EPA 600/R/12/514 | May 2012 | www.epa.gov/ord
United States
Environmental Protection
Agency
Chemical Contaminant
Persistence and
Decontamination in Drinking
Water Pipes
Results using the EPA Standardized
Persistence and Decontamination
Experimental Design Protocol
Office of Research and Development
National Homeland Security Research Center
-------
EPA/600/R/12/514
May 2012
Chemical Contaminant Persistence
and Decontamination in Drinking
Water Pipes
Results using the EPA Standardized
Persistence and Decontamination
Experimental Design Protocol
-------
TABLE OF CONTENTS
Page
TABLE OF CONTENTS 3
LIST OF ABBREVIATIONS 6
EXECUTIVE SUMMARY 7
INTRODUCTION 10
1 SUMMARY OF EXPERIMENTAL DESIGN PROTOCOL 11
1.1 Experimental Reactor System 11
1.2 Coupon Biofilm Growth 12
1.3 Pipe Coupon Contamination Method Verification Experiments 13
1.3.1 Method Verification Step 1: Surface Contamination Extraction 13
1.3.2 Method Verification Step 2: Surface Contamination 14
1.4 Evaluation of Contaminant Persistence 15
1.5 Evaluation of Decontamination Approaches 16
1.6 Analytical Methods 18
1.6.1 Chlordane 18
1.6.2 Sodium Fluoroacetate 19
1.7 Quality Control 20
2 RESULTS REPORT 22
2.1 Results from Testing with Chlordane on Cement Pipe Coupons 22
2.1.1 Chlordane on Cement Quality Control Results 22
2.1.2 Method Verification Step 1: Chlordane on Cement Surface Extraction 24
2.1.3 Method Verification Step 2: Chlordane on Cement Surface Contamination 25
2.1.4 Chlordane on Cement Persistence Evaluation 26
2.1.5 Chlordane on Cement Flushing Evaluation 28
2.1.6 Chlordane on Cement Hyperchlorination Evaluation 29
2.2 Results from Testing with Chlordane on PVC Pipe Coupons 31
2.2.1 Chlordane on PVC Quality Control Results 31
2.2.2 Method Verification Step 1: Chlordane on PVC Surface Extraction 32
2.2.3 Method Verification Step 2: Chlordane on PVC Surface Contamination 33
2.2.4 Chlordane on PVC Persistence Evaluation 33
2.2.5 Chlordane on PVC Flushing Evaluation 35
2.2.6 Chlordane on PVC Hyperchlorination Evaluation 36
2.3 Results from Testing with Sodium Fluoroacetate on Cement Pipe Coupons 37
2.3.1 SFA on Cement Quality Control Results 37
2.3.2 Method Verification Step 1: SFA on Cement Surface Extraction 38
2.3.3 Method Verification Step 2: SFA on Cement Surface Contamination 39
2.3.4 SFA on Cement Persistence Evaluation 40
2.3.5 SFA on Cement Flushing Evaluation 41
2.3.6 SFA on Cement Hyperchlorination Evaluation 43
3 RESULTS SUMMARY 45
3.1 Experimental Design Protocol Development 45
3.2 Persistence and Decontamination Testing 46
3.2.1 Chlordane on Cement 46
3.2.2 Chlordane on PVC 46
-------
3.2.3 Sodium Fluoroacetate on Cement 46
3.3 Future Research Opportunities 47
REFERENCES 48
APPENDIX 49
Figures
Figure 1. Persistence evaluation - percent persistence and chlordane remaining on cement (left)
and backing (right) 26
Figure 2. Flushing evaluation - percent persistence and chlordane remaining on cement (left) and
backing (right) 28
Figure 3. Hyperchlorination evaluation - percent persistence and chlordane remaining on cement
(left) and backing (right) 30
Figure 4. Persistence evaluation - percent persistence and chlordane remaining on PVC 34
Figure 5. Flushing evaluation - percent persistence and chlordane remaining on PVC 35
Figure 6. Hyperchlorination evaluation - percent persistence and chlordane remaining on PVC 36
Figure 7. Persistence evaluation - percent persistence and SFA remaining on cement (left) and
backing (right) 41
Figure 8. Flushing evaluation - percent persistence and SFA remaining on cement (left) and
backing (right) 42
Figure 9. Hyperchlorination evaluation - percent persistence and SFA remaining on cement (left)
and backing (right) 43
Tables
Table 1. Contaminant Analytical Techniques, Limit of Quantisation, and Stock Solution
Concentrations 14
Table 2. Persistence Evaluation 16
Table 3. Evaluation of Flushing as Decontamination Approach 17
Table 4. Evaluation of Hyperchlorination as Decontamination Approach 18
TableS. Information about the GC-MS 18
Table 6. Information about the Ion Chromatography Method 20
Table 7. Data Quality Objectives for Contaminant Reference Methods 21
Table 8. Chlordane on Cement GC-MS Continuing Calibration Verification Results 23
Table 9. Chlordane on Cement Laboratory Fortification Matrix Sample Results 24
Table 10. Chlordane on Cement Surface Contamination Extraction 25
Table 11. Chlordane on Cement Surface Contamination 26
Table 12. Chlordane on Cement - Probability Value Matrix for Persistence Evaluation 27
Table 13. Chlordane on Cement - Probability Value Matrix for Flushing Evaluation 29
Table 14. Chlordane on Cement - Probability Value Matrix for Hyperchlorination Evaluation 30
Table 15. Chlordane on PVC GC-MS Continuing Calibration Verification Results 31
Table 16. Chlordane on PVC Laboratory Fortification Matrix Sample Results 32
Table 17. Chlordane on PVC Surface Contamination Extraction 33
Table 18. Chlordane on PVC Surface Contamination 33
Table 19. Chlordane on PVC - Probability Value Matrix for Persistence Evaluation 35
Table 20. Chlordane on PVC -Probability Value Matrix for Flushing Evaluation 36
Table 21. Chlordane on PVC - Probability Value Matrix for Hyperchlorination Evaluation .... 37
Table 22. SFA on Cement Laboratory Fortification Matrix Sample Results 38
-------
Table 23. SFA on Cement Surface Contamination Extraction 39
Table 24. SFA on Cement Surface Contamination 40
Table 25. SFA on Cement - Probability Value Matrix for Persistence Evaluation 41
Table 26. SFA on Cement - Probability Value Matrix for Flushing Evaluation 43
Table 27. SFA on Cement - Probability Value Matrix for Flushing Evaluation 44
-------
LIST OF ABBREVIATIONS
AR annular reactor
ASTM ASTM International
AWWA American Water Works Association
cm centimeters
cfu colony forming units
EPA U.S. Environmental Protection Agency
°C degrees Celsius
F flushing
ft/s feet per second
h hour
HC hyperchlorination
HPC heterotrophic plate counts
1C ion chromatography
IS internal standard
in. inch
g gram
GC-MS gas chromatographic mass spectrometry
KOW octanol-water partitioning coefficient
KOH potassium hydroxide
LOQ limit of quantitation
LFM laboratory fortified matrix
L liter
Lpm liter per minute
jiL microliter
jig microgram
mA milliamp
mM millimolar
mg milligrams
mm millimeters
mL milliliters
min minute
ng nanogram
NHSRC National Homeland Security Research Center
QAPP Quality Assurance Proj ect Plan
QC quality control
%R percent recovery
%P percent persistence
PE persistence evaluation
PDEDP Persistence and Decontamination Experimental Design Protocol
PVC polyvinyl chloride
rpm revolutions per minute
s second
SEVI selected ion monitoring
SPME solid phase micro extraction
-------
EXECUTIVE SUMMARY
The objective of this project was to develop
and test a standardized Persistence and
Decontamination Experimental Design
Protocol (PDEDP) that could be used across
laboratories to perform pipe
decontamination research. To test the
protocol for chemical contaminants, data
were collected pertaining to the adsorption,
persistence, and possible decontamination
approaches for chlordane and sodium
fluoroacetate (SFA) on cement-lined and
polyvinyl chloride (PVC) pipe material.
Experimental Design Protocol.
Implementation of the PDEDP simulated
conditions within operational drinking water
pipes using annular reactors (AR). The ARs
consist of a glass outer cylinder and a
rotating polycarbonate inner cylinder with
20 flush mounted rectangular coupons that
are made of materials that simulate drinking
water pipe materials. The annular reactor
was selected because it is relatively
inexpensive, permits the protocol to be
easily reproduced across different
laboratories, and eliminates potential
variability among various studies.
For this work, cement-lined and PVC
coupons were used. Shear stress was
applied to the coupon surfaces by setting the
AR inner cylinder rotation to 100
revolutions per minute (rpm), which
produces flow similar to 1 foot per second
(ft/s) (30.5 centimeters (cm)/s) in a 6 inch
(15.2 cm) diameter pipe . For the flushing
evaluation, the AR inner cylinder rotation
was set to 200 rpm (1.64 ft/s) (50.3 cm/s)
and subsequently 250 rpm (1.91 ft/s) (58.2
cm/s) to simulate increased flow . During
Based on calculations provided in the User Manual of the
BioSurface Technologies (421 Griffin Drive #2, Bozeman, MT
58715) Model 1120/1320 LS Biofilm Annular Reactor. Assumes a
Hazen-Williams coefficient of 120. Corresponding Reynolds
normal operation, the flow of drinking water
through the AR (connected directly to the
tap) was maintained at a mean velocity of
200 milliliters (mL) per minute so that the
mean residence time of the water in the AR
was 5 minutes. Prior to use of any pipe
material coupons, a biofilm was grown on
all of the coupons.
The PDEDP includes five components:
• Surface extraction method
verification - determines if a
contaminant could be extracted
from a pipe material surface
• Surface contamination method
verification - determines if the
pipe material coupon would be
contaminated when exposed to
bulk solution of contaminated
water
• Persistence evaluation - pipe
material coupons contaminated
and then exposed to fresh tap
water in ARs operating at 100
rpm (1 ft/s)
• Flushing evaluation - pipe
material coupons contaminated
and then exposed to fresh tap
water in ARs operating at 200
rpm (1.64 ft/s) or 250 rpm (1.91
ft/s)
• Hyperchlorination evaluation -
pipe material coupons
contaminated and then exposed
to solutions of 25 mg/L and 50
mg/L of free chlorine in ARs
with no rotation
Chlordane on Cement Results. The surface
extraction method confirmed that chlordane
can be extracted from the cement after direct
Numbers calculated for velocities of 1.0, 1.64, and 1.91 fVs are
53800, 88771, and 102759 respectively (all Re are turbulent flow).
-------
contamination of the coupon. The surface
contamination method verification
confirmed that a cement coupon can be
contaminated with chlordane by exposing it
to a solution of contaminated water. The
results from the persistence and flushing
evaluations exhibited results that were very
similar to one another. The percent
persistence (%P) after 24 h for the
persistence evaluation (AR operated at 100
rpm) was 9% ± 3% and the %P after 24 h
during the flushing evaluation (AR operated
at 200 rpm) was 6% ± 1%. Results from the
hyperchlorination evaluation showed that
hyperchlorination without increased flow is
not an effective means of decontaminating
chlordane from cement.
Chlordane on Polyvinyl Chloride Results.
The surface extraction method verification
confirmed that chlordane can be extracted
from the PVC surface after direct
contamination of the PVC coupon. The
surface contamination method verification
confirmed that a PVC coupon can be
contaminated with chlordane by exposing it
to a solution of contaminated water. The
results from the persistence and flushing
evaluations for the PVC exhibited very
similar results. The %P after 24 h for the
persistence evaluation (AR operated at 100
rpm (1 ft/s)) was 14% ± 4% and the %P
after 24 h during the flushing evaluation
(AR operated at 200 rpm (1.64 ft/s)) was
14% ± 6%. Again, as for the chlordane on
cement testing, results from the
hyperchlorination evaluation showed that
hyperchlorination without flow is not an
effective means of decontaminating
chlordane from PVC.
Sodium Fluoroacetate on Cement Results.
The surface extraction method confirmed
that SFA can be extracted from the cement
after direct contamination of the coupon.
The surface contamination method
verification confirmed that a cement coupon
can be contaminated with SFA by exposing
it to a solution of contaminated water. The
results from the persistence, evaluation, and
hyperchlorination evaluations showed that
SFA was persistent in each of these
experimental scenarios.
Future Research Needs. This work has laid
the foundation for a PDEDP that can be
adapted to accommodate additional research
priorities. Below are a few possible areas
for further study:
• Importance of biofilm to pipe
decontamination research - During
the SFA surface contamination
method verification step, two
cement coupons without biofilm
(only two because of the limited
capacity of the AR and the fact that
this impromptu experiment was
outside the context of the PDEDP)
were contaminated with SFA along
with the coupons covered with
biofilm. For these two coupons, five
times as much SFA was adsorbed as
the coupons with biofilm. This very
limited data set suggested that the
presence or absence of biofilm could
significantly impact the results of
pipe adsorption/decontamination
research. More rigorous
experimentation would need to be
performed to better characterize the
role of biofilm, which is typically
expected in actual field studies.
• Broadening of
adsorption/decontamination data set
by expanding on list of chemical
contaminants tested using the
PDEDP (e.g., organophosphates as
available toxic chemicals and
simulated chemical agents, metals to
simulate heavy metal, or
radiological contamination).
• Study of
adsorption/decontamination of
-------
biological organisms using the
PDEDP.
Use of additional pipe materials with
additional chemicals and biological
organisms as well as additional
chemical pipe cleaning materials as
possible decontamination agents.
Research on shearing stress,
dynamic pressure, and the effects
that laminar, transient, and turbulent
flow has on biofilm removal as part
of the PDEDP on different diameter
pipes.
Scaling up of AR experiments into
experiments with real pipe using a
pipe loop in order to study how well
the AR experiments translate into
scenarios with real pipe.
Study of risk assessment questions
addressing how much persistence of
various chemicals is acceptable.
-------
INTRODUCTION
The U.S. Environmental Protection
Agency's (EPA) National Homeland
Security Research Center (NHSRC)
conducts research to protect, detect, respond
to, and recover from terrorist attacks on the
nation's water and wastewater
infrastructure. The objective of this project
was the development and testing of a
standardized Persistence and
Decontamination Experimental Design
Protocol (PDEDP) to quantitatively
determine the persistence of individual
priority contaminants to various drinking
water pipe materials as well as the testing of
techniques for decontaminating affected
pipe surfaces if the contaminant persists.
This report provides a summary of the
results from the testing that was performed
following the development of the
experimental design protocol, which is
included in Appendix A. As thoroughly
described in the PDEDP, testing included
use of an annular reactor (AR) as the device
used to simulate flow past materials from
which drinking water pipe is made. The
annular reactor was selected because it is
relatively inexpensive, permits the protocol
to be easily reproduced across different
laboratories, and eliminates potential
variability associated with laboratories
constructing their own apparatus, which
would likely occur even with detailed
instructions. Annular reactors have been
used for several previous EPA persistence
and decontamination studies4'5.
The drinking water pipe materials used for
the study included cement-lined (with
contaminants chlordane and sodium
fluoroacetate [SFA]) and polyvinyl chloride
(PVC) with only chlordane. These two
contaminants were selected based in part on
their absorption properties; chlordane is a
low solubility organic while SFA is ionic.
Specifically, the absorption characteristics
of a chemical can be described by its
octanol-water partitioning coefficient (Kow).
Chemicals with high Kow values are more
likely to partition out of the water and onto
the pipe surface and chemicals with low Kow
values are more likely to remain in the water
than absorb onto the pipe. Of the two
contaminants, chlordane is the high Kow
value contaminant (log Kow of 6.2) and
sodium fluoroacetate is a chlorine resistant
contaminant with ion-exchange (log Kow of -
0.061) sorption characteristics. The
following report includes a summary of the
experimental design as well as study results,
with one section dedicated for each pipe
material and contaminant combination that
was tested.
10
-------
1.
SUMMARY OF EXPERIMENTAL DESIGN PROTOCOL
This project included five components of
testing that were completed for each
combination of pipe material and
contaminant. They included 1) the surface
extraction method verification, 2) the
surface contamination method verification,
3) the persistence evaluation, 4) the flushing
evaluation, and 5) the hyperchlorination
evaluation. Summaries of the experimental
set up, each component of the experimental
design, and details of the analytical methods
are provided below.
1.1. Experimental Reactor System
For the persistence and decontamination
experiments described in this experimental
design, the conditions within operational
drinking water pipes were simulated in
annular reactors (AR) (Bio Surf ace
Technologies Corporation, Bozeman, MT).
The ARs consist of a glass outer cylinder
and a rotating polycarbonate inner cylinder
with 20 flush mounted rectangular coupons
that are made of materials that simulate
drinking water pipe materials. For this
testing, cement-lined and PVC coupons
(BioSurface Technologies Corporation,
Bozeman, MT) were used. For the cement-
lined coupons, the cement used for the
coupons met the requirements of the C150-
07 American Society for Testing and
Materials (ASTM) Standard Specification
for Portland Cement and the thickness of
the cement was approximately 1.3 mm,
slightly less than as specified in American
Waterworks Association (AWWA) C104-
03 Standard for Cement-Mortar Lining for
Ductile-Iron Pipe and Fittings for Water3.
The cement coupons were made from a
polycarbonate backing with the cement
applied at the above thickness. Because of
the porosity of the cement, some of the
contaminants passed through the cement and
adsorbed to the polycarbonate backing.
Therefore, the cement was separated from
the polycarbonate and the two components
were analyzed separately. The PVC
coupons were made entirely of PVC so no
separation was required. In this manner, the
adsorption to the infrastructure material
could be investigated independent of other
adsorption processes occurring in the AR
set-up.
The coupons had surfaces that were 0.55
inch (in.) (14 millimeters (mm)) x 5.8 in.
(148 mm). Shear stress was applied to the
coupon surfaces by setting the inner AR
cylinder rotation to 100 revolutions per
minute (rpm), which produces shear similar
to 1 foot (ft)/second (s) (30.5 centimeter
(cm)/s) flow in a 6 inch (in.) (15.2 cm)
pipe4. For the flushing evaluation, the AR
inner cylinder rotation was set to 200 rpm
(1.64 ft/s) (50.3 cm/s) and subsequently 250
rpm (1.91 ft/s) (58.2 cm/s) to simulate
increased flow. During normal operation,
the flow of drinking water through the AR
(connected directly to the tap) was
maintained at a mean velocity of 200
milliliters (mL) per minute, so the mean
residence time of the water in the AR was
five minutes. This flow velocity prevented
the depletion of chlorine level over the
course of the experiments. The short
residence time decreased the chance that
desorbing contaminant could re-contaminate
a surface.
Columbus, Ohio tap water from the
laboratory faucet was used for the study and
no range of water quality parameters was
specified. However, experience in the same
laboratory has shown that the free chlorine
level is typically between 1.0 mg/L and 2.0
11
-------
mg/L, the pH between 7.5 and 8.0, and the
temperature between 22 and 25 degrees
Celsius (°C). The pH, temperature, and free
chlorine concentration of the drinking water
was measured daily using a multi-parameter
water monitor (Rosemount Analytical
Model WQS, Rosemount Analytical, Irvine,
CA). The ARs were always operated in the
dark by covering them completely with
aluminum foil. Because some contaminant
was likely to adsorb onto the non-coupon
components of the AR and affect the amount
of contaminant that was available for
coupon contamination, the concentration of
the bulk contamination solutions was
measured to ensure that an adequate
concentration of contaminant was
maintained to achieve coupon
contamination.
1.2 Coupon Biofilm Growth
Prior to performing each component of the
PDEDP, a biofilm was grown on all of the
coupons by submerging the required number
of coupons into a container (an 8L plastic
tub) that allowed recirculation of
dechlorinated tap water (outlet near the top
of the container and inlet near the bottom of
the container) fortified with 1 gram (g) of
yeast extract as a nutrient to stimulate more
rapid biofilm growth. This container was
filled with water and kept in the dark (to
better simulate biofilm growth in an
enclosed pipe) and recirculated using a
pump for at least four days with an
additional 1 g of yeast added after every two
days. The biofilm growth was measured,
using heterotrophic plate counts (HPC),
from one of the coupons in the biofilm
growth container. However, there was not a
strict biofilm density required for use in
experiments. Following the detailed
procedure included in the PDEDP, coupons
to be measured for HPC were centrifuged in
a Triton X solution, mixed using a vortex
mixer, and then decanted. Two tenfold
dilutions of that decanted solution were
prepared and plated in triplicate on tryptic
soy agar plates (Rainin L200, LI9304,
Rainin Instrument LLC, Oakland, CA).
After incubation for 48 hours at 35-37 °C,
the distinguishable colonies on each plate
were counted and surface density of HPC
was calculated by dividing the number of
colonies by the surface area of the coupons.
Throughout the cement and PVC
experiments, seven sets of coupons were
used and the HPC densities were determined
for six of the seven sets. On average, the
HPC densities were 1.6 x 106 colony
forming units (cfu)/cm2. The standard
deviation of the HPC densities was 1.3 x io6
cfu/cm2. The HPC concentration in the
biofilm growth water was determined for all
seven sets of coupons. The average HPC
concentration was 6.3 x IO5 cfu/mL with a
standard deviation of 1.1 x io6 cfu/mL.
While there was not a target HPC density to
be grown on the pipe material coupons, the
consistent growth of biofilm (densities
within one log of one another) provided a
means to simulate pipe conditions on pipe
material coupons.
The one set of coupons for which no HPC
measurement was made was used for the
cement-chlordane persistence evaluation.
The HPC measurement was not made
because the colonies on the enumeration
plates of the dilution level used were too few
to count. A more concentrated dilution (that
had been refrigerated for two days) was then
plated and incubated, but there were again
too few colonies to count. The
concentration of HPC in the water used for
biofilm growth on that set of coupons was
9.0 x IO4 cfu/mL, which was similar to the
water HPC concentrations measured in the
biofilm growth water for the rest of the
coupons. Because none of the other biofilm
12
-------
growth conditions had been altered and there
were similar levels of HPC in the biofilm
growth water, it was determined that colony
growth from the more concentrated dilution
was apparently inhibited by storage during
the incubation of the original plate and it
was likely that there had been biofilm on
that set of coupons.
1.3 Pipe Coupon Contamination
Method Verification Experiments
The generation of persistence and
decontamination data from this experimental
design included contamination of coupons
by exposing them to bulk solutions of
chlordane and SFA. Thereafter, the
persistence of each contaminant on the
coupons and/or the application of a
decontamination approach were investigated
to determine both the propensity of each
contaminant to persist on the coupons and
the effectiveness of decontamination
approaches in removing the applicable
contaminant from the coupon surface. The
usefulness of results from such experiments
relies on the accuracy of the required
contaminant measurements. In order to be
confident in these measurements, two
important questions needed to be answered
about the approach to contaminant
measurement.
• When adsorbed to the coupon
surface, how well can a
contaminant be extracted
from that surface?
• When a coupon has been
exposed to a bulk solution at
a given concentration, how
much of the contaminant is
adsorbed to the coupon
surface?
To answer these two questions, two method
verification steps were conducted as the first
two steps of the experimental design. First,
the surface contamination extraction method
was validated. Second, the coupon surface
contamination method was validated.
1.3.1 Method Verification Step 1: Surface
Contamination Extraction
The purpose of this step is to determine
whether it is possible to extract the
contaminant if adsorbed to a pipe material
surface. The extraction must be statistically
quantifiable in order to make valid
conclusions about contaminant removal;
otherwise the extraction procedure must be
further developed. The verification required
20 half coupons of the applicable material
type with a biofilm developed as described
in Section 1.1. These coupons were
removed from the biofilm growth container
and allowed to air dry until water droplets
were not visible on the surface, but the
surface was still damp (mean time of seven
minutes). This drying step ensured that the
contaminant was added to the coupon
surface and not to the water remaining on
the coupon surface following the time period
that the coupon was immersed in water
during biofilm growth.
Each coupon, including blanks, was cut in
half with scissors and five drops of
contaminant solution were applied directly
to each half coupon using a micropipette
(Eppendorf Research Plus, Eppendorf
International, Hauppauge, NY)
approximately 10 mm apart. For chlordane,
the volume of each drop was 5 jiL and for
sodium fluoroacetate, the drop volume was
15 jiL. This verification included low,
medium, and high spike levels to determine
the effectiveness of the extraction at various
contamination levels. Table 1 gives the
concentration of the three chlordane and
SFA spiking solutions.
13
-------
Table 1. Contaminant Analytical Techniques, Limit of Quantitation, and Stock Solution
Concentrations
Contaminant
Chlordane
Sodium Fluoroacetate
Analytical Technique
Gas Chromatographic
Mass Spectrometer
Ion Chromatography
Approx. Limit
of Quantitation
0.002 mg/L
O.lmg/L
Concentration of
Spike Solutions
0.8, 4, and 40 mg/L
133, 667, 6,667 mg/L
Each coupon received drops of a different
contaminant concentration and each
concentration was applied to five coupons
(for a total of 15 coupons per contaminant).
The drops were allowed to air dry until they
were not visible on the surface (mean of
seven minutes) to ensure that the
contaminant was being extracted from the
surface of the coupon (and not from a
droplet of spiking solution). Five non-
contaminated coupons were also extracted
as blanks.
The surface contamination extraction
method included the extraction of the entire
coupon, both the cement surface and the
polycarbonate backing supporting the
cement. The cement coupons were
extracted (for separate analysis) by
removing the cement from the
polycarbonate backing and placing the
cement and polycarbonate backing into
separate test tubes (Kimble #73785-50,
VWR, West Chester, PA or Fisherbrand
#03-337-14, Fisherbrand, Pittsburgh, PA)
filled with an appropriate extraction solvent.
The extraction solvent for chlordane was 9:1
hexane:acetone and for SFA, ASTM
International (ASTM) Type I water. For the
chlordane extractions, after inserting both
components of the coupons into separate test
tubes, the test tubes were sealed with a cap
and sonicated for 10 minutes, solvent
decanted and replaced with fresh solvent,
and then sonicated for another 10 minutes.
The decanted solvent was combined. The
resulting solution was centrifuged and
supernatant solution collected for analysis.
The SFA coupons were extracted in a
similar manner but only one sonication step
was performed. The PVC coupons required
no separation and were extracted following
the same method as for the polycarbonate
backing of the cement coupons. For
chlordane, the extraction solution was
concentrated using nitrogen evaporation
prior to analysis using a gas chromatograph-
mass spectrometer (GC-MS). For SFA, ion
chromatography (1C) was used as the
measurement technique without sample
concentration.
The percent recovery (%R) was calculated
using the following equation
%R = -£ x 100
^o
where CR is the mass of contaminant
recovered from the coupon surface (area
22.5 cm2) and C0is the mass of contaminant
originally dispensed onto the coupon
surface.
7.3.2 Method Verification Step 2: Surface
Contamination
Step 2 involved validating a method to
contaminate the surface of the coupons in a
way that simulates an actual intentional
contamination of a water distribution
system. The surface contamination method
to be validated incorporated:
• Preparing coupons with biofilm
• Exposing the coupons to
contaminated water (1 mg/ liter [L] -
14
-------
chlordane and 500 mg/L - SFA) in
the AR without flow (batch mode)
• Extraction of the contaminant from
the coupon using the method
validated in Step 1.
To begin the verification, 10 coupons were
prepared with a biofilm. Then, the AR was
filled with contaminated water at the above
concentration levels and five of the coupons
were added to the AR and five were
collected as blank samples. Then, the AR
was operated at 100 rpm (1 ft/s), but the
flow of tap water through the AR was
stopped to increase the contact time between
the contaminated water and the coupons.
Two hours following the contamination of
the water, the coupons were removed, rinsed
twice with 25 mL of ASTM Type I water
(which was then discarded), and then
extracted and analyzed following the surface
contamination extraction and measurement
method described in Section 1.3.1. This
rinse step was to ensure that the contaminant
is extracted from the surface of the coupon
and not just an artifact of residual
contamination solution on the surface of the
coupon. The bulk solution was sampled at
the start of the contamination time period, at
the half-way point, and at the end and the
concentration of contaminant was measured
via the applicable measurement technique to
confirm the availability of the contaminant
for adsorption.
1.4 Evaluation of Contaminant
Persistence
This section describes the approach to
evaluating the persistence of a contaminant
on various pipe coupon materials. Table 2
provides an overview of the persistence
evaluation (PE). For each combination of
coupon material and contaminant, biofilm
was grown on 20 coupons as described in
Section 1.2. Two coupons with biofilm
were the non-contaminated blank coupons
and the rest of the coupons were
contaminated with a bulk solution following
the surface contamination method.
Immediately following the coupon
contamination step, three coupons were
removed to serve as control coupons. The
amount of contaminant on the surface of
these control coupons were compared with
the amount remaining on the coupons that
were left in the AR for various lengths of
time following the removal of the control
coupons.
Thereafter, a stopped flow scenario was
evaluated by stopping the rotation of the AR
and stopping the flow of water through the
AR (after the contaminant water is replaced
by uncontaminated drinking water). This
stopped flow scenario was held for 24 hours
after which three PE coupons were removed.
After that 24 hour period, the flow of
drinking water and AR rotation was resumed
to normal operating conditions (AR rotating
at 100 rpm (1 ft/s) and tap water flow
through the AR at 200 mL/min, with a mean
hydraulic retention time of 5 minutes.
15
-------
Table 2. Persistence Evaluation
PE Step
PE 1
PE2
PE3
PE4
PE5
PE6
PE7
Description
Developed biofilm (confirmed with heterotrophic plate count) on 20
coupons; remove two coupons as blank control coupons
Stopped flow through AR, filled AR with contaminated bulk solution
concentration, inserted 18 coupons into AR, operated AR at 100 rpm,
waited 2 hours
Sampled bulk contamination solution at start, half-way point, and end of
contamination period
Following 2 hour contamination period, removed three coupons as
contaminated control coupons
Stopped AR rotation to simulate stopped flow. Replaced bulk
contamination solution with uncontaminated water and remained at
stopped flow for 24 hours; collected three coupons
Restarted the AR rotation and flow through the AR. Removed three
coupons at 4 hours, 1 day, 3 days, and 7 days after restart of AR rotation
and flow
Measured amount of contaminant remaining on coupons and compared to
amount remaining on contaminated control coupons
Coupons
removed
(20 total)
2
0
0
3
3
12
0
Following the stopped flow scenario, sets of
three PE coupons were collected from the
AR at four different time increments (4
hours, 1 day, 3 days, and 7 days) following
the resumption of flow. Following the
removal of each of these sets of PE coupons,
they were extracted and the amount of
contaminant on the coupon surfaces
compared with the amount on the control
coupons collected just after the coupon
contamination step.
This comparison was made by calculating
the percent persistence (%P) of the
contaminant on the coupons as described by
the following equation.
£
%p = -™ x 100
Q
where CPE is the mass of contaminant
recovered from the coupon surface and Cc is
the average mass of contaminant originally
measured from the surfaces of the control
coupon surfaces.
1.5 Evaluation of Decontamination
Approaches
This section describes the evaluation of two
approaches to decontaminating pipe,
flushing (F) and hyperchlorination (HC).
Table 3 provides an overview of the flushing
evaluation and Table 4 provides an overview
of the HC evaluation. As was the case for
the persistence evaluation, a biofilm was
grown on 20 coupons of the desired material
and 18 were loaded in the AR and
contaminated using the validated surface
contamination method. Then three
contaminated coupons were removed to
serve as the control coupons. The amount of
contaminant on the surface of these control
coupons were compared with the amount
remaining on the coupons that were left in
the AR (operated under increased flow
conditions to simulate flushing).
For the flushing evaluation, following
coupon contamination, the AR inner
cylinder rotation was raised from 100 rpm (1
ft/s) to 200 rpm (1.64 ft/s), which
16
-------
-1
corresponded to a water velocity of 0.5 ms
in a 15.2 cm (6 in.) pipe3. This increased
rotational speed was held for one day. Sets
of three coupons were collected from the
AR at three different time increments (1
hour, 4 hours, and 1 day) following the
coupon contamination. Then, the rotational
speed was increased again to 250 rpm (1.91
ft/s) and held for another day, with the
collection of three coupons after 4 hours and
after 1 day of 250 rpm (1.91 ft/s) conditions.
Following the removal of each set of three
coupons, the coupons were extracted and the
amount of contaminant on the coupon was
compared with the amount on the control
coupons collected just after the surface
contamination step. Comparisons were
made using a recognized statistical
approach, as illustrated in the study results.
The evaluation of hyperchlorination as a
decontamination approach was performed as
shown in Table 4. The evaluation was
started in a similar way as for the flushing
evaluation. However, instead of increasing
the rotational velocity of the AR, the
rotation of the AR was stopped and the
drinking water flow through the AR was
also stopped to simulate a stopped flow
scenario. The free chlorine concentration
was then increased first to 25 mg/L and then
to 50 mg/L after several increments of time
after which coupons were collected from the
AR. This comparison was made by
calculating the %P as described in the
previous section.
Table 3. Evaluation of Flushing as Decontamination Approach
Step
F 1
F2
F3
F4
F5
F6
F7
F8
F9
Description
Developed biofilm (confirm with heterotrophic plate count) on 20 coupons
of same material; removed two coupons as blanks
Injected enough contaminant into ARto make desired bulk concentration
within AR; inserted 18 coupons and operated AR at 100 rpm, waited 2
hours
Sampled bulk contaminant solution at start, half-way point, and end of
contamination time and sample bulk contamination solution
Following 2 hour contamination period, replaced bulk contamination
solution with uncontaminated water and removed three coupons as
contaminated control coupons
Increased AR rotational velocity to 200 rpm (1 .64 ft/s) from original
velocity of 100 rpm (1 ft/s)
Removed three coupons at 2 hours, 4 hours, and 1 day following increase in
rotational velocity
Increased AR rotational velocity to 250 rpm (1.91 ft/s) from 200 rpm
Removed three coupons at 4 hours and 1 day following increase in
rotational velocity to 250 rpm
Measured amount of contaminant remaining on coupons and compared to
amount remaining on contaminated control coupons
Coupons
removed
(20 total)
2
0
0
3
0
9
0
6
0
17
-------
Table 4. Evaluation of Hyperchlorination as Decontamination Approach
Step
HC 1
HC2
HC3
HC4
HC5
HC6
HC7
HC8
HC9
Description
Developed biofilm (confirm with heterotrophic plate count) on 20 coupons
of same material; removed two coupons as blanks
Injected enough contaminant into ARto make desired bulk concentration
within AR; inserted 18 coupons and operate AR at 100 rpm, waited 2 hours
Sampled bulk contaminant solution at start, half-way point, and end of
contamination time and sampled bulk contamination solution
Following the 2 hour contamination period, replaced bulk contamination
solution with uncontaminated water and removed three coupons as
contaminated control coupons
Stopped flow through AR and stopped rotation of AR; increased free
chlorine concentration to 25 mg/L
Removed three coupons at 2 hours, 4 hours, and 1 day following increase in
free chlorine concentration
Increased free chlorine concentration to 50 mg/L
Removed three coupons at 4 hours and 1 day following increase in free
chlorine concentration to 50 mg/L
Calculated percent persistence for all coupons by comparing residual
contaminant on the surface with contaminated control coupons
Coupons
removed
(20 total)
2
0
0
3
0
9
0
6
0
1.6 Analytical Methods
1.6.1 Chlordane
The analytical standard for chlordane (Chem
Service, West Chester, PA) was a mixture of
the isomers alpha-chlordane (30%), beta-
chlordane (37%), gamma-chlordane (10%),
and trans nonachlor (23%). The relative
abundances were determined through
evaluation of peak areas during repeated
Table 5. Information about the GC-MS
analysis of a 100 nanogram (ng)/mL
calibration standard. The standard solutions
for chlordane were made in hexane.
Trichlornate (ChemService, West Chester,
PA) was used as the internal standard (IS).
The samples were analyzed by GC-MS
(Agilent 5973, Agilent, Santa Clara, CA)
operating in the selected ion monitoring
(SIM) mode. Table 5 gives details
pertaining to the GC-MS:
Component
Analytical column
Helium flow rate
Injection volume
Injection port
Oven temperature
program
Transfer line
temperature
Quantitation Ions
Description
Rtx-5MS (Restek, Bellefonte, PA), 30m x 0.25 mm
(\m\) film or equivalent
x 0.25 micrometer
1 mL/min
1-2 jiL
300°C, splitless for 0.75 min
120°C for 1 min, 120-300°C at 9°/min, hold 300 °C
for hold for 10 min
300°C
Chlordane 373/375/377
Trichlornate 297/299/269
18
-------
Calibration standards were prepared at total
chlordane concentrations from 2-1000
ng/mL. Each calibration standard contained
the IS at a constant level. The calibration
curve was analyzed followed by a blank and
then the samples. The limit of quantitation
(LOQ) for this method was 2 ng/mL. If the
concentration of a sample exceeded the
highest calibration point, that sample was
diluted into the calibration range and re-
analyzed.
Two continuing calibration check solutions
(lowest and middle calibration levels,
respectively) were analyzed after every 10
samples and at the end of the sequence in
order to verify instrument sensitivity and
calibration throughout the analysis. The
results of these samples were targeted to be
between 70 -130% of the known
concentration. A laboratory reagent blank
consisting of hexane was analyzed at the
beginning of the sequence and bracketed all
calibration and check standards in order to
verify system cleanliness and prevent
carryover. In addition, 200 ng/mL
chlordane was added to a split sample of
10% of the total samples analyzed to create
laboratory fortified matrix (LFM) samples.
Target recoveries for the LFM samples were
from 70-130%.
The coupon extracts for chlordane were
concentrated to 1 mL and transferred to a
GC-MS analysis vial for direct analysis.
Sample concentration was performed using
a TurboVap LV (Biotage, Charlotte, NC).
In summary, the sample was transferred to
the TurboVap LV tubes by rinsing the
original extraction test tube. The nitrogen
was turned on to 4 pounds per square inch
and then the solution was checked
periodically to determine remaining volume,
taking care to avoid concentrating the
sample below the target volume. Each
sample was removed from the concentrator
as the sample reached a final 1 mL volume.
The bulk contamination solution samples
were analyzed using solid phase micro
extraction (SPME, Supelco 57341-U, 3-
pack) to extract the contaminants out of the
aqueous solution. A 3 mL volume of the
bulk contamination solution was extracted
by placing the samples in SPME vials
(ChromSys, 18 03 1309-lOmL) and
analyzed directly by GC-MS. A relative
determination of peak areas was used to
evaluate if there was chlordane available for
binding throughout the time period of
contamination.
1.6.2 Sodium Fluoroacetate
SFA calibration standards were prepared in
ASTM Type I water from a high purity
(>99%) standard from Riedel-de Haen
PESTANAL® Analytical Standard Catalog
#36755 (Sigma-Aldrich, St. Louis, MO).
The 1C system consisted of a Dionex LC 20
with EG40 Eluent Generator, AS3500
Autosampler, GP40 Gradient Pump, and
Dionex lonpac ASH analytical column (4
x 250 mm) (Dionex, Bannockburn, IL).
Table 6 gives a few details pertaining to the
1C method. QC criteria are described
subsequently.
19
-------
Table 6. Information about the Ion Chromatograph
Component
Detector
Mobile phase
Elution
Description
ED40 Electrochemical Detector working in conductivity mode with 5
milliamp (mA) suppression current
0.5 millimolar (mM) potassium hydroxide (KOH) in ASTM Type I
water at a 2.00 mL/minute (min) flow rate
Gradient starting at 0.5 mM KOH for 1.5 min followed by linear ramp
from 0.5 mM to 10.5 mM KOH over next five minutes (2 mM/min).
Next three minutes consist of cleanout step where mobile phase
increases to 40 mM KOH. System re-equilibration then achieved by
decreasing eluent concentration to 0.5 mM KOH for 14.5 min resulting
in 24 minute run time
Quantitative analysis was performed using
external standards. A five-point calibration
curve was generated at the beginning of the
sequence. The calibration levels ranged
from 0.1 mg/ liter (L) to 2.5 mg/L. The
LOQ for this method was 0.1 mg/L. If the
concentration of a sample exceeded the
highest calibration point, that sample was
diluted into the calibration range and re-
analyzed.
One continuing calibration check solution
(0.5 mg/L) was analyzed after every 10
samples and at the end of the sequence in
order to verify instrument sensitivity and
calibration throughout the analysis. The
acceptable recovery these samples was for
their concentration to be between 90-110%
of the known concentration. A laboratory
reagent blank consisting of ASTM Type I
water was analyzed at the beginning of each
sequence to verify system cleanliness. In
addition, 0.5 mg/L of sodium fluoroacetate
was added to a split sample of 10% of the
total samples analyzed to create LFM
samples. Acceptable recoveries for the
LFM samples ranged from 75-125%. The
calibration standards, water samples, and
sample extracts were directly injected onto
the 1C at a volume of 100 jiL.
1.7 Quality Control
Quality control samples for the contaminant
reference methods including continuing
calibration checks, laboratory blanks, and
laboratory fortified matrix samples are
described in Section 2. The data quality
objectives for each of these samples are
provided in Table 7. The acceptable ranges
were intended to limit the error introduced
into the experimental work.
20
-------
Table 7. Data Quality Objectives for Contaminant Reference Methods
Method
GC-MS analysis
of chlordane
1C analysis of
sodium
fluoroacetate
(similar to EPA
Method 300.0)
Sample Type
Continuing calibration
check at lowest and
middle calibration
levels
Laboratory reagent
blank
Laboratory fortified
matrix samples
Continuing calibration
check at middle
calibration level
Laboratory reagent
blank
Laboratory fortified
matrix samples
QC Target
70- 130% of known
concentration, include with
each batch of 10 samples
-------
2.
RESULTS REPORT
Testing of the PDEDP included use of
chlordane and SFA with cement-lined AR
coupons as well as chlordane with PVC AR
coupons. The results are divided into
separate sections for each combination of
contaminant and coupon type.
2.1 Results from Testing with
Chlordane on Cement Pipe Coupons
The following sections describe results from
performing quality control, verification, and
evaluation experimental design procedures
for chlordane on cement pipe coupons.
2.1.1 Chlordane on Cement Quality
Control Results
Continuing calibration verification (CCV)
samples were analyzed on the GC-MS
throughout each analysis set. After every 10
samples analyzed, a low concentration
calibration solution (2 ng/mL or 5 ng/mL)
and a middle concentration calibration
solution (100 ng/mL) were reanalyzed. In
addition, 10% of the samples were split and
200 ng of chlordane was spiked into the
sample extract to create a laboratory
fortified matrix (LFM) samples. Target
recoveries for each of these QC samples
were between 70% and 130%. Tables 8 and
9 show the results obtained during testing.
For the CCV samples, the recoveries of the
low concentration samples ranged from 84%
to 246%. These low concentration samples
were very close to the LOQ so small
changes in peak area greatly impacted the
percent recoveries of the CCV samples.
Specifically all four Step 1 surface
extraction method verification low
concentration CCV samples (recovery
range: 185%-198%) and two out of the four
low concentration CCV samples exceeded
the acceptable range of recoveries during the
analyses applicable to both the flushing
(210% and 246%) and hyperchlorination
(132% and 149%) evaluations. However, no
corrective action was taken with these
results (i.e. results were used) because the
peak areas measured during these
components of the evaluation were closer to
the middle and higher parts of the
calibration curve. The recoveries of the
middle concentration (100 mg/mL) CCV
samples ranged from 64% to 118% with an
average recovery of 79% with a standard
deviation of 16%. The middle concentration
CCV was never more than 6% outside the
targeted acceptable range and within each
sample set there was at least one CCV
sample that was within the targeted range.
For the LFM samples, the recoveries ranged
from 86% to 209%. All but two of the LFM
samples that were outside of the targeted
range of recovered occurred during the Step
1 and Step 2 method verification
experiments, which were used to
qualitatively determine the feasibility of
extracting chlordane from the surface of the
coupon as well as contaminating the surface
from a bulk solution. One LFM each from
the persistence and hyperchlorination
evaluations were the only other LFM
samples to be outside of the target range.
No corrective action was made because the
LFM samples that were outside of the
targeted recovery range were paired with
several other LFM samples that were
recovered within the targeted range.
22
-------
Table 8. Chlordane on Cement GC-MS Continuing Calibration Verification Results
Component of Testing
Step 1 - Surface Extraction Method
Verification
Step 2 - Surface Contamination
Method Verification
Persistence Evaluation
Flushing Evaluation
Hyperchlorination Evaluation
Average
Standard Deviation
Low Calibration Standard
(%R of Expected)
198%
196%
185%
199%
186%
121%
94%
87%
85%
84%
210%
246%
121%
100%
103%
121%
132%
149%
145%
51%
Mid Calibration Standard
(%R of Expected)
70%
67%
68%
70%
66%
79%
75%
73%
70%
69%
97%
68%
64%
75%
95%
92%
118%
110%
79%
16%
23
-------
Table 9. Chlordane on Cement Laboratory Fortification Matrix Sample Results
Component of Testing
Step 1 - Surface Extraction
Method Verification
Step 2 - Surface Contamination
Method Verification
Persistence Evaluation
Flushing Evaluation
Hyperchlorination Evaluation
Average
Standard Deviation
Laboratory Fortified Matrix
(%R)
161%
142%
171%
131%
113%
201%
145%
150%
110%
129%
86%
119%
121%
106%
86%
128%
116%
209%
135%
34%
The bulk contamination solution was
sampled at the beginning, middle, and end
of the 2 h contamination time during the
Step 2 method verification, the persistence
evaluation, the flushing evaluation, and the
hyperchlorination evaluation. These
samples were analyzed by direct SPME
injection and a relative comparison of
chlordane isomer peak areas was used to
evaluate if there was chlordane available for
binding throughout the time period of
contamination. Across those four
experiments, the peak areas of the initial 1
milligram (mg)/L contamination solution
were considered the 100% chlordane levels.
The samples collected at the 1 h point of the
contamination step retained a 22%±2% of
the peak area and the sample collected at the
end of the contamination period retained
17%±2% of the peak area. Therefore, while
there was a considerable loss of chlordane
during the first hour, chlordane was
available throughout the entire 2 h
contamination period.
2.1.2 Method Verification Step 1:
Chlordane on Cement Surface Extraction
The objective of this component of testing
was to determine if chlordane could be
extracted from the surface of the coupon.
Cement coupons were spiked with 20 ng,
100 ng, and 1,000 ng of chlordane. When
the chlordane was spiked onto the cement
coupon, some of the chlordane adsorbed to
the cement surface and some flowed through
the cement and adsorbed to the
polycarbonate backing on which the cement
was mounted. The cement and backing
were extracted separately using the method
24
-------
described in Section 1.3.1 and the results
were reported for both the cement and the
backing for all five components of the
testing. Table 10 gives the results including
the amount of chlordane spiked onto the
coupons, the amount extracted from the
backing and cement, the total recovery, and
the standard deviation. Overall, the total
recovery ranged from 44% to 68% with
standard deviations of the total percent
recovered across the five replicates of less
than 5%. This indicates that chlordane
could be reproducibly extracted and
measured from both the cement surface and
the polycarbonate backing of the coupons.
The amounts recovered from the backing
and cement show that considerably more
chlordane adsorbed to the cement surface
than passing through the cement and
adsorbing to the polycarbonate backing.
The concentration of each spiking solution
was confirmed using GC-MS. The low,
middle, and high spiking solutions had
percent recoveries of 76%, 83%, and 87% of
the target concentration levels.
Table 10. Chlordane on Cement Surface Contamination Extraction
Low level
Mid level
High level
Amount
spiked
(ng)
20
100
1000
Avg. amount
recovered from
cement (ng)
8.8
34
340
Avg. amount
recovered from
backing (ng)
4.8
13
100
Avg. total
recovered (ng)
14
47
440
Total %
Recovery
68%
47%
44%
SD
2%
2%
3%
Five replicates were spiked and extracted at each concentration level.
2.1.3 Method Verification Step 2:
Chlordane on Cement Surface
Contamination
This verification indicates if a contaminant
will adsorb to the cement surface containing
biofilm in the event that it is exposed to a
bulk solution. Table 11 gives the results
from the surface contamination method
verification for chlordane on cement
including the amount of chlordane extracted
from each part of the coupon after a two
hour exposure to 1 mg/L chlordane.
Overall, an average of 3.1 jig ± 0.4 jig was
adsorbed to the coupon surfaces (cement and
backing combined) out of a total of 1,000 jig
of chlordane (0.31%) that was available in
the bulk contamination solution. Albeit to a
small percentage, the results show that
chlordane reproducibly adsorbed to the
surface of the cement coupons as well as the
polycarbonate backing.
During this experiment, very similar
amounts of chlordane adsorbed to the
cement surface and the polycarbonate
backing. This was compared to the previous
experiment during which the chlordane was
spiked directly onto the coupons and more
chlordane ended up adsorbing to the cement
surface. It is not entirely clear what caused
this, but it likely has something to do with
the duration of contaminant exposure. In the
first experiment, only five drops of
contaminated solution were added to the
coupon while in the second experiment, the
coupon was equilibrated with the
contaminated solution for two hours,
providing more opportunity for the
chlordane to come to equilibrium between
the two components of the coupon.
25
-------
Table 11. Chlordane on Cement Surface Contamination
Contaminated
Coupon
#1
#2
#3
#4
#5
Avg.
St. Dev.
%RSD
Amount Recovered
from Cement (jig)
.0
.4
.3
.7
.3
.4
0.3
19%
Amount Recovered
from Backing (jig)
2.3
1.1
2.1
1.4
1.4
1.7
0.5
30%
Total Amount
Recovered from
Coupon (jig)
3.3
2.5
3.4
3.1
2.8
3.1
0.4
12%
2.1.4 Chlordane on Cement Persistence
Evaluation
Figure 1 shows the results from the
persistence evaluation for chlordane on the
cement coupon surfaces as well as the
polycarbonate backing. The vertical axes
show the amount of chlordane remaining on
the coupons after each time period (shown
across the horizontal axis) during which
fresh tap water is flowing through the AR
and the AR is rotating at 100 rpm (1 ft/s).
The average free chlorine concentration in
the tap water during this evaluation was 1.54
mg/L ±0.17 mg/L, the average pH was 7.6 ±
0.1, and average temperature was 24.5°C
±0.7 °C. The columns at the far left side of
the graphs represent the initial
contamination level (as measured on the
contaminated control coupons) and each
successive column represents the time
periods and experimental conditions defined
by the PDEDP. The error bars on the graphs
are the standard deviations of the remaining
chlordane on the three coupons. The %P
that corresponds with each time period is
given across the top of each graph.
:100±55 53±21 25±12 9±3 9±5 4±2
Oh 24 h 4 h 24 h 72 h 168 h
hold
:100±16 128±55 58±17 50±27 49±13 38±10
Oh 24 h 4 h 24 h 72 h 168 h
hold
Figure 1. Persistence evaluation - percent persistence and chlordane remaining on cement
(left) and backing (right)
In order to further clarify the data, t-tests
were performed to determine what time
periods exhibited significant differences
from one another at the 95% confidence
interval. The null hypotheses of the t-tests
were that the difference in amount of
chlordane remaining on the coupons across
the various time periods was zero. The
26
-------
probabilities (p) generated by the t-test were
the probabilities of the null hypothesis being
confirmed. Therefore, p-values less than
0.05 indicated a small likelihood that the
difference between the two data sets was
zero, and thus, are considered to be
significantly different from one another.
Table 12 gives the p-values for comparisons
of each possible set of coupons collected at
the various time periods. The data that
exhibited significant differences are
highlighted in gray. For the cement, the
initial contamination level was not
significantly different from the 24 h hold
level (largely due to the rather high
variability in the initial concentration
chlordane level), but the chlordane levels at
the initial contamination, after the 24 h hold,
and 4 h after resumption of flow were all
significantly different from the chlordane
levels collected 24 h, 72 h, and 168 h after
the resumption of flow. Therefore, after the
initial 24 h hold, the residual chlordane
decreased until 24 h after the resumption of
flow and then the chlordane residuals
became steady. The cement was initially
contaminated with 500 ng ± 200 ng of
chlordane and 24 h after the resumption of
flow, the chlordane levels had decreased to
45 ng ± 16 ng which was not significantly
different from the levels at 72 h (46ng ± 14
ng) or 168 h (20ng ± 4 ng). The %P after 24
h of flow (after which there was no
additional decrease) was 9% ± 3%.
For the backing, the chlordane residual
decreased through 4 h after the flow was
resumed and then there was no further
decrease until 168 h. The 24 h, 72 h, and
168 h samples were not different from one
another, indicating the steady residual after
the 4 h sample. The backing was initially
contaminated with 1,700 ng ± 230 ng of
chlordane and 4 h after the resumption of
flow, the chlordane levels had decreased to
950 ng ± 250 ng and no further significant
decrease was noted until 168 h when the
chlordane levels were 640 ng ± 140 ng. The
%P after 168 h of flow was 38% ± 10%.
The increased %P for the backing with
respect to the cement was likely due to the
fact that the shear of the flowing water more
directly impacted the cement surface which
served to shield the backing.
Table 12. Chlordane on Cement - Probability Value Matrix for Persistence Evaluation
Persistence
Evaluation Times
probability (p) values (< 0.05 - significant difference)
Oh
24 h hold
Cement
4h
24 h
72 h
Oh
24 h hold
Backing
4h
24 h
72 h
Read as matrix, for times at left, read right for p-value to determine possible differences.
Light shading - significant differences
27
-------
2.7.5 Chlordane on Cement Flushing
Evaluation
Figure 2 shows the results from the flushing
evaluation for chlordane on the cement
coupon surfaces as well as the polycarbonate
backing. As was the case for the persistence
evaluation, the vertical axes show the
amount of chlordane remaining on the
coupons after each time period and flushing
condition that is shown across the horizontal
axes. The average free chlorine
concentration in the tap water during this
evaluation was 1.45 mg/L ±0.17 mg/L, the
average pH was 7.6 ±0.1, and average
temperature was 25.4°C ± 0.3 °C. The
columns at the far left side of the graphs
represent the initial contamination level (as
measured on the contaminated control
coupons) and each successive column
represents the time periods and experimental
conditions defined by the PDEDP. The
error bars on the graphs are the standard
deviations of the remaining chlordane on the
three coupons that were collected at each
time period. The %P that corresponds with
each time period is given across the top of
each graph.
: 100±27 24±6 23±6 6±1 2±1 3±1
P:100±39 64±19 53±16 25±8 15±6 47±22
2000
Figure 2. Flushing evaluation - percent persistence and chlordane remaining on cement
(left) and backing (right)
Similar to the persistence evaluation,
statistical analyses were performed using t-
tests to further clarify any differences
between the data from each flushing
scenario. Table 13 gives the p-values for
comparisons of each possible set of coupons
collected at the various flushing conditions.
The significant differences are highlighted
in gray. For the cement, the initial
contamination level was significantly
different from all of the other scenarios. In
addition, while the residual chlordane after 2
h and 4 h of 200 rpm (1.64 ft/s) was not
different, the residual chlordane decreased
with each scenario until there was no change
between the 4h and 24 h 250 rpm (1.91 ft/s)
samples. The cement was initially
contaminated with 530 ng ± 70 ng of
chlordane and it decreased to 130 ng ± 27 ng
after 2 h at 200 rpm where it held steady for
the next 2 h and decreased to 29 ng ± 2 ng
after 24 h at 200 rpm. Another significant
decrease took place after the rotation of the
AR was increased to 250 rpm for 4 h (11 ng
± 4 ng) which was not significantly different
that the chlordane levels after 24 h at 250
rpm (18 ng ± 7 ng). The %P after the time
period including 24 h of 200 rpm and 4 h of
250 rpm (after which there was no
additional decrease) was 2%± 1%.
28
-------
Table 13. Chlordane on Cement - Probability Value Matrix for Flushing Evaluation
Flushing
Evaluation
Conditions
probability (p) values (< 0.05 - significant difference)
Cement
Backing
Oh
2hr - 200 rpm
4hr - 200 rpm
24 hr - 200 rpm
4 hr - 250 rpm
Oh
2hr - 200 rpm
4hr - 200 rpm
24 hr - 200 rpm
4 hr - 250 rpm
Read as matrix, for conditions at left, read right for p-value to determine possible differences.
Light shading - significant differences
For the backing, the chlordane residual
decreased from the initial contamination to
the 2 h and 4 h 200 rpm (1.64 ft/s) samples
(that were not different from one another)
and then the chlordane residual decreased
after 24 h at 200 rpm and then again after 4
h at 250 rpm. However, then the chlordane
residual increased in the 24 h 250 rpm
samples. The chlordane level on the
backing decreased from an initial
concentration of 1,500 ng ± 400 ng to 930
ng ± 95 ng after 2 h at 200 rpm where it held
steady for the next 2 h and decreased to 370
ng ± 51 ng after 24 h at 200 rpm. Another
significant decrease took place after the
rotation of the AR was increased to 250 rpm
for 4 h (220 ng ± 58 ng), but then the
observed chlordane level unexpectedly
increased after 24 h at 250 rpm. There was
no apparent reason for this increase. The
%P after 24 h of 200 rpm and 4 h of 250
rpm was 15% ± 6%. The increased %P for
the backing with respect to the cement was
likely for the same reasons as the similar
observation during the persistence
evaluation.
2.1.6 Chlordane on Cement
Hyperchlormatron Evaluation
Figure 3 shows the results from the
hyperchlorination evaluation for chlordane
on the cement coupon surfaces as well as the
polycarbonate backing in a similar way as
was done for the persistence and flushing
evaluations. The columns at the far left side
of the graphs represent the initial
contamination level (as measured on the
contaminated control coupons) and each
successive column represents the time
periods and experimental conditions defined
by the PDEDP, specifically, the amount of
time that the coupons were exposed to either
25 mg/L or 50 mg/L free chlorine. The error
bars on the graphs are the standard
deviations of the remaining
29
-------
P: 100±35 103±26 105±43 126±78 71±21 70±21
: 100±30 117±44 118±36 106±32 110±28 113±28
2h25
mg/L
4 h 25 24 h 25 4 h 50 24 h 50
mg/L mg/L mg/L mg/L
2h25
mg/L
4h25
mg/L
24 h 25 4 h 50 24 h 50
mg/L mg/L mg/L
Figure 3. Hyperchlorination evaluation - percent persistence and chlordane remaining on
cement (left) and backing (right)
chlordane on the three coupons that were
collected at each time period. The %P that
corresponds with each time period is given
across the top of each graph.
As for the persistence and flushing
evaluations, statistical analyses were
performed using t-tests to further clarify any
differences between the data from each
flushing scenario. Table 14 gives the p-
values for comparisons of each possible set
of coupons collected at the various
Table 14. Chlordane on Cement - Probability Value Matrix for Hyperchlorination
Evaluation
Hyperchlorination
Evaluation
Conditions
probability (p) values (< 0.05 - significant difference)
Oh
2 h 25 mg/L FC
Cement
4 h 25 mg/L FC
24 h 25 mg/L FC
4 h 50 mg/L FC
Oh
Backing
2 h 25 mg/L FC
4 h 25 mg/L FC
24 h 25 mg/L FC
4 h 50 mg/L FC
Read as matrix, for conditions at left, read right for p-value to determine possible differences.
Light shading - significant differences
FC-free chlorine
hyperchlorination conditions. The
significant differences are highlighted in
gray. For the cement, the only significant
differences occurred between the residual
chlordane concentration after 2 h exposure
to 25 mg/L free chlorine and the residual
chlordane present after exposure to both 4 h
and 24 h of 50 mg/L free chorine. These
data suggested that hyperchlorination with
no flow is not an effective decontamination
30
-------
approach for chlordane on cement.
Similarly, for the backing, there were no
differences in residual chlordane
concentration through the duration of the
hyperchlorination experiment.
2.2 Results from Testing with
Chlordane on PVC Pipe Coupons
The following sections describe results from
performing quality control, verification, and
evaluation experimental design procedures
for chlordane on PVC coupons.
2.2.1 Chlordane on PVC Quality Control
Results
The same QC procedures were followed for
these measurements as in the previous
section. Tables 15 and 16 show the results
obtained during testing.
Table 15. Chlordane on PVC GC-MS Continuing Calibration Verification Results
Component of Testing
Step 1 - Surface Extraction
Method Verification
Step 2 - Surface Contamination
Method Verification
Persistence Evaluation
Flushing Evaluation
Hyperchlorination Evaluation
Average
Standard Deviation
Low Calibration Standard
(%R of Expected)
80%
72%
141%
113%
42%
0%
0%
0%
32%
174%
137%
140%
130%
124%
85%
61%
Mid Calibration Standard
(%R of Expected)
45%
55%
95%
97%
81%
61%
69%
79%
77%
92%
83%
84%
91%
90%
78%
16%
For the CCV samples, the recoveries of the
low concentration samples ranged from 0%
to 174%. These low concentration samples
were very close to the LOQ so small
changes in peak area greatly impacted the
%Rs. During the persistence and flushing
measurement, the 5 ng/mL standard was not
detectable during the analysis set. However,
the low end of the concentration range was
not applicable to these samples. Throughout
testing, the peak areas that most of the
samples were measured at were in the
middle and higher parts of the calibration
curve and often the samples had to be
diluted to bring the peak areas into the linear
range of the calibration curve. Therefore, no
corrective action was taken in response to
these CCV results. The recoveries of the
middle concentration CCV samples (100
ng/mL) ranged from 45% to 97% with an
average recovery of 78% with a standard
deviation of 16%. The two lowest
recoveries (45% and 55%) were during the
Step 1 surface extraction method
verification which was meant to determine if
the chlordane could be extracted from the
31
-------
surface of the cement coupons. No
corrective action was taken because of the
qualitative nature of the question being
explored in Step 1. For the rest of the 100
ng/mL CCV samples, only two were outside
of the acceptable range (61% and 69%) and
those were both in the same sample sets with
another 100 ng/mL CCV sample that was
within the acceptable range of recoveries.
Therefore, no corrective action was taken.
Table 16. Chlordane on PVC Laboratory Fortification Matrix Sample Results
Component of Testing
Step 2 - Surface
Contamination Method
Verification
Persistence Evaluation
Flushing Evaluation
Hyperchlorination
Evaluation
Average
Standard Deviation
Laboratory Fortified Matrix
(%R)
98%
98%
151%
105%
237%
106%
118%
130%
51%
For the LFM samples, with the exception of
two samples with recoveries of 151% and
237%, the recoveries ranged from 98% to
118%. The two outlying samples occurred
during analysis of the flushing evaluation.
These samples were in an analysis set with
one other LFM samples recovered at 105%
and two CCV samples that were within the
acceptable range. There was no clear reason
why these two samples were over recovered.
Because of the reasons stated, and because
the flushing data is interpreted based on the
relative change in concentration over time,
no corrective action was made.
As for the chlordane testing on the cement
coupons, the bulk contamination solution
was sampled at the beginning, middle, and
end of the 2 h contamination time during the
Step 2 method verification, the persistence
evaluation, the flushing evaluation, and the
hyperchlorination evaluation and analyzed
as described above. Across those four
experiments, the peak areas of the initial 1
milligram (mg)/L contamination solution
were considered the 100% chlordane levels.
The samples collected at the 1 h point of the
contamination step retained a 25%±1% of
the peak area and the sample collected at the
end of the contamination period retained
20%±2% of the peak area. Therefore, as in
the previous example using the cement
coupons, while there was a considerable loss
of chlordane during the first hour, chlordane
was available throughout the entire 2 h
contamination period.
2.2.2 Method Verification Step 1:
Chlordane on PVC Surface Extraction
Table 17 gives the results from the surface
contamination extraction method
verification for chlordane on PVC. Overall,
the total recovery ranged from 35% to 62%
with standard deviations across the five
replicates of less than 14%, indicating that
chlordane could be reproducibly extracted
and measured from the PVC coupons. The
concentration of each spiking solution was
confirmed using GC-MS. The low, middle,
32
-------
and high spiking solutions had percent
recoveries of 61%, 53%, and 68% of the
target concentration levels.
Table 17. Chlordane on PVC Surface Contamination Extraction
Low level
Mid level
High level
Amount
spiked
(ng)
20
100
1000
Avg. amount
recovered from
PVC (ng)
9.4
35
620
Total %
Recovery
47%
35%
62%
SD
14%
5%
7%
Five replicates were spiked and extracted at each concentration level.
2.2.3 Method Verification Step 2:
Chlordane on PVC Surface Contamination
This verification indicated if a contaminant
would adsorb to the PVC surface containing
biofilm in the event that it is exposed to a
bulk solution. Table 18 gives the results
from the surface contamination method
verification for chlordane on PVC including
the amount of chlordane extracted from each
coupon after a two hour exposure to 1 mg/L
chlordane. Overall, an average of 3.6 jig ±
0.6 jig was adsorbed to the coupon surfaces
out of a total of 1,000 jig of chlordane that
was available in the bulk contamination
solution (0.36%). As for the cement, albeit
a small percentage, these results indicate
that chlordane did adsorb to the PVC
coupon following exposure to the bulk
contamination solution.
Table 18. Chlordane on PVC Surface Contamination
Contaminated
Coupon
#1
#2
#3
#4
#5
Avg.
St. Dev.
%RSD
Amount Recovered
from PVC (jig)
3.4
2.8
4.4
3.4
3.8
3.6
0.6
17%
2.2.4 Chlordane on P VC Persistence
Evaluation
Figure 4 shows the results from the
persistence evaluation for chlordane on the
PVC coupon. The vertical axes show the
amount of chlordane remaining on the
coupons after each time period (shown
across the horizontal axis) during which
fresh tap water is flowing through the AR
and the AR is rotating at 100 rpm (1 ft/s).
The average free chlorine concentration in
the tap water during this evaluation was 1.34
mg/L ±0.11 mg/L, the average pH was 7.8
±0.1, and average temperature was 25.5°C
± 0.0 °C. The columns at the far left side of
the graphs represent the initial
contamination level (as measured on the
33
-------
contaminated control coupons) and each
successive column represents the time
periods and experimental conditions defined
by the PDEDP. The error bars on the graphs
are the standard deviations of the remaining
chlordane on the three coupons that were
collected at each time period. The %P that
corresponds with each time period is given
across the top of the graph.
94±6
51±5
14±4
7±2
4±3
Oh 24 h hold 4h 24 h 72 h 168 h
Figure 4. Persistence evaluation - percent persistence and chlordane remaining on PVC
As for the chlordane cement evaluations
described above, statistical analyses were
performed using t-tests to further clarify any
differences between the data from each
flushing scenario. Table 19 gives the p-
values for comparisons of each possible set
of coupons collected at the various flushing
conditions. The data that exhibit significant
differences are highlighted in gray. The
initial contamination level was not
significantly different from the 24 h hold
level, but the chlordane levels at the initial
contamination and 24 h hold were
significantly different from the chlordane
levels on the rest of the coupons. The
chlordane levels dropped significantly from
the initial and 24 h hold levels after 4 h and
again after 24 h of resumed flow.
Thereafter, the chlordane concentration
steadied with only another significant
difference between the 24 h and 168 h
chlordane levels. The PVC was initially
contaminated with 3,900 ng ± 200 ng of
chlordane and 4 h after the resumption of
AR rotation at 100 rpm (1 ft/s), the
chlordane levels had decreased to 2,000 ng ±
150 ng and after 24 h the levels decreased to
540 ng ± 130 ng which was not significantly
different from the levels at 72 h (260 ng ±
94 ng) and the 72 h chlordane levels were
not different from the 168 h chlordane levels
(180 ng ± 110 ng), but the 168 h levels had
decreased in comparison to the 24 h levels.
The %P after 24 h of flow was 14% ± 4%
and after 168h, 5% ±3%.
34
-------
Table 19. Chlordane on PVC - Probability Value Matrix for Persistence Evaluation
Persistence
Evaluation Times
Oh
probability (p) values (< 0.05 - significant difference)
24 h hold
4h
24 h
72 h
Read as matrix, for times at left, read right for p-value to determine possible differences.
Light shading - significant differences
2.2.5 Chlordane on P VC Flushing
Evaluation
Figure 5 shows the results from the flushing
evaluation for chlordane on the PVC
coupons. As was the case for the
persistence evaluation, the vertical axes
show the amount of chlordane remaining on
the coupons after each time period and
flushing condition that is shown across the
horizontal axes. The columns at the far left
side of the graphs represent the initial
contamination level (as measured on the
contaminated control coupons) and each
successive column represents the time
periods and experimental conditions defined
by the PDEDP. The average free chlorine
concentration in the tap water during this
evaluation was 1.34 mg/L ± 0.11 mg/L, the
average pH was 7.8 ± 0.1, and average
temperature was 25.5°C ± 0.0 °C. The error
bars on the graphs are the standard
deviations of the remaining chlordane on the
three coupons that were collected at each
time period. The %P that corresponds with
each time period is given across the top of
the graph.
68±20
37±14
14±6
11±4
8±6
Oh 2h 200 rpm 4h 200 rpm 24 h 200 rpm 4 h 250 rpm 24 h 250 rpm
Figure 5. Flushing evaluation - percent persistence and chlordane remaining on PVC
Statistical analyses were performed using t-
tests to further clarify any differences
between the data from each flushing
scenario. Table 20 gives the p-values for
comparisons of each possible set of coupons
collected at the various flushing conditions.
The data that exhibit significant differences
are highlighted in gray. The initial
contamination level (2,300 ng ± 630 ng)
was significantly different from all of the
other scenarios and the residual chlordane
levels decreased significantly until the
significant decreases in residual chlordane
stopped after the 24 h of the AR rotating at
35
-------
200 rpm (1.64 ft/s) (320 ng ± 96 ng).
Increasing the AR rotation to 250 rpm (1.91
ft/s) did not cause additional decreases in the
residual chlordane levels. The %P after 24 h
of 200 rpm (1.64 ft/s) rotation (after which
there was no additional decrease) was 14%
± 6%.
Table 20. Chlordane on PVC - Probability Value Matrix for Flushing Evaluation
Flushing
Evaluation
Conditions
Oh
probability (p) values (< 0.05 - significant difference)
2hr - 200 rpm
4hr - 200 rpm
24 hr - 200 rpm
4 hr - 250 rpm
Read as matrix, for conditions at left, read right for p-value to determine possible differences.
Light shading - significant differences
2.2.6 Chlordane on PVC
Hyperchlormatron Evaluation
Figure 6 shows the results from the
hyperchlorination evaluation for chlordane
on PVC coupons. The columns at the far
left side of the graphs represent the initial
contamination level (as measured on the
contaminated control coupons) and each
successive column represents the time
periods and experimental conditions defined
by the PDEDP, specifically, the amount of
time that the coupons were exposed to either
25 mg/L or 50 mg/L free chlorine. The error
bars on the graphs are the standard
deviations of the remaining chlordane on the
three coupons that were collected at each
time period. The %P that corresponds with
each time period is given across the top of
the graph.
2000
1500
P: 100±21
71+11
82±20
88±15
71±11
72±20
1000
500
n
Initial cont. 2 h 25 mg/L 4 h 25 mg/L 24 h 25 mg/L 4 h 50 mg/L 24 h 50 mg/L
Figure 6. Hyperchlorination evaluation - percent persistence and chlordane remaining on
PVC
As for the persistence and flushing
evaluations, statistical analyses were
performed using t-tests to further clarify any
differences between the data from each
flushing scenario. Table 21 gives the p-
values for comparisons of each possible set
of coupons collected at the various
hyperchlorination conditions. The data
exhibiting significant differences are
highlighted in gray. Overall, the statistical
evaluation confirmed the visual observation
of the data in the graphs. There were several
36
-------
significant differences, but no two that were
in succession to clarify the effect of the
hyperchlorination. Instead the data seem to
be indicating that hyperchlorination does not
cause significant and repeatable
decontamination of chlordane from PVC.
Table 21. Chlordane on PVC - Probability Value Matrix for Hyperchlorination
Evaluation
Hyperchlorination
Evaluation
Conditions
Oh
probability (p) values (< 0.05 - significant difference)
2 h 25 mg/L FC
4 h 25 mg/L FC
24 h 25 mg/L FC
4 h 50 mg/L FC
Read as matrix, for conditions at left, read right for p-value to determine possible differences.
Light shading - significant differences
FC-free chlorine
2.3 Results from Testing with Sodium
Fluoroacetate on Cement Pipe Coupons
The following sections describe results from
performing quality control, verification, and
evaluation experimental design procedures
for SFA on cement pipe coupons.
2.3.1 SFA on Cement Quality Control
Results
Continuing calibration verification (CCV)
samples were analyzed on the 1C throughout
each analysis set. After every 10 samples
analyzed, a middle concentration calibration
solution (0.5 mg/L) was reanalyzed and
following each analysis set, the low
calibration solution (0.1 mg/L) were
reanalyzed. There were 34 middle
concentration CCV samples analyzed and
the recoveries ranged from 95% to 105%
with an average of 99% and a standard
deviation of 2%. Ten low calibration CCV
samples were analyzed and the recoveries
ranged from 95% to 126% with an average
of 106% with a standard deviation of 12%.
Overall, none of the middle level CCV
samples were outside of the targeted range
of recoveries and only two of the low level
CCV samples were outside of the targeted
range. In addition, 10% of the samples were
split and 0.5 mg/L of chlordane was spiked
into the sample extract to create LFM
samples. Target recoveries for each of these
QC samples were between 90% and 110%.
The recoveries of the LFM samples are
shown in Table 22.
For the LFM samples, the recoveries ranged
from 78% to 238% with an average recovery
of 119% and a standard deviation of 44%.
Only five out of the 22 LFM samples were
outside of the targeted range of 70% to
130% recovery and LFM results outside of
the acceptable range were always
accompanied with at least three other LFM
samples that were within the targeted range.
If the five outlying LFM results were
removed, the average recovery would be
96% with a standard deviation of 11%.
There was not a clear explanation as to why
those five samples were over-recovered, but
because of the number of samples that were
within the acceptable range, no corrective
action was made.
37
-------
Table 22. SFA on Cement Laboratory Fortification Matrix Sample Results
Component of Testing
Step 1 - Surface Extraction
Method Verification
Step 2 - Surface Contamination
Method Verification
Persistence Evaluation
Flushing Evaluation
Hyperchlorination Evaluation
Average
Standard Deviation
Laboratory Fortified Matrix
(%R)
92%
98%
78%
94%
90%
88%
86%
184%
90%
191%
238%
109%
117%
103%
184%
97%
88%
119%
174%
95%
88%
106%
119%
44%
The bulk contamination solution was
sampled at the beginning, middle, and end
of the 2 h contamination time during the
persistence evaluation, the flushing
evaluation, and the hyperchlorination
evaluation and the SFA measured
quantitatively. Across those three
experiments and three collection times
during each experiment, the recovery of
SFA from the 500 mg/L bulk contamination
solution was 89%±2%. Therefore, most of
the SFA remained available for adsorption
throughout the duration of the 2 h
contamination time period.
2.3.2 Method Verification Step 1: SFA on
Cement Surface Extraction
Table 23 gives the results from the surface
contamination extraction method
verification for SFA on cement. When the
SFA was spiked onto the cement coupon,
some of the SFA adsorbed to the cement
surface and some flowed through the cement
and adsorbed to the polycarbonate backing
on which the cement was mounted. The
cement and backing were extracted
separately and the results were reported for
both the cement and the backing for all five
components of the testing. Table 23 gives
the results including the amount of SFA
38
-------
spiked onto the coupons, the amount
extracted from the backing and cement, the
total recovery, and the standard deviation.
Overall, the total recovery ranged from 68%
to 91% with standard deviations across the
five replicates of less than 28%, indicating
that SFA could be extracted and measured
from the cement coupons. The amounts
recovered from the backing and cement
show that considerably more SFA adsorbed
to the cement surface than being adsorbed to
the polycarbonate backing. This is
consistent with the chemical characteristics
of SFA, as preferential adsorption would be
expected from a highly non-polar organic
chemical as opposed to SFA, a salt. The
concentration of each spiking solution was
confirmed using 1C. The low, middle, and
high spiking solutions had average percent
recoveries of 93% ±1% of the target
concentration levels.
Table 23. SFA on Cement Surface Contamination Extraction
Spike
Level
Low level
Mid level
High level
Amount
spiked
(HZ)
10
50
500
Avg. amount
recovered from
cement(jig)
6.9
25
220
Avg. amount
recovered from
backing (ug)
2.2
11
130
Avg. total
recovered (jig)
9.1
36
340
Total %
Recovery
91%
72%
68%
SD
28%
10%
3%
Five replicates were spiked and extracted at each concentration level.
2.3.3 Method Verification Step 2: SFA on
Cement Surface Contamination
This verification indicates if a contaminant
will adsorb to the cement surface containing
biofilm in the event that it is exposed to a
bulk solution. Table 24 gives the results
from the surface contamination method
verification for SFA on cement including the
amount of SFA extracted from each part of
the coupon after a two hour exposure to 1
mg/L SFA. Overall, an average of 55 jig ±
17 ng was adsorbed to the coupon surfaces
(cement and backing combined) out of a
total of 500,000 |ig of SFA that was
available in the bulk contamination solution
(0.011%). These data indicated that SFA
adsorbed to the cement coupon following
exposure to the bulk contamination solution
in a similar way as it did during the surface
extraction method verification, with more
SFA having adsorbed to the cement surface
than to the polycarbonate backing.
During this verification, one set of coupons
was contaminated as described above only
with a 100 mg/L SFA contamination
solution (which had been used to
contaminate coupons at a detectable level
during some method development work),
but the SFA was not able to be detected
following extraction. This set of coupons
had been contaminated following growth of
biofilm as described in Section 1.2. Prior to
that, another method was being used to grow
biofilm. Because of the lack of measured
SFA, it was suspected that in prior
experiments, biofilm had not been grown on
the coupons and now that it had been, SFA
was not adsorbing as readily. The method
verification was repeated using a 500 mg/L
SFA contamination solution. Within this set
of coupons, two coupons were included
(only two because of the limited capacity of
the AR) that had no biofilm growth in order
to get some indication as to whether biofilm
growth played a role in the lack of
adsorption of SFA. The two coupons
without biofilm had 4-5 times the amount of
39
-------
SFA (225 |ig SFA) adsorbed onto the
cement surface of the biofilmed coupons (49
jig). While this observation is based on very
little data, it suggests that at least for
Table 24. SFA on Cement Surface Contamination
Contaminated
Coupon
#1
#2
#3
#4
#5
Avg.
St. Dev.
%RSD
Amount Recovered
from Cement (ug)
37
73
40
48
46
49
14
29%
Amount Recovered
from Backing (ug)
3.0
10
5.8
6.7
5.2
6.2
2.7
43%
Total Amount
Recovered from
Coupon (jig)
40
83
46
55
51
55
17
31%
SFA, an ionic bonding chemical, that
biofilm hinders its adsorption to cement
surfaces. More research would be required
to further characterize the behavior of this
and other contaminants with biofilms.
2.3.4 SFA on Cement Persistence
Evaluation
Figure 7 shows the results from the
persistence evaluation for SFA on the
cement coupon surfaces as well as the
polycarbonate backing. The vertical axes
show the amount of SFA remaining on the
coupons after each time period (shown
across the horizontal axis) during which
fresh tap water is flowing through the AR
and the AR is rotating at 100 rpm (1 ft/s).
The average free chlorine concentration in
the tap water during this evaluation was 1.46
mg/L ±0.12 mg/L, the average pH was 7.9
± 0.2, and average temperature was 24.0°C
± 1.0 °C. The columns at the far left side of
the graphs represent the initial
contamination level (as measured on the
contaminated control coupons) and each
successive column represents the time
periods and experimental conditions defined
by the PDEDP. The error bars on the graphs
are the standard deviations of the remaining
SFA on the three coupons that were
collected at each time period. The %P that
corresponds with each time period is given
across the top of the graphs.
40
-------
:100±74 102±69 138±93 179±124 16±69 96±50
: 100±53 105±46 139±92 100±42 145±83 55±22
Oh 24 h 4 h 24 h 72 h 168 h
hold
Oh 24 h 4 h 24 h 72 h 168 h
hold
Figure 7. Persistence evaluation - percent persistence and SFA remaining on cement (left)
and backing (right)
Table 25 gives the p-values for comparisons of each possible set of coupons collected at the
various time periods. There was only one significant difference across all of the
Table 25. SFA on Cement - Probability Value Matrix for Persistence Evaluation
Persistence
Evaluation Times
probability (p) values (< 0.05 - significant difference)
Oh
24 h hold
Cement
4h
24 h
72 h
Oh
24 h hold
Backing
4h
24 h
72 h
Read as matrix, for times at left, read right for p-value to determine possible differences.
Light shading - significant differences
combinations of data sets and it is
highlighted in gray. For neither cement nor
the backing did the levels of residual SFA
change significantly due to the scenarios
tested during this evaluation. The only
significant difference between coupon
collection periods was a decrease in residual
SFA on the backing between the 24 h after
flow was resumed and the 168 h sample.
However, the 72 h sample collected in
between those two did not exhibit a
significant difference, further exemplifying
the scattered nature of the results. The %P
after the persistence evaluation was 96%
±50% for the cement and 55% ±22% for the
backing.
2.3.5 SFA on Cement Flushing Evaluation
Figure 8 shows the results from the flushing
evaluation for SFA on the cement coupon
surfaces as well as the polycarbonate
41
-------
backing. As was the case for the persistence
evaluation, the vertical axes show the
amount of SFA remaining on the coupons
after each time period and flushing condition
that is shown across the horizontal axes.
The average free chlorine concentration in
the tap water during this evaluation was 1.62
mg/L ±0.18 mg/L, the average pH was 7.8
± 0.1, and average temperature was 23.8 °C
± 0.9 °C. The columns at the far left side of
the graphs represent the initial
contamination level (as measured on the
contaminated control coupons) and each
successive column represents the time
periods and experimental conditions defined
by the PDEDP. The error bars on the graphs
are the standard deviations of the remaining
P:100±44 110±36 82±26 132±64 123±61 101±34
P: 100±45 59±24 83±27 83±50 110±36 116±45
2hr 4hr 24hr200 4hr 24hr250
200 rpm 200 rpm rpm 250 rpm rpm
Oh 2hr 4hr 24 hr 4hr 24 hr
200 rpm 200 rpm 200 rpm 250 rpm 250 rpm
Figure 8. Flushing evaluation - percent persistence and SFA remaining on cement (left) and
backing (right).
SFA on the three coupons that were
collected at each time period. The %P that
corresponds with each time period is given
across the top of the graphs.
As for the persistence evaluation, statistical
analyses were performed using t-tests to
further clarify any differences between the
data from each flushing scenario. Table 26
gives the p-values for comparisons of each
possible set of coupons collected at the
various flushing conditions. There was only
one significant difference across the various
flushing scenarios and it was highlighted in
gray. The statistical data indicated that there
was only one significant difference across
the cement and backing data. This data
suggests that SFA is not decontaminated
effectively by increasing the duration of
flushing and flow velocity past the cement
pipe coupons. The %P after the flushing
evaluation was 101% ±34% for the cement
and 116% ±45% for the backing.
42
-------
Table 26. SFA on Cement - Probability Value Matrix for Flushing Evaluation
Flushing
Evaluation
Conditions
probability (p) values (< 0.05 - significant difference)
Cement
Backing
Oh
2hr - 200 rpm
4hr - 200 rpm
24 hr - 200 rpm
4 hr - 250 rpm
Oh
2hr - 200 rpm
4hr - 200 rpm
24 hr - 200 rpm
4 hr - 250 rpm
Read as matrix, for conditions at left, read right for p-value to determine possible differences.
Light shading - significant differences
2.3.6 SFA on Cement Hyperchlorination
Evaluation
Figure 9 shows the results from the
hyperchlorination evaluation for SFA on the
cement coupon surfaces as well as the
polycarbonate backing as was done for the
persistence and flushing evaluations. The
columns at the far left side of the graphs
represent the initial contamination level (as
measured on the contaminated control
coupons) and each successive column
represents the time periods and experimental
conditions defined by the PDEDP,
specifically, the amount of time that the
coupons were exposed to either 25 mg/L or
50 mg/L free chlorine. The error bars on the
graphs are the standard deviations of the
remaining SFA on the three coupons that
were collected at each time period. The %P
that corresponds with each time period is
given across the top of the graphs.
: 100±34 132±82 125±77 141±69 88±48 152±88
:100±107 110±96 98±92 177±140 178±140 152±119
4h25
mg/L
24 h 25 4 h 50
mg/L mg/L
24H50
mg/L
2h25
mg/L
4 h 25 24 h 25 4 h 50 24 h 50
mg/L mg/L mg/L mg/L
Figure 9. Hyperchlorination evaluation - percent persistence and SFA remaining on cement
(left) and backing (right)
As for the persistence and flushing evaluations, statistical analyses were performed using t-tests
to further clarify any differences between the data from each flushing scenario. Table 27
43
-------
Table 27. SFA on Cement - Probability Value Matrix for Hyperchlorination Evaluation
Hyperchlorination
Evaluation
Conditions
probability (p) values (< 0.05 - significant difference)
Oh
Cement
2 h 25 mg/L FC
4 h 25 mg/L FC
24 h 25 mg/L FC
4 h 50 mg/L FC
Oh
Backing
2 h 25 mg/L FC
4 h 25 mg/L FC
24 h 25 mg/L FC
4 h 50 mg/L FC
Read as matrix, for conditions at left, read right for p-value to determine possible differences.
Light shading - significant differences
FC - free chlorine
gives the p-values for comparisons of each
possible set of coupons collected at the
various hyperchlorination conditions. The
data exhibiting significant differences are
highlighted in gray. There were several
significant differences between the data sets
from some of the experimental scenarios,
but no clear trends indicating that
hyperchlorination was an effective means
for decontaminating SFA from the surface
of cement pipes. The %P after the
hyperchlorination evaluation was 152%
±88% for the cement and 152% ±119% for
the backing.
44
-------
RESULTS SUMMARY
The objective of this project was to develop
a PDEDP that could be used across
laboratories to performed pipe
decontamination research. In addition, data
was to be collected pertaining to the
adsorption, persistence, and possible
decontamination approaches to chlordane
and sodium fluoroacetate on cement-line
pipe and/or PVC. Several key points of
summary are given below.
3.1 Experimental Design Protocol
Development
The development and testing of the PDEDP
was successfully accomplished. Use of the
annular reactor proved to be an effective
means of reproducibly simulating the flow
of water past pipe materials. The surface
extraction and surface contamination
method verification steps were necessary to
demonstrate whether or not a selected
contaminant can be studied (if it cannot be
extracted it will be difficult to study its
decontamination behavior) and if it is a
viable threat (if a contaminant will not
partition onto a pipe from an aqueous
solution, it may not be a decontamination
concern). These method verification steps
were demonstrated with a limited number of
replicates for chlordane and SFA. Each of
these method verifications could be more
rigorously tested by including more
replicates and additional separate
experiments and optimized (sonication time,
solvent, etc.) in order to provide additional
information on the reproducibility of the
pipe material coupon extraction for the
selected pipe material type and contaminant
as well as to more accurately determine the
extent of and reproducibility of the
contamination step.
The persistence evaluation was a beneficial
component of the PDEDP as it mimicked
rather typical conditions in a water
distribution system and it was compared
with the flushing evaluation at higher flow
velocities to determine if there was
increased efficacy at higher flow velocities.
Additional information could be gleaned
during this evaluation by controlling the
water quality parameters in order to study
how water quality parameters impact
contaminant adsorption and
decontamination efficacy. Lastly, the
hyperchlorination evaluation allowed for
collection of data using a chemical
decontamination approach. These results
were compared with the persistence and
flushing evaluations. Additional work could
be performed to include multiple other pipe
decontamination chemicals to compare the
effectiveness of those approaches with
hyperchlorination. Regardless of the
additional work that could be performed,
each of the PDEDP steps was successfully
demonstrated and the combined results
proved to be a useful data set.
45
-------
3.2 Persistence and Decontamination
Testing
3.2.7 Chlordane on Cement
The surface extraction method verification
confirmed that chlordane could be extracted
from the surface of cement after direct
contamination of the cement coupon and the
surface contamination method verification
confirmed that a cement coupon could be
contaminated with chlordane by exposing to
a solution of contaminated water. The
results from the persistence and flushing
evaluations for the cement exhibited very
similar results. The %P after 24 h for the
persistence evaluation (AR operated at 100
rpm (1 ft/s)) was 9% ± 3% and the %P after
24 h during the flushing evaluation (AR
operated at 200 rpm (1.64 ft/s)) was 6% ±
1%. However, during the flushing
evaluation, a further decrease was noted
during the next 4 h of the AR operating at
250 rpm (1.91 ft/s), taking the %P to 2% ±
1% for the flushing evaluation. These
results suggest that the flow velocity past the
pipe materials may have less to do with the
decontamination efficacy than the duration
of the flow past the contaminated pipe.
Results from the hyperchlorination
evaluation showed that hyperchlorination
without flow is not an effective means of
decontaminating chlordane from cement.
This result was unexpected as free chlorine
would be expected to oxidize the chlordane
from the surface of the cement. These data
suggest oxidation was not occurring to the
extent that was anticipated and flushing with
water with a concentration of 1-2 mg/L of
free chlorine was much more effective at
decontaminating the pipe materials than
water with a free chlorine concentration of
25 mg/L and 50 mg/L.
3.2.2 Chlordane onPVC
The surface extraction method verification
confirmed that chlordane could be extracted
from the PVC surface after direct
contamination of the PVC coupon and the
surface contamination method verification
confirmed that a PVC coupon could be
contaminated with chlordane by exposing to
a solution of contaminated water. The
results from the persistence and flushing
evaluations for the PVC exhibited very
similar results. The %P after 24 h for the
persistence evaluation (AR operated at 100
rpm (1 ft/s)) was 14% ± 4% and the %P
after 24 h during the flushing evaluation
(AR operated at 200 rpm (1.64 ft/s)) was
14% ± 6%. However, during the persistence
evaluation, a further decrease was noted
between 24 and 168 h, taking the %P to 5%
± 3% for the overall persistence evaluation.
As for the chlordane on cement results, these
results suggest that the flow velocity past the
pipe materials may have less to do with the
decontamination efficacy than the duration
of the flow past the contaminated pipe.
Again, as for the chlordane on cement
testing, results from the hyperchlorination
evaluation unexpectedly showed that
hyperchlorination without flow is not an
effective means of decontaminating
chlordane from PVC.
3.2.3 Sodium Fluoroacetate on Cement
The surface extraction method verification
confirmed that SFA could be extracted from
the surface of cement after direct
contamination of the cement coupon and the
surface contamination method verification
confirmed that a cement coupon could be
contaminated with SFA by exposing to a
solution of contaminated water. The results
from the persistence, evaluation, and
hyperchlorination evaluations suggest that
these approaches were not effective in
46
-------
decontaminating SFA from cement. These
results are exemplified by the %Ps. After
the persistence evaluation (AR operated at
100 rpm (1 ft/s)) the %P was 96% ± 50%,
after the flushing evaluation, 101% ±34%,
and after the hyperchlorination study, 152%
±88%.
3.3 Future Research Needs
The water system decontamination research
area is one with many facets to be explored.
This work has laid the framework for a
PDEDP that can be adapted to accommodate
other research priorities. Below are a few
possible areas for further study:
• Importance of biofilm to pipe
decontamination research - During
the SFA surface contamination
method verification step, two
cement coupons without biofilm
(only two because of the limited
capacity of the AR and that the
impromptu experiment was outside
the context of the PDEDP) were
contaminated with SFA along with
the coupons containing biofilm. For
these two coupons, approximately
five times as much SFA was
adsorbed to the non-biofilm
coupons. This very limited data set
suggested that the presence or
absence of biofilm could
significantly impact the results of
pipe adsorption/decontamination
research. More rigorous
experimentation could be performed
to better characterize the role of
biofilm.
Broadening of
adsorption/decontamination data set
by expanding on list of chemical
contaminants tested using the
PDEDP (e.g., organophosphates as
available toxic chemicals and
simulated chemical agents, metals to
simulate heavy metal or radiological
contamination).
Study of
adsorption/decontamination of
biological organisms using the
PDEDP.
Use of additional pipe materials with
additional chemicals and biological
organisms as well as additional
chemical pipe cleaning materials as
possible decontamination agents.
Scaling up of AR experiments into
experiments with real pipe using a
pipe loop in order to study how well
the AR experiments translate into
scenarios with real pipe.
Study of risk assessment questions
addressing how much persistence of
various chemicals is acceptable.
47
-------
REFERENCES
1. Welter, G., M. Lechevallier, S.
Spangler, J. Cotruvo, R. Moser,
Guidance for Decontamination of
Water System Infrastructure.
Denver, CO: AWWA Research
Foundation, 2007.
2. ASTM C 150-07 Standard, 2007,
"Standard Specification for Portland
Cement," ASTM International, West
Conshohocken, PA, www.astm.org.
3. American Water Works Association
C104-03 Standard, 2004, "Standard
for Cement-Mortar Lining for
Ductile-Iron Pipe and Fittings for
Water" Denver, CO,
www.awwa.org.
Szabo, J.G., Persistence and
decontamination of surrogate
radioisotopes in a model drinking
water distribution system, Water
Research (2009),
doi:10.1016/j.watres.2009.08.012.
USEPA, Pilot-Scale Tests and
Systems Evaluation for the
Containment, Treatment, and
Decontamination of Selected
Materials from T&E Building Pipe
Loop Equipment
(www.epa.gov/nhsrc/pubs/600r0801
6.pdf)
48
-------
APPENDIX
Experimental Design Protocol
49
-------
United States Environmental Protection
Agency
National Homeland Security
Research Center
Water Infrastructure
Protection Division
Experimental Design Protocol for the Study
of Chemical Contaminant Persistence and
Decontamination in Drinking Water Pipes
50
-------
Experimental Design Protocol for the Study of Chemical
Contaminant Persistence and Decontamination in Drinking Water
Pipes
February 10, 2012
51
-------
TABLE OF CONTENTS
Page
TABLE OF CONTENTS 52
LIST OF ABBREVIATIONS 53
INTRODUCTION 54
Al EXPERIMENTAL DESIGN 55
Al.l Experimental Reactor System 55
Al .2 Pipe Coupon Contamination Method Verification Experiments 56
Al .2.1 Method Verification Step 1: Surface Contamination Extraction 56
Al .2.2 Method Verification Step 2: Surface Contamination 59
A1.3 Evaluation of Contaminant Persistence 59
A1.4 Evaluation of Decontamination Approaches 61
Sections A2-A10 63
REFERENCES 64
Figure
Figure 1. Schematic of drops of contaminant solution across coupon surface 58
Tables
Table 1. Surface Contamination Extraction Method Verification (Step 1) 58
Table 2. Persistence Evaluation 60
Table 3. Evaluation of Flushing as a Decontamination Approach 61
Table 4. Evaluation of Hyperchlorination as a Decontamination Approach 63
52
-------
LIST OF ABBREVIATIONS
AR
ASTM
AWWA
cm
cfu
EPA
°C
F
ft/s
h
HC
HPC
1C
IS
in.
g
GC-MS
KOW
KOH
LOQ
LFM
L
Lpm
,iL
US
mA
mM
mg
mm
mL
min
ng
NHSRC
QAPP
QC
%R
%P
PE
PDEDP
PVC
rpm
s
SIM
SPME
annular reactor
ASTM International
American Water Works Association
centimeters
colony forming units
U.S. Environmental Protection Agency
degrees Celsius
flushing
feet per second
hour
hyperchlorination
heterotrophic plate counts
ion chromatography
internal standard
inch
gram
gas chromatographic mass spectrometry
octanol-water partitioning coefficient
potassium hydroxide
limit of quantitation
laboratory fortified matrix
liter
liter per minute
microliter
microgram
milliamp
millimolar
milligrams
millimeters
milliliters
minute
nanogram
National Homeland Security Research Center
Quality Assurance Project Plan
quality control
percent recovery
percent persistence
persistence evaluation
Persistence and Decontamination Experimental Design Protocol
polyvinyl chloride
revolutions per minute
second
selected ion monitoring
solid phase micro extraction
53
-------
INTRODUCTION
The U.S. Environmental Protection Agency's (EPA) National Homeland Security Research
Center (NHSRC) conducts research to protect, detect, respond to, and recover from terrorist
attacks on the nation's water and wastewater infrastructure. The objective of this project was the
development and testing of a standardized Persistence and Decontamination Experimental
Design Protocol (PDEDP) to quantitatively determine the adherence and persistence of
individual priority contaminants to the wetted surfaces of various drinking water pipe materials.
This experimental design also addresses testing of techniques for decontaminating affected pipe
surfaces if the contaminant persists. The experimental design can be implemented in a
consistently reproducible fashion across different laboratories for various contaminants and pipe
materials. The PDEDP is used to gain additional experimental information about the adsorption
of specific contaminants to various drinking water pipe materials and to test various methods to
destroy, reduce, or remove adsorbed contaminants.
Multiple research studies have already been conducted to determine the adsorption of particular
chemical, biological, and radiological contaminants to drinking water pipe materials and test
various methods to destroy, reduce, or remove adsorbed contaminants3"5. While useful data have
resulted from studies conducted to date, often the differing designs of previous studies limit the
usability and comparability of the data. This document describes a proposed experimental
design that could be used to generate contaminant persistence and decontamination data for
water utilities and other decision-makers with decontamination responsibility in the instance of
an intentional or natural contamination of a drinking water system. This experimental design
could also provide a means to generate data that are comparable to that which has been published
in the peer-review literature.
One of the most significant factors in this experimental design is the use of an annular reactor
(AR) as the device used to simulate flow past coupons of materials that represent drinking water
pipe surfaces. The AR simulates pipe flow with a variable speed motor that drives an inner
rotating cylinder, providing surface shear between pipe surface coupons and water within the
AR. Twenty removable slide coupons of relevant materials can be mounted within the reactor.
There are benefits and drawbacks of using the AR as the flow simulator. The main drawback of
using the AR is that actual pipe sections cannot be used as in some previous studies; pipe
material coupons either need to be purchased from the AR manufacturer or pipe materials need
to be attached to a standard backing that can be inserted into the AR.
Several benefits of using the AR outweigh these drawbacks, including the following:
• Provides option of altering rotational speed to simulate various flow velocities,
and therefore shear, to allow simulation of both flushing and decontamination
conditions
• Injection ports facilitate the precise alteration of water chemistry
• The AR manufacturer offers coupons with several common pipe materials, such
as cement lined and polyvinyl chloride (PVC). Cement lined coupons meet
requirements of the C150-07 American Society for Testing and Materials
(ASTM) Standard Specification for Portland Cement1 and the thickness of the
concrete is at least 1.6 millimeters (mm), as specified in American Water Works
54
-------
Association (AWWA) C104-03 Standard for Cement-Mortar Lining for Ductile-
Iron Pipe and Fittings for Water2.
• ARs are commercially available, providing ease of repeatability across
laboratories, as opposed to requiring the fabrication of flow cells at each
laboratory
• Several decontamination projects described in the literature have used the AR,2"5
making it possible to replicate the experimental conditions found in the literature
Overall, the measurement of persistence and decontamination of contaminants from pipe
material coupons is going to be challenging because of the small amounts of contaminant that are
to be recovered from coupon surfaces. To ensure the accuracy and precision of persistence and
decontamination data, it is important that as many experimental factors as possible be controlled.
The AR provides the best approach to providing experimental conditions that are adequately
controlled to attain usable persistence and decontamination data.
The following experimental design is meant to be generic, since it is intended for use with
various contaminants and pipe materials. Note that before following this experimental design,
the laboratory being used must be capable of measuring the contaminant used for contaminating
the pipe material and have at least one AR and an adequate number of AR coupons of the desired
pipe material.
Al EXPERIMENTAL DESIGN
Al.l Experimental Reactor System
For the persistence and decontamination experiments described in this experimental design, the
conditions within operational drinking water pipes are to be simulated in annular reactors (AR)
(BioSurface Technologies Corporation, Bozeman, MT). The ARs consist of a glass outer
cylinder and a rotating polycarbonate inner cylinder with 20 flush mounted rectangular coupons
that can be manufactured from materials such as polyvinyl chloride (PVC), steel, and concrete
and obtained from the manufacturer of the AR. These pipe material coupons, which have
surfaces that are .55 inch (in) (14 millimeters (mm)) x 5.8 in. (148 mm), simulate the inner
surface of drinking water pipes. Shear stress is to be applied to the coupon surfaces by setting
the inner cylinder rotation to 100 revolutions per minute (rpm), which produces shear similar to 1
foot (ft)/second (s) (30.5 centimeter (cm)/s) flow in a 6 in. (15.2 cm) pipe5. During normal
operation, the flow of drinking water through the AR (connected directly to the tap) is to be
maintained at a mean velocity of 200 mL/min so that mean the residence time of the water in the
AR is five minutes. This rapid flow velocity prevents the depletion of chlorine level over the
course of the experiments. The short residence time decreases the chance that desorbing
contaminant could re-contaminate an AR surface. The pH, temperature, and free chlorine
concentration of the drinking water are to be measured daily. The ARs are to always be operated
in the dark by covering them completely with aluminum foil or another opaque material. Some
contaminants may adsorb onto the polycarbonate components of the AR and affect the amount of
contaminant that is available for coupon contamination. To control against this adsorption
negatively impacting experiments, the bulk contamination solution is to be monitored to ensure
55
-------
that an adequate concentration of contaminant is maintained to achieve pipe coupon
contamination.
Prior to any persistence or decontamination experiments, a biofilm is to be grown on the
coupons by submerging the required number of coupons into a container that allows
recirculation of dechlorinated tap water (outlet near the top of the container and inlet near the
bottom of the container) fortified with 1 gram (g) of yeast extract. This water is to be kept in the
dark and be recirculated using a pump for four days with an additional 1 g of yeast added after
two days. The biofilm growth is to be measured, using heterotrophic plate counts (HPC), on one
of the 20 pipe material coupons in the AR. The four-day time period for biofilm growth also
serves to condition the pipe material coupons in flowing water prior to coupon contamination.
Note that the extent of biofilm growth on the pipe material coupons can have a significant effect
on how much contaminant is adsorbed to the pipe coupon so it is important to confirm its
presence.
A1.2 Pipe Coupon Contamination Method Verification Experiments
The generation of persistence and decontamination data from this experimental design includes
contamination of coupons by exposing them to a bulk solution of at least one contaminant.
Thereafter, the persistence of that contaminant on the coupons and/or the application of a
decontamination approach are to be investigated to determine both the propensity of the
contaminant to persist on the coupons and the effectiveness of decontamination approaches in
removing the contaminant from the coupon surface. The usefulness of results from such
experiments relies on the accuracy of the required contaminant measurements. In order to be
confident in these measurements, two important questions need to be answered about the
approach to contaminant measurement.
• When adsorbed to the coupon surface, how well can the contaminant be extracted from
that surface?
• When a coupon has been exposed to a bulk solution at a given concentration, how much
of the contaminant is adsorbed to the coupon surface?
To answer these two questions, two method verification steps make up the first two steps of the
experimental design. First, the surface contamination extraction method is to be validated.
Second, the coupon surface contamination method is to be validated. If the contaminant is able
to be extracted from the surface of the coupon and it is able to be deposited onto the coupon
surface from the bulk solution, the experimental design can proceed to experiments that seek
information about contaminant persistence and, if the contaminant is persistent, the effectiveness
of various decontamination approaches.
A 1.2.1 Method Verification Step 1: Surface Contamination Extraction
The purpose of this step is to determine whether it is possible to extract the contaminant if
adsorbed to a pipe material surface. The surface contamination extraction method verification
includes the extraction of the entire coupon by placing each coupon in a test tube (BD Falcon
56
-------
#352045, BD Biosciences, San Jose, CA) filled with an appropriate extraction solution,
depending on the characteristics of the contaminant. If the contaminant requires an organic
solvent, a glass test tube may need to be used (Fisher #14-962-26H Fisherbrand, Pittsburgh, PA).
After inserting the coupon, the test tube is to be sealed with a cap and sonicated for 10 minutes,
solvent decanted and replaced with fresh solvent, and then sonicated for another 10 minutes.
The decanted solvents are to be combined. For pipe material coupons with a significant amount
of corrosion or other loose particles, the contaminant may be bound to that component of the
pipe that could separate from the coupon during sonication. The coupon is to be removed and
the resulting solution is to be centrifuged and supernatant solution collected for analysis. For
organic chemicals, the extraction solution is to be an organic solvent that may be concentrated
using nitrogen evaporation prior to analysis using a gas chromatographic mass spectrometer
(GC-MS) or other appropriate detection device. For biological organisms, ATSM Type I water
should be the extraction solvent and membrane filtration should be used to measure the
biological organisms via plate enumeration.
The verification requires 20 coupons of the applicable material type with a biofilm developed as
described in Section ALL These coupons are to be removed from the biofilm growth container
after the four day long biofilm development (in uncontaminated water) and allowed to air dry
until water droplets are not visible on the surface, but the surface is still damp. This drying step
is to ensure that the contaminant is added to the coupon surface and not the water remaining on
the coupon surface. The required drying time is to be documented and used for other surface
contamination extraction and measurement verifications. For this phase of the evaluation, each
coupon (including blanks) is to be cut approximately in half with scissors and five drops of stock
solution applied directly to each smaller coupon (total volume of 15 jiL) using a micropipette
(Eppendorf Research Plus, Eppendorf International, Hauppauge, NY or equivalent)
approximately 10 mm apart. If the contaminant is water soluble, the stock solution should be
prepared in ASTM Type I water (for contaminants insoluble in water, an appropriate solvent is to
be used). The concentration of the stock solution depends on the quantitation limit of the
analytical technique that is available for the contaminant. For example, if the quantitation limit
of the applicable analytical technique is 0.1 |ig/mL, and the final extraction solution is
concentrated to 10 mL, then the minimum amount of contaminant that would be removed and
measured from the coupon surface would be 1 jig in 0.075 mL; which corresponds to a
contaminant stock concentration (from which the drops originate) of 0.33 |ig/mL. Because
measuring the contaminant in this scenario would require a 100% recovery and the results would
still be at the quantitation limit, this scenario would not be preferable as measurements near the
detection limit are likely to be imprecise. Instead, the contaminant stock solutions are to be
prepared at concentrations 10, 50, and 500 times higher to provide data that indicates what
concentration range provides the best likelihood of precise measurements which corresponds
with precise extraction recoveries. Precise extraction recoveries allow for the determination of
any differences between experimental conditions (i.e., in this case, contaminant concentration).
Using a range of stock solution concentration also provides information about how the extraction
recovery varies with concentration. The concentration of the stock solution is to be confirmed
with the appropriate analytical method. The drops of contaminant stock solution are to be
applied to each coupon as shown in Figure 1.
57
-------
0
0
0
0
0
Figure 1. Schematic of drops of contaminant solution across coupon surface
Each concentration is to be applied to five coupons (for a total of 15 coupons). The coupon
should air dry until the drops are not visible on the surface. This drying step ensures that the
contaminant is on the surface of the coupon (and not still in a droplet of solution) prior to the
extraction procedure. The required drying time is to be documented and used for other surface
contamination extraction and measurement verifications. Five non-contaminated coupons should
also be measured to determine any possible interference. Table 1 gives an overview of the steps
included in the surface contamination extraction and measurement method verification.
Table 1. Surface Contamination Extraction Method Verification (Step 1)
Step
Description
, . Develop biofilm on 20 pipe material coupons (confirm with heterotrophic plate count)
and allow coupons to air dry
IB
Determine contaminant stock solution concentration required for detection with 100%
contaminant recovery (depends on quantitation limit of contaminant measurement
technique)
1C
Prepare contaminant stock solutions at 10, 50, and 500 times (*) the concentration
required for attaining detection limit with 100% recovery and confirm the concentration
1D.1
Leave five coupons unspiked for blank analyses
1D.2
Spike five drops of the 10* stock solution on five coupons and air dry
1D.3
Spike five drops of the 50* stock solution on five coupons and air dry
1D.4
Spike five drops of the 50Qx stock solution on five coupons and air dry
IE
Extract contaminant from all coupons and calculate recovery
The percent recovery (%R) should be calculated using the following equation
CR
%R = JL x 100
^o
where CR is the mass of contaminant (or number of organisms) recovered from the coupon
surface and C0is the mass of contaminant (or number of organisms) originally dispensed onto the
coupon surface. The percent recovery data is to be evaluated to determine if the extraction
recovery is adequate for obtaining useful contaminant persistence and decontamination data and
how the extraction recovery varies with the concentration level of the contaminant applied to the
coupons. Following evaluation of the data, it may be necessary to repeat experiments with
additional replicates to clarify the results.
58
-------
A 1.2.2 Method Verification Step 2: Surface Contamination
Step 2 involves validating a method to contaminate the surface of the pipe material coupons
in a way that simulates an actual intentional contamination of a water distribution system.
The surface contamination method to be validated incorporates:
• Preparing coupons with biofilm
• Exposing the coupons to contaminated water (100 mg/ liter (L) or 106 CFU/mL,
depending on contaminant) in the AR without flow (batch mode)
• Extraction of the contaminant from the coupon using the method validated in Step 1.
To begin the verification, 10 coupons are to be prepared with a biofilm. The coupons are to be
loaded in the AR. Then, contaminant is to be added to the AR so that the bulk solution becomes
contaminated to the above-stated concentration levels. During this time, the AR is to be
operating as described in Section Al.l, but the flow through the AR is to be stopped to increase
the contact time between the contaminated water and the coupons. Two hours following the
contamination of the water, the coupons are to be removed, rinsed twice with 25 mL of ASTM
Type I water, and then extracted and analyzed following the surface contamination extraction
and measurement method validated as described in Section Al .2.1. This rinse step is to ensure
that the contaminant is extracted from the surface of the coupon and is not just an artifact of
residual contamination solution on the surface of the coupon. It is possible that a slow adsorbing
contaminant would have to be exposed to the coupons for a longer time or that a higher
concentration contamination solution would need to be used. The bulk solution is to be sampled
at the start of the contamination time period, at the half-way point, and at the end and the
concentration of contaminant confirmed via the appropriate measurement technique to confirm
the availability of the contaminant for adsorption.
The extent of surface contamination is to be evaluated to determine whether the level of
contamination and precision of these results are adequate for obtaining useful contaminant
persistence and decontamination data. Following evaluation of the data, it may be necessary to
repeat experiments with additional replicates, increased contamination times, or increased
contamination solution concentrations to clarify the results. This verification may have to be
repeated for additional coupon material and/or contaminant combinations.
A1.3 Evaluation of Contaminant Persistence
This section describes the approach to evaluating the persistence of a contaminant on various
pipe coupon materials. Table 2 provides an overview of the persistence evaluation (PE). Once
validated that a contaminant can be extracted from the surface of a coupon and a pipe coupon
can be contaminated with contact with a bulk contaminant solution, the persistence of that
contaminant on the pipe surface can be evaluated. For each combination of coupon material and
contaminant, 20 coupons should be prepared with biofilm as described in Section Al.l.
59
-------
Table 2. Persistence Evaluation
PE Step
PE 1
PE2
PE3
PE4
PE5
PE6
PE7
Description
Develop biofilm (confirm with heterotrophic plate count) on 20 coupons;
remove two coupons as blanks
Stop flow through the AR, inject enough contaminant into the ARto
make the bulk concentration within the AR 100 mg/L of contaminant;
wait 2 hours (concentration and time could vary depending on results of
surface contamination verification)
Sample bulk contaminant solution at start, half-way point, and end of
contamination period and measure bulk water contaminant concentrations
Following 2 hour contamination period, remove three coupons as control
coupons; extract and determine residual surface contaminant
concentration
Stop AR rotation to simulate stopped flow. Replace bulk contamination
solution with uncontaminated water and remain at stopped flow for 24
hours; collect three coupons, extract and determine residual surface
contaminant concentration.
Restart the AR rotation and flow through the AR. Remove three coupons
at 4 hours, 1 day, 3 days, and 7 days after restart of AR rotation and flow;
extract and determine residual surface contaminant concentration
Calculate percent persistence for all coupons by comparing to control
coupons
Coupons
removed
(20 total)
2
0
0
3
3
12
0
Two coupons with biofilm should be collected as non-contaminated blanks and the rest of the
coupons contaminated with a bulk solution following the validated surface contamination
method as described in Section ALL Immediately following the coupon contamination step,
three coupons are to be removed to serve as control coupons. The amount of contaminant on the
surface of these control coupons is to be compared with the amount remaining on the coupons
that are left in the AR for various lengths of time following the removal of the control coupons.
Collectively, the coupons removed from the AR during this part of the evaluation are to be
referred to as the persistence evaluation (PE) coupons.
Thereafter, a stopped flow scenario is to be evaluated by stopping the rotation of the AR and
stopping the flow of water through the AR (after the contaminant water is replaced by
uncontaminated drinking water). This stopped flow scenario is to be held for 24 hours, which is
when three PE coupons are to be removed. After that 24 hour period, the flow of drinking water
and AR rotation should be resumed to normal operating conditions as described in
Section ALL Following the stopped flow scenario, sets of three PE coupons are to be collected
from the AR at four different time increments (4 hours, 1 day, 3 days, and 7 days) following the
resumption of flow. Following the removal of each of these sets of PE coupons, they are to be
extracted and the amount of contaminant on the coupon surfaces compared with the amount on
the control coupons collected just after the coupon contamination step. This comparison can be
made by calculating the percent persistence (%P) of the contaminant on the coupons as described
by the following equation:
o/0p =
Cr
. x 100
60
-------
where CPE is the mass of contaminant (or number of organisms) recovered from the PE coupon
surface and Ccis the average mass of contaminant (or number of organisms) originally measured
from the surfaces of the control coupon surfaces. The %P data should be evaluated to determine
whether the %P at the various time periods is adequate to consider evaluation using various
approaches to decontamination of contaminants that are persistent on pipe surfaces. It should be
noted that the evaluation of persistence needs to be performed separately for each combination of
contaminant and coupon material. In addition, the uncertainty of each of the individual
measurements required to calculate the %P (i.e., uncertainty in the analytical measurements
required to determine CPE and Cc) is to be used to propagate the uncertainty in the %P
calculation. The uncertainty is to be used to determine the adequacy of the %P in making
comparisons between the various time increments evaluated during the persistence evaluation.
Upon evaluation of the %P, additional replicates may need to be evaluated in order to attain low
enough relative uncertainties in order to determine significant differences.
A1.4 Evaluation of Decontamination Approaches
For those contaminant and pipe material combinations that are determined to be persistent, this
section describes the evaluation of two approaches to decontaminating pipe, flushing (F) and
hyperchlorination (HC). Table 3 provides an overview of the flushing evaluation and Table 4
provides an overview of the HC evaluation. However, the same general evaluation could be
performed for other decontamination approaches that alter the makeup of the available tap water.
As was the case for the persistence evaluation, a biofilm is to be grown on
Table 3. Evaluation of Flushing as a Decontamination Approach
Step
F 1
F2
F3
F4
F5
F6
F7
F8
F9
Description
Develop biofilm (confirm with heterotrophic plate count) on 20 coupons of
the same material; remove two coupons as blanks
Inject enough contaminant into the ARto make the bulk concentration
within the AR 100 mg/L of contaminant; wait 2 hours (concentration and
time could vary depending on results of surface contamination verification)
Sample bulk contaminant solution at start, half-way point, and end of
contamination time and measure bulk water contaminant concentrations
Following 2 hour contamination period, replace bulk contamination solution
with uncontaminated water and remove three coupons as contaminated
control coupons
Increase AR rotational velocity to 200 rpm (1 .64 ft/s) from original
velocity of 100 rpm (1 ft/s)
Remove three coupons at 2 hours, 4 hours, and 1 day following increase in
rotational velocity
Increase AR rotational velocity to 250 rpm from 200 rpm
Remove three coupons at 4 hours and 1 day following increase in rotational
velocity to 250 rpm (1 .91 ft/s)
Calculate percent persistence for all coupons by comparing with control
coupons
Coupons
removed
(20 total)
2
0
0
3
0
9
0
6
0
61
-------
20 coupons of the desired material as described in Section ALL Thereafter, two coupons are to
be collected as blanks and 18 coupons are to be contaminated using the validated surface
contamination method. Following contamination, three contaminated coupons are to be removed
to serve as the control coupons. The amount of contaminant on the surface of these control
coupons should be compared with the amount remaining on the coupons that are left in the AR
(operated under increased flow conditions to simulate flushing). These coupons are to be
referred to as the decontamination evaluation (DE) coupons.
Specifically, following coupon contamination, the AR inner cylinder rotation is to be raised from
100 rpm (1 ft/s) to 200 rpm (1.64 ft/s), which corresponds to a water velocity of 0.5 ms"1 (1.64
ft/s) in a 15.2 cm (6 in.) pipe3. This increased rotational speed is to be held for one day. Sets of
three DE coupons are to be collected from the AR at three different time increments (2 hour, 4
hours, and 1 day) following the coupon contamination. Then, the rotational speed is to be
increased again to 250 rpm (1.91 ft/s) and held for another day, with the collection of three DE
coupons after 4 hours and after 1 day of 250 rpm conditions. Following the removal of each set
of three DE coupons, the coupons are to be extracted and the amount of contaminant on the
coupon compared with the amount on the control coupons collected just after the surface
contamination step. This comparison is to be made by calculating the %P of the contaminant
originally on the coupons, as described in the previous section. As was the case for the
persistence evaluation, the evaluation of decontamination approaches needs to be performed
separately for each combination of contaminant and coupon material.
The evaluation of hyperchlorination as a decontamination approach is to be performed as shown
in Table 4. The evaluation is to start in a similar way as for the flushing evaluation. However,
instead of increasing the rotational velocity of the AR, the rotation of the AR is to be stopped and
the drinking water flow through the AR is to also be stopped to simulate a stopped flow scenario.
The free chlorine concentration is to then be increased first to 25 mg/L and then to 50 mg/L after
several increments of time after which DE coupons are to be collected from the AR. Note that
other chemical decontamination approaches could be evaluated in the same way as
hyperchlorination if that decontaminant was added in place of the increased free chlorine.
62
-------
Table 4. Evaluation of Hyperchlorination as a Decontamination Approach
Step
HC 1
HC2
HC3
HC4
HC5
HC6
HC7
HC8
HC9
Description
Develop biofilm (confirm with heterotrophic plate count) on 20 coupons of
same material; remove two coupons as blanks
Inject enough contaminant into the ARto make bulk concentration within
AR 100 mg/L of contaminant; wait 2 hours (contaminant concentration and
time could vary depending on results of surface contamination verification)
Sample bulk contaminant solution at start, half-way point, and end of
contamination time and measure bulk water contaminant concentrations
Following the 2 hour contamination period, remove three coupons as control
coupons; extract and determine residual surface contaminant concentration
Following 2 hour contamination period, stop flow through AR and stop
rotation of the AR; increase the free chlorine concentration to 25 mg/L from
original concentration of 1 mg/L
Remove three coupons at 2 hours, 4 hours, and 1 day following increase in
free chlorine concentration
Increase free chlorine concentration to 50 mg/L
Remove three coupons at 4 hours and 1 day following increase in free
chlorine concentration to 50 mg/L
Calculate %P for all coupons by comparing with control coupons
Coupons
removed
(20 total)
2
0
0
3
0
9
0
6
0
Sections A2-A10
Sections A2-A10 of the prospective QAPP will be very dependent on the selection of the
contaminant that is to be used for the testing of this experimental design. The section headings
are shown below:
• Sampling Methods
• Sample Handling and Custody
• Analytical Methods
• Quality Control
• Instrument/Equipment Testing, Inspection, and Maintenance
• Instrument/Equipment Calibration and Frequency
• Inspection/Acceptance of Supplies and Consumables
• Non-direct Measurements
• Data Management.
Therefore, these sections will need to be completed pending selection of a contaminant (or
contaminants) to be tested.
63
-------
REFERENCES
1. ASTM C 150-07 Standard, 2007, "Standard Specification for Portland Cement," ASTM
International, West Conshohocken, PA, www.astm.org.
2. American Water Works Association C104-03Standard, 2004, "Standard for Cement-
Mortar Lining for Ductile-Iron Pipe and Fittings for Water" Denver, CO,
www.awwa.org.
3. Szabo, J. G., E. W. Rice, and P. L. Bishop. 2006. Persistence of Klebsiellapneumoniae
on simulated biofilm in a model drinking water system. Environ. Sci. Technol. 40:4996-
5002.
4. Szabo, J. G., E. W. Rice, and P. L. Bishop. 2007. Persistence and Decontamination of
Bacillus atrophaeus subsp. globigii Spores on Corroded Iron in a Model Drinking Water
System. Applied and Environmental Microbiology, 73, 8, 2451-2457.
5. Szabo, J.G., Persistence and decontamination of surrogate radioisotopes in a model
drinking water distribution system, Water Research (2009),
doi:10.1016/j.watres.2009.08.012.
6. EPA Requirements for Quality Assurance Project Plans, EPA/240/B-01/003,
Washington, D.C., March 2001.
64
-------
United States
Environmental Protection
Agency
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT NO. G-35
Offce of Research and Development (8101R)
Washington, DC 20460
Off cial Business
Penalty for Private Use
$300
------- |