air through a pressure swing adsorption process. The
pressure swing adsorption process uses a molecular sieve (a synthetic zeolite), which adsorbs
nitrogen and other impurities from the air at high pressure and desorbs them at low pressure.
The O2 concentrator is designed for a maximum airflow rate of 6.6 standard cubic feet per hour
9
-------
(scfh) (187 L/hr). The O2 is then fed into the O3 generator. In the reaction chamber of the O3
generator, the feed gas is exposed to multiple high-voltage electrical discharges, producing O3.
The O3 is injected into the system through a venturi-type, differential pressure injector (Mazzei*
3/4-inch (1.9 cm) MNPT Model 684, Mazzei Injector Company, LLC., Bakersfield, CA) located
on the discharge side of the system recirculation pump (3/4-horsepower (0.56 kW) G&L Pump
NPE/NPE-F, Xylem Inc., Rye Brook, NY). When the contaminated water enters the injector
inlet, it is constricted towards the injection chamber and emerges as a high-velocity jet stream.
The increase in velocity through the injection chamber results in a decrease in pressure, thereby
enabling O3 to be drawn through the suction port and entrained into the motive stream. The
venturi is assisted by an ozone compressor (Dia-Vac® pump, Air Dimensions, Inc., Deerfield
Beach, FL) to allow the system to operate at lower differential pressures while maintaining a
high ozone concentration in the system. The ozone concentrations are further increased by the
use of an ozone cone diffuser shown in Figure 3. Excess O3 is converted back to O2 using an O3
destruct unit before it is vented into the atmosphere. The recirculation pump is connected to a
variable-speed controller (1AB2 AquaBoost® II Controller, Goulds Pumps, Seneca Falls, NY),
which enables the flow rate in the loop to be set to any desired value. Maximum flow through the
system is 12 gpm.
Figure 2: Medium pressure UV lamp system.
10
-------
8" PVC
Pipe
Existing 3/4"
^NPTSS Pipe
3/i" Cone
Opening
7" x 4'4" ft
cone
Downward
Flow Direction
Existing 3/4"
"NPTSS Pipe
Figure 3: Cone diffuser used for ozone concentration with
national pipe threat taper (NPT) stainless steel (SS) pipe.
2.2 AOP Treatment Process
The medium pressure mercury vapor UV lamp installed in the AOP system provided UV-C
radiation at an emission spectrum between 200 nm and 300 nm with a power requirement of 0.9
kW and a UV dose >10 mJ/cm2 The UV unit had one setting, so UV conditions were constant
for all experiments in the study. Preliminary tests were performed by running carbon-filtered tap
water and ozone through the AOP system to test the capacity of the ozone generator and to
determine the ozone concentration in the AOP system. The setting of the ozone generator was
adjusted during the preliminary tests to achieve the target ozone concentration of approximately
4 to 6 mg/L in the AOP system. The levels of ozone in effluent samples were never definitively
established due to the high reactivity of O3 and the presence of turbidity and organic constituents
in the samples. Settings for the ozone generator were left at the highest values possible and the
presence of ozone in the effluent was recorded throughout testing.
The AOP disinfection technology is UV irradiation combined with O3. Due to the high molar
extinction coefficient of ozone, UV radiation can be applied to ozonated water to form highly
11
-------
reactive hydroxyl radical (*OH). Because photolysis of O3 generates H2O2, the UV/O3 process
involves the disinfection mechanisms present in O3/H2O2 and UV/H2O2 AOPs. For instance,
H2O2 in conjunction with O3 can enhance the formation of *OH. H2O2 is a weak acid that
partially dissociates into hydro-peroxide ion (HO2") in water. The HO2" ion can rapidly react
with O3 to form *OH. Meanwhile, hydroxyl radicals are produced from the photolytic
dissociation of H2O2 in water by UV radiation. Disinfection can occur either by direct photolysis
or by reactions with *OH.
2.3 Experimental Procedure
The investigation consisted of the treatment of E. coli in clean tap water, dirty water from Fort
Riley wash racks, and naturally sourced water from a local runoff collection pond using the
mobile AOP trailer. All tests were conducted on the Kansas State University campus in the
Biological and Agricultural Engineering Department workshop. Water from the wash racks was
used directly without dilution to establish whether turbidity interfered with disinfection. Carboys
of water from the wash racks at Ft. Riley were collected as needed along with water from a local
runoff pond. For reference, Ft. Riley is approximately 15 miles from the Kansas State University
campus.
Preparing and running the AOP trailer for an individual test required approximately 2 hours per
run with 24 hours of preparation between tests for bacteria propagation and final enumeration.
Before experiments began, the AOP system was flushed for 20 minutes with tap water.
Experiments were conducted by filling the feed tank with E. coli at an initial microbial density of
lxlO6 most probable number (mpn)/ml. Water was fed to the AOP unit at two different rates, 6
gallons per minute (gpm) (22.7 L/min) or 4 gpm (15.1 gpm). During experiments where flow
was maintained at4 gpm (15.1 L/min), flow periodically dipped to between 3 and 3.5 gpm (11.4
to 13.2 L/min). Therefore, results at the 4 gpm (15.1 L/min) flowrate are reported as 4 gpm
(15.1 L/min) or less.
One sample was taken from the feed tank to determine the initial E. coli concentration (Ti).
Water exiting the AOP unit (effluent samples) were sampled at the last sampling point before the
water left the AOP unit (Ce). Samples were taken at 1, 5, 10, 15, and 20 minutes after the
contaminated water feed to the AOP unit had started. Disinfection was assessed by examining
the log reduction (LR) of samples taken at 1, 5, 10, 15, and 20 minutes compared to the initial
microbial density in the feed tank using the following equation:
Ce
LR = -log —
* i
12
-------
Disinfection was also assessed by calculating percent reduction of E. coli in water through the
following equation:
T- — C
%Reduction = —- x 100
* i
Table 1 summarizes the primary experimental design parameters for AOP disinfection of E. coli.
Table 1: Experimental Design Parameters
Parameters
Designed Values
Source water
Pond/Lagoon water, Dechlorinated tap water
Dilution water
Pond/Lagoon water, Dechlorinated tap water
Target contaminant
Escherichia coli
Concentration of contaminant
106 to 107mpn/ml
AOP method
UV irradiation/03
Type of UV lamp
Medium-pressure UV lamp
UV intensity
preset level kept constant
Ozone concentration in the AOP unit
4 to 6 mg/L
Temperature range
20 to 23 °C
Flow rates
4 gpm (15.1 L/min) (or less) and 6 gpm (22.7 L/min)
Recirculation ratio
None (once-through flow)
Test Duration
20 minutes
gpm = gallons per minute; mpn = most probable number
2.4 Evaluation objectives
Critical measurements (those key to the study), sampling location, reporting units, and sampling
frequency are summarized in Table 2.
Table 2: Summary of Critical Experimental Parameters
Measurement
Reporting
Unit
Sampling
Location
Measurement Purpose
E. coli
One sample from the blend
tank (T); AOP System
mpn/ml Effluent (E) at 0, 1, 5, 10,
15, and 20 minutes after the
start of a test run.
Primary microbial
contaminant for study
Ozone
Outlet sampling port, 2 grab
sampling events per test run
mg/L (at the beginning and end of
the test run); ozone sample
port
Disinfectant concentration
E = effluent; mpn = most probably number; T = blend tank
13
-------
The information in Table 2 highlights critical parameters for experiment. The initial
concentration of E. coli was compared to concentrations in the effluent to evaluate how much
inactivation took place. The presence of ozone, while difficult to measure precisely in the
effluent, was also measured from a sampling port in the AOP system when clean tap water was
flowing. This verified that ozone was being produced at 5.8 mg/L in the AOP unit, which was
within the expected range of 4 to 6 mg/L.
Table 3 summarizes the AOP system operating parameters, reporting units, sampling type,
sample locations, and sample frequencies.
Table 3: AOP System Operating Parameters
Inline AOP
Measurements
Reporting
Unit
Sample
Type
Sampling
Location
Sampling Frequency
Temperature
°C
Analog gauge
reading
On-line gauge
2 readings per test run (at the beginning and
end of the test run)
Flow rate
gpm
Digital flow meter
reading
On-line meter
2 readings per test run (at the beginning and
end of the test run)
Water pressure
psi
Analog gauge
reading
On-line gauge
2 readings per test run (at the beginning and
end of the test run)
Air flow into the ozone
scfh
Flow meter
On-line meter
2 readings per test run involving ozone (at
generator
the beginning and end of the test run)
C = AOP influent; E = AOP effluent; gpm = gallons per minute; psi = pounds per square inch; scfh = standard cubic
feet per hour, T = Blend Tank
Table 3 lists AOP system operation parameters that were monitored during test run. Temperature
was influenced by the daily ambient conditions, but was stable between 20 and 23 °C. Flowrate,
water pressure, and air flow were determined by inline sensors on the AOP system. Maintaining
consistent pressure, temperature and ozone generator air flow between each experiment provided
uniformity by which to compare results.
The measurements in Table 4 are indicative of the influent water quality used during each test
(tap water, pond water or wash rack water). Correlations of these measurements with inactivation
were used to compare the impact of water quality on inactivation.
Table 4: Water Quality Parameters
Measurement
Reporting
Unit
Sample
Type
Sampling
Location
Sampling Frequency
*TDS
mg/L
Sample from supply
tank
Mixing Tank
1 sample per test run
*Conductivity
m S/cm
Sample from supply
tank
Mixing Tank
1 sample per test run
*Total N
ppm
Sample from supply
Mixing Tank
1 sample per test run
14
-------
tank
*Total P
ppm
Sample from supply
tank
Mixing Tank
1 sample per test run
COD
mg/L
Sample from supply
tank
Mixing Tank
1 sample per test run
pH
Standard
unit
Sample from supply
tank
Mixing Tank
1 sample per test run
TSS
mg/L
Sample from blend
tank, influent and
effluent
Mixing Tank
10 sampling events per test run (T, CO, C5,
CIO, C15, C20, E5, E10, E15, E20)
C = AOP influent; COD = chemical oxygen demand; E = AOP effluent; m S/cm = micro Siemens per centimeter;
T=Blend Tank; TDS = total dissolved solids; TSS = total suspended solids
*Conducted by Kansas State University's Soil Testing Lab
3.0 Sampling and Measurement Procedures
3.1 Sampling Containers, Holding Times and Preservation
Sampling containers, preservation techniques, and holding times for grab sample measurement
are presented in Table 5. Aliquots of each sample were deposited into the proper containers and
the appropriate preservation technique were applied in accordance with the guidelines in Table 5.
Table 5: Sample Containers, Preservation Methods, and Holding Times for Grab Samples
Parameter
Sample Container
Preservation
Method
Holding Time
E. coli
Sterile 200-ml glass
sample bottle
Cool to 4 ± 2 °C
24 hours from collection
Ozone
200-ml glass bottle
None
Samples analyzed immediately
in the field
pH
200-ml glass bottle
Cool to 4 ± 2 °C
Samples analyzed immediately,
or held for no more than 4 hours
TSS, TDS, Total N,
Total P, Conductivity
200-ml glass sampling
bottle
Cool to 4 ± 2 °C
Samples analyzed immediately,
or held for no more than 48
hours
TDS = total dissolved solids; TSS = total suspended solids
3.2 Preservation Procedure for Microbial Samples
Microbial samples from the supply tank and AOP unit influent/effluent were collected in 200-ml
glass sampling bottles. Once the bottles were full the samples were immediately analyzed or
placed in a refrigerator at 4 ± 2 °C until analysis within 24 hours (see table 5).
3.3 Analytical Laboratories
All analyses and measurements listed in Tables 2 and 3 were conducted at the Kansas State
15
-------
University Soil Testing Lab. E. coli strain K-12 was obtained from EPA. The stock culture
obtained from EPA was stored at 4 °C, and sub-cultured in tryptic soy broth before experiments.
3.4 Sampling and Analytical Procedures
Analytical procedures are summarized in Table 6. When collecting a grab sample, the sample
tap was opened and water allowed to flow for approximately 10 seconds to flush the sampling
port.
Table 6: Analytical Methods Used to Analyze Grab Samples
Parameter
Units
Method
Citation
E. coli
mpn/ml
9221 B,C
Rice EW, Baird RB, Eaton
AD, Clesceri LS (editors).
Standard Methods for
Examination of Water and
Wastewater, 22nd Edition.
Washington DC: APHA,
AWWA, WEF.
Ozone
mg/L
4500-03-B
Standard Methods for
Examination of Water and
Wastewater, 22nd Edition
PH
pH units
150.1
U.S. Environmental Protection
Agency (EPA), Methods for
the Chemical Analysis of
Water and Waste, March
1983. Cincinnati, OH: EPA.
EPA/600/4-79-020
TSS
mg/L
SM2540D
Standard Methods for
Examination of Water and
Wastewater, 22nd Edition
*TDS
mg/L
SM 2540C
Standard Methods for
Examination of Water and
Wastewater, 22nd Edition
COD
mg/L
SM 5200D/Hach 8000
Standard Methods for
Examination of Water and
Wastewater, 22nd Edition
*Conductivity
US/cm
SM 2510
Standard Methods for
Examination of Water and
Wastewater, 22nd Edition
*Total N
mg/L
USGSWRIR 03-4174
Patton CJ, Kryskalla JR.
Methods of Analysis by the
U.S. Geological Survey
National Water Quality
Laboratory—Evaluation of
Alkaline Persulfate Digestion
as an Alternative to Kjeldahl
Digestion for Determination of
Total and Dissolved Nitrogen
and Phosphorus in Water.
Denver: USGS. USGS WRIR
03-4174
*Total P
mg/L
USGS WRIR 03-
4174/EPA 365.2
USGS WRIR 03-4174
COD = chemical oxygen demand; mpn = most probable number
16
-------
* Conducted at the Kansas State University Soil Testing Lab ('http://www.agronomv.k-state.edu/services/soiltesting/'l
Samples were labeled in accordance with the following identification scheme: date, sample
location, sample time, and experiment number. Temperature, flow and pressure readings were
recorded 2 times per test run (at the beginning and the end of the test run). The information in
Table 7 lists the source for each test batch of water, its characteristic properties, and flowrate
maintained during treatment.
Table 7: Summary of Experiments
Supply
Test
Source
Flowrate
TSS
Volume
Run
Run
Water
(gpm)
(mg/L)
(gal)
Time
TW2
Tap
6
0
100
10
LW1
Lagoon
4
197
100
20
LW2
Lagoon
4
121
100
20
LW3
Lagoon
3.5-4
70
100
20
PW10
Pond
6
52
150
20
PW11
Pond
6
110
150
20
PW12
Pond
6
70
150
20
PW2
Pond
6
49
100
10
PW3
Pond
5.5-6
65
150
20
PW5
Pond
6
682
150
20
PW6
Pond
3
155
150
20
PW7
Pond
6
50
150
20
PW8
Pond
3
278
100
20
PW9
Pond
3
176
100
20
TW 1
Tap
4
67
100
20
TW3
Tap
4
210
100
20
gpm = gallons per minute; LW = lagoon water (wash rack), PW = pond water, TSS = total suspended solids; TW =
tap water
17
-------
4.0 Results and Discussion
4.1 Water Quality and E. coii Disinfection Results
Table 8 shows the water quality data for each sample reported by the Kanas State University Soil
Testing Lab. Note that the lagoon water samples from the Ft. Riley wash racks were on average
lower in TDS, TSS, Total N and Total P. On average, TSS and TDS were 3-4 times higher in
pond water than lagoon water, and total N and P were 8-12 times higher. Before experiments
began, it was assumed that the parameters in Table 8 could contribute to disinfectant demand and
inhibit disinfection. Since the wash rack water was more dilute in these constituent (TDS, TSS,
Total N, Total P) than the pond water, more pond water experiments were conducted so that the
impact of more concentrated samples could be evaluated.
Table 8: Water Quality Sample Results for Tap Water, Pond Water, and Lagoon Water
Test
TSS
(mg/L)
TDS
(mg/L)
Conductivity
(m S/cm)
Total N
(ppm)
Total P
(ppm)
COD
(mg/L)
pH
TW2
0
0
NA
NA
NA
NA
7
PW2
38
NA
NA
NA
NA
NA
_
PW3
65
648
0.93
11.0
0.9
123
8
PW5
682
569
0.81
15.9
1.66
150
8
PW6
155
616
0.88
13.5
1.22
142
8
PW7
52
571
0.82
15.7
1.17
143
8
PW8
278
591
0.84
17.4
1.36
150
7
PW9
176
601
0.86
16.0
1.23
150
8
LW1
197
356
0.51
4.12
0.33
47
8
LW2
121
368
0.53
4.41
0.34
60
8
LW3
70
365
0.52
4.00
0.29
37
8
PW10
52
573
0.82
10.0
1.01
145
8
PW11
110
591
0.84
12.7
1.46
150
8
PW12
70
604
0.86
12.2
1.31
155
8
TW1
67
NA
NA
NA
NA
NA
NA
TW3
120
NA
NA
NA
NA
NA
NA
COD = chemical oxygen demand; LW = lagoon water (from the wash racks); PW = pond water; TDS = total
18
-------
dissolved solids; TSS = total suspended solids; TW = tap water; NA = Not Analyzed
Figures 4 and 5 show the amount of E. coli inactivation achieved in the AOP system at high (6
gpm [22.7 L/min]) and low (4 gpm [15.1 L/min] or less) flowrates. Note that these figures do
not include the E. coli inactivation sample taken from the AOP system at 1 minute. In 11 out of
the 16 experiments, it was noticed that 4 to 5 log reductions of E. coli was observed in the
volume of water that had been treated at the 1 minute time point, with consistently lower log
removal occurring in the samples from the remaining treated water. It was eventually
determined that flushing the AOP system with tap water before the experiment resulted in
residual chloraminated water lingering in the AOP system. Chloramine levels in the local tap
water were typically 2 to 2.5 mg/L. E. coli samples taken at 1 min had been in contact with this
water before it was flushed out, and the high log reductions observed resulted from disinfection
with monochloramine, not the AOP process. Samples after 1 min reflect E. coli disinfected by
the AOP process only.
The data in Figure 4 shows the disinfection performance of the AOP process for flow rates at 6
gpm (22.7 L/min). Except for the PW7 experiment, E. coli inactivation ranged from 1 to 2 log
over the course of the 20 minute experiment. Figure 5 shows inactivation of E. coli at a flowrate
of 4 gpm (15.1 L/min) or less. E. coli inactivation ranged from 4 to 6 log over the course of the
20 minute experiment. The lower flowrates may have permitted more E. coli contact with UV,
ozone and hydroxy radicals in the AOP system.
In Figures 4 and 5, the experiments using PW 6 and 7 resulted in log reductions of approximately
9.5 log, which was higher than the other studies. These samples has TSS levels of 155 and 50
mg/L respectively, which are the first highest and third lowest TSS value tested. It is possible
that mixing in the AOP unit was more efficient during these tests or that the ozone generator
produced more ozone in the piping that in other studies. It should also be noted that calculation
of 9+ log removal was possible due to the nature of the Colitert test used to detect coliforms.
The test requires 100 ml of water, in tests including these two, the initial E. coli density in the
feed tank was between 107 to 108 MPN/ml, which is slightly higher than the desired target
concentration. When analyzing these 100 ml volume with 108 MPN/ml, the total sample has
1010 MPN. The Colitert test can detect down to 1 MPN in a 100 ml volume, which allowed
calculation of 9-log removal in the PW 6 and 7 tests.
19
-------
Figure 4: Log inactivation of E. coli at high flowrate (6 gpm) for tap water and pond water
(refer to Table 7 for definition of the samples).
PW 6
PW 9
LW 1
LW 2
LW 3
TW 1
TW 3
5 min
10 min
15 min
20 min
Figure 5: Log inactivation of E. coli at low flowrate (4 gpm or less) for tap water, pond
water, and lagoon water (refer to Table 7 for definition of the samples).
20
-------
4.2 Comparison of data grouped by solids and flow rate
Figure 6 shows the percent inactivation for water samples grouped by TSS. The various groups
represent 3-5 tests with each grouping containing experiments at both flow rate ranges. For 0-60
mg/L the reduction averaged 93.2% with similar rates for 60-100 mg/L and 100-160 mg/L at
96.0% and 95.2%, respectively. The highest TSS levels or >160 mg/L experienced 98.4%
inactivation. The error bars represent standard deviation within the data groups. Reduction values
were based on the average inactivation for all associated effluent sampling times. TSS values
ranged from 0 mg/L to 682 mg/L among the 16 water samples used in this study. The
overlapping standard deviation for log reduction within each grouping of TSS levels suggests
that suspended solids were not a significant factor that influenced inactivation.
Figures 7 shows E. coli inactivation experiments grouped by the flow rate. By reducing the
flowrate to 4 gpm (15.1 L/min) (or less) the level of inactivation was increased by 2 to 3 logs.
The two groupings, 6 gpm (22.7 L/min) and 4 gpm (15.1 L/min), each consist of 8 individual
tests and their average reduction. The error bars represent the standard deviation of inactivation
within the two groups. The standard deviation error bars do not overlap, which suggests that
there is a significant difference between the two groups.
105%
100%
•¦P 95%
OJ
>
"+¦»
u
(D
= 90%
"o
u
uj
S? 85%
80%
0-60 mg/L 60-100 mg/L 100-160 mg/L
TSS Concentration (mg/L)
>160 mg/L
Figure 6: Percent inactivation of E. coli at 0-60, 60-100,100-160 and >160 mg/L total
suspended solids (TSS).
21
-------
Figure 7: Percent inactivation of E. coli at high (6 gpm) and low (4 gpm or less) flow rates.
22
-------
4.3 Statistical Analysis using ANOVA
In Table 9, flowrate (independent variable) was compared against the level of inactivation
(response variable) to determine if inactivation is dependent on flow rate using single factor
Analysis of Variance (ANOVA). The high F value indicates greater variation between the two
groups rather than within the sample groups indicating flowrate is a significant contributor to
inactivation. Flowrate impacts contact time, or the time of exposure to the AOP treatment
process. The connection between increased disinfectant contact time and higher inactivation is a
well-established principle. By reducing the flowrate, even marginally by 2 gpm (7.6 L/min), the
rate of inactivation increased substantially. In a system requiring additional contact time, the
alternative to reducing flowrate would be re-treating a batch of water, or recirculating water
through the treatment system.
Table 9: Single Factor ANOVA Results for Flowrate
SUMMARY
Groups
Count)
Sum
Average
Variance
Low (4 gpm or less)
8
37.42194
4.677743
1.360293
High (6 gpm)
8
12.54557
1.568197
1.484671
ANOVA
Source of Variation SS Df MS F P-value F crit
Between Groups 38.67711 1 38.67711 27.18988 0.000131 4.60011
Within Groups 19.91475 14 1.422482
Total 58.59186 15
Count = # of experiments tested; Df = degrees of freedom; F = F statistic; F crit = critical F value MS = mean
square; P-value = probability; SS = sum of squares
Variation between groups is lower than variation within groups of high and low TSS (Table 10).
The low F statistic illustrates this relationship indicating that TSS does not have a significant
influence on inactivation. The variation within the data shows that whether or not TSS is
elevated does not influence effectiveness of the AOP process to inactivate E. coli. This is
interesting to note since increased turbidity in the water should block UV light to some extent.
The data suggests that the UV light was still able to react with ozone to create sufficient amount
of hydroxyl radical, or that ozone alone is the dominant disinfectant in the AOP process and was
unaffected by the increased turbidity.
23
-------
Table 10: Single Factor ANOVA Results for Total Suspended Solids
SUMMARY
Groups
Count
Sum
Average
Variance
Low (0-70 mg/L, Avg:
52 mg/L)
High (110-682 mg/L,
Avg: 230 mg/L
00 00
19.15929
30.80822
2.394912
3.851028
3.220706
3.937974
ANOVA
Source of Variation
SS
df
MS
F
P-value F crit
Between Groups
Within Groups
8.481098
50.11076
1
14
8.481098
3.57934
2.369459
0.146022 4.60011
Total
58.59186
15
Count = # of experiments tested; df = degrees of freedom; F = F statistic; F crit = critical F value; MS = mean
square; P-value = probability; SS = sum of squares
24
-------
5.0 Conclusions
When operated at lower flow rates (4 gpm [15.1 L/min] or less), inactivation of E. coli by AOP
treatment ranged from 4 to 6 log. Increasing flow rate to 6 gpm reduced inactivation to 1 to 2
log. Decreased flow rates resulted in a longer periods of E. coli exposure to the AOP treatment,
and increased the time available for 'OH to form. The decrease in flow rates from 6 gpm (22.7
L/min) to 4 gpm (15.1 L/min) irrespective of the TSS appears small, but the results demonstrate
a significant change in the rate of E. coli inactivation. The data suggests that increasing contact
time via recirculation may be necessary for flow rates above 6 gpm (22.7 L/min).
Statistical analyses conducted using ANOVA suggests no significant relationship between the
level of E. coli inactivation and the amount of suspended solids in the source water. However,
the ANOVA results do suggest a significant relationship between inactivation and flow rate,
which can be influenced by disinfectant contact time. Although E. coli inactivation was not
complete at either flowrate, the once-through flow system could be adjusted to recirculate water
for additional treatment. If the AOP was utilized in the field to disinfect and reuse wash water
for vehicle washing, adequate contact time would be needed to ensure personnel are not
contaminated. Due to time constraints, additional testing was not possible, but the benefits
increased contact time could be examined in future research. Furthermore, the equipment used in
this study is scalable. Treated volumes were 100 to 150 gallons in this study, but the UV and
ozone units could be sized to handle more flow if larger volumes were generated during washing
activities.
The original objective of this study was to help decision makers in the Army determine if
vegetative biological agents are disinfected to a degree that wash rack or other wash water could
be reused for further vehicle washing or some other use. Since the wash rack water is
representative of water washed from cars or structures after an outdoor contamination event, this
data may be applicable a wide area biological contamination event, where dirty water must be
disinfected before being disposed of in a sewer. For both scenarios, the data in this study shows
that 4 to 6 log reduction of E. coli is possible if the flow rate used results in appropriate contact
time (4 gpm in this case). Therefore, if E. coli contamination is 4 log or less, the AOP
technology used in this study should be considered a tool that could be used for reuse and/or
disposal of wash rack or other dirty water.
25
-------
vvEPA
United States
Environmental Protection
Agency
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT NO. G-35
Office of Research and Development (8101R)
Washington, DC 20460
Official Business
Penalty for Private Use
$300
------- |