EPA/600/R-12/543 | November 2012 | www.epa.gov/ord
United States
Environmental Protection
Agency
Technology Evaluation Report
Ol Analytical, Inc.
Model 9210m (Alpha Version)
Online Total Organic Carbon
Monitor
Office of Research and Development
National Homeland Security Research Center
-------
Technology Evaluation Report
Ol Analytical, Inc.
Model 9210m (Alpha Version)
Online Total Organic Carbon Monitor
U.S. Environmental Protection Agency
Cincinnati, OH 45268
-------
Contents
List of Figures iii
List of Tables iii
Disclaimer iv
Acknowledgements v
Abbreviations/Acronyms vi
Executive Summary vii
1.0 Introduction 1
2.0 Technology Description 2
3.0 Experimental Details 3
3.1 Portable Pipe Loop and Experimental Conditions 3
3.2 Baseline Conditions 4
3.3 Portable Pipe Loop (PPL) Contaminant Injections 5
3.4 Discrete Sample Analysis 7
3.5 Contaminant Concentrations 8
3.6 Data Analysis 8
4.0 Quality Assurance/Quality Control 11
4.1 Reference Method 11
4.2 Instrument Calibration 11
4.3 Audits 12
5.0 Evaluation Results 15
5.1 Toxic Industrial Chemicals (TICs) in Drinking Water 15
5.2 Biological Contaminants (BCs) in Drinking Water 22
5.3 Discrete Sample Analyses 26
5.4 Additional Tests 30
5.5 Operational Characteristics 33
6.0 Performance Summary 35
6.1 9210m Response to Contaminant Injections 35
6.2 Accuracy of 9210m Measurements 35
6.3 Operational Characteristics 36
7.0 References 37
11
-------
List of Figures
Figure 2-1. OI 9210m 2
Figure 3-2. EPA's portable pipe loop 4
Figure 5-1. Change in 9210m Total Organic Carbon (TOC) in response to injections of
colchicine 16
Figure 5-2. Change in 9210m Total Organic Carbon (TOC) in response to injections of
Bacillus globigii 22
List of Tables
Table 3-1 Source and Purity of Contaminants 6
Table 3-2. Contaminant List 9
Table 4-1. Summary of Total Organic Carbon (TOC) Reference Method 11
Table 4-2. Performance Evaluation Audit Results 12
Table 5-1. Change in 9210m Total Organic Carbon (TOC) from Injections of Toxic
Industrial Chemicals (TICs) into Chlorinated Water 17
Table 5-2. Change in 9210m Total Organic Carbon (TOC) from Injections of Toxic
Industrial Chemicals (TICs) into Chloraminated Water 18
Table 5-3. 9210m Accuracy for Total Organic Carbon (TOC) from Injections of Toxic
Industrial Chemicals (TICs) in Chlorinated Water 20
Table 5-4. 9210m Accuracy for Total Organic Carbon (TOC) from Injections of Toxic
Industrial Chemicals (TICs) in Chloraminated Water 21
Table 5-5. Change in 9210m Total Organic Carbon (TOC) from Injections of Biological
Contaminants (BCs) into Chlorinated Water 23
Table 5-6. Change in 9210m Total Organic Carbon (TOC) from Injections of BCs into
Chloraminated Water 24
Table 5-7. 9210m Accuracy for Total Organic Carbon (TOC) from Injections of
Biological Contaminants (BCs) in Chlorinated Water 25
Table 5-8. 9210m Accuracy for Total Organic Carbon (TOC) from Injections of
Biological Contaminants (BCs) in Chloraminated Water 26
Table 5-9. Change in 9210m Total Organic Carbon (TOC) for Discrete Sample Analyses
in Chlorinated Water 27
Table 5-10. Change in 9210m Total Organic Carbon (TOC) for Discrete Sample
Analyses in Chloraminated Water 28
Table 5-11. 9210m Accuracy for Total Organic Carbon (TOC) from Discrete Sample
Analyses of Contaminant Injections in Chlorinated Water 30
Table 5-12. 9210m Accuracy for Total Organic Carbon (TOC) from Discrete Sample
Analyses of Contaminant Injections in Chloraminated Water 30
Table 5-13. Change in 9210m Total Organic Carbon (TOC) with Elevated TOC
Concentrations 31
Table 5-14. Change in 9210m Total Organic Carbon (TOC) with Elevated Ionic Strength
32
Table 5-15. Change in 9210m Total Organic Carbon (TOC) with Varying
Monochloramine Concentrations 32
iii
-------
Disclaimer
The U.S. Environmental Protection Agency (EPA), through its Office of Research and
Development's National Homeland Security Research Center, funded and managed this
technology evaluation through a blanket purchase agreement under General Services
Administration contract number GS23F0011L-3 with Battelle. This report has been peer
and administratively reviewed and has been approved for publication as an EPA
document, but does not necessarily reflect the Agency's views. Mention of trade names
or commercial products does not constitute official EPA endorsement or recommendation
for use of a specific product. EPA does not endorse the purchase or sale of any
commercial products or services.
The evaluation reported herein was conducted by Battelle as part of the TTEP program.
Information on NHSRC and TTEP can be found at http://www.epa.gov/nhsrc/ttep.html.
Questions concerning this document should be addressed to:
Shannon D. Serre, Ph.D.
National Homeland Security Research Center
Office of Research and Development
U.S. Environmental Protection Agency
109 T.W. Alexander Drive, E343-01
Research Triangle Park, NC 27711
919-541-3817
IV
-------
Acknowledgements
Contributions of the following individuals and organizations to the development of this
document are gratefully acknowledged.
United States Environmental Protection Agency (EPA)
Jeff Szabo
Matthew Magnuson
Mike Henrie
New York City Department of Environmental Protection
Yves Mikol
Battelle (contract number GS23F0011L-3)
-------
Abbreviations/Acronyms
BC
cm
%D
EPA
L
m
m/s
mg/L
NDIR
NHSRC
PBS
PE
PPL
ppm
QA
QC
QMP
SCADA
T&E
TIC
TOC
ISA
TTEP
%U
UV
UVS
biological contaminant
centimeter
the percent difference between the 9210m and the reference measurements
U.S. Environmental Protection Agency
liter
meter
meter per second
milligram per liter
non-dispersive infrared
National Homeland Security Research Center
phosphate buffered saline
performance evaluation
portable pipe loop
parts per million
quality assurance
quality control
quality management plan
supervisory control and data acquisition
testing and evaluation
toxic industrial chemical
total organic carbon
technical systems audit
Technology Testing and Evaluation Program
uncertainty
ultraviolet
ultraviolet spectroscopy
microgram per liter
VI
-------
Executive Summary
The U.S. Environmental Protection Agency's (EPA's) National Homeland Security
Research Center (NHSRC) Technology Testing and Evaluation Program is protecting
human health and the environment from adverse impacts resulting from acts of terror
testing homeland security technologies. The primary objective of this series of
evaluations was to determine the response of the TOC analyzers and ultraviolet
spectrometers upon the introduction of contaminants such as toxic industrial chemicals
(TICs) and biological contaminants in drinking water. Under the program, the
performance of several online total organic carbon (TOC) analyzers and ultraviolet
spectrometers (UVS) was evaluated. In this study, the OI Analytical, Inc. (College
Station, TX), Model 9210m Online TOC Monitor (hereafter referred to as the 9210m)
was tested in conjunction with EPA's portable pipe loop (PPL), which simulates a
drinking water distribution system. A total of 14 different contaminants were injected
into both chlorinated and chloraminated water. Relatively low contaminant concentration
levels (0.01-10 milligrams/liter (mg/L) were purposefully chosen for this evaluation of
online TOC analyzers because many of the contaminants posed health risks at the low
drinking water concentrations tested.
For the purposes of this 9210m study, one criterion for a "response" (or change) was
subjectively determined to be a post-injection change in TOC measurement of at least
three times the standard deviation of the baseline TOC measurement for the thirty
minutes prior to and after the contaminant injection. Measurements of TOC were made
daily using a laboratory reference method and compared with the results from the 9210m
to evaluate the accuracy of the 9210m measurements. Deployment and operational
factors were also documented and reported. The development schedule of the 9210m did
not allow for a final production unit to be ready for the TTEP testing, but upon request
from EPA, OI Analytical Inc. submitted an "Alpha" development unit for testing.
Therefore, this testing could reveal performance issues in the "Alpha" unit that would be
corrected in the final production units.
9210m Response to Contaminant Injections
Contaminant injections were performed for aldicarb, carbofuran, colchicine, diesel fuel,
disulfoton, mevinphos, nicotine, potassium cyanide, sodium fluoroacetate, Bacillus
globigii, Bacillus thuringiensis, Chlorella, ovalbumin, and ricin. The contaminant
injection solutions were prepared within 24 hours (most within 8 hours) in the same water
that was within the PPL. Since this water contained disinfectants it could cause
degradation or transformation of the injected contaminants prior to injection. The 9210m
measured a change in TOC in response to aldicarb, carbofuran, colchicine, mevinphos,
nicotine, and sodium fluoroacetate at a minimum concentration of 1 mg/L in both
chlorinated and chloraminated water. The 9210m measured a change in TOC in response
to potassium cyanide only at 10 mg/L for both chlorinated and chloraminated water. The
9210m did not measure a change in TOC in response to injections of Bacillus globigii,
vii
-------
Bacillus thuringiensis, or Chlorella at any injected concentration. Ovalbumin injected at
1 mg/L and 10 mg/L in both chlorinated and chloraminated water was measured by the
9210m as an increase in TOC.
Disulfoton, diesel fuel, and ricin, were analyzed only as discrete samples and carbofuran
was analyzed both using the continuous PPL and the discrete mode. For disulfoton, the
9210m measured a change in TOC in response to 1 mg/L in chloraminated water, but not
in chlorinated water. While diesel fuel was not soluble, it was added to water and
analyzed using the 9210m, but the results were inconsistent and difficult to interpret. For
ricin, the 9210m measured a change in TOC in response to 10 mg/L in chlorinated and
chloraminated water. In addition to these measurements, limited experiments were
performed to examine the effect of elevated TOC, ionic strength, and monochloramine
concentrations on the 9210m TOC measurements.
Accuracy of 9210m Measurements
The TOC measurements from the 9210m were compared with those from a commonly
used reference method during all of the contaminant injections performed during the
evaluation. These comparisons should be interpreted with the awareness the different
TOC instruments and oxidation methods can respond differently to various contaminants.
Overall, the average absolute value of the %Ds (the percent difference between the
9210m and the reference measurements) for all the comparisons across the evaluation
was 6% plus or minus (±) a standard deviation (SD) of 5%. For TICs in chlorinated and
chloraminated water, the %Ds averaged 5% ± 6% and 6% ± 4%, respectively. For the
biological contaminants (BC), the %D averaged 16% ± 8% and 13% ± 7%, respectively
for chlorinated and chloraminated water. For each individual comparison, the
experimental error was propagated using the uncertainty of both measurements. During
the TIC injections, the differences between the 9210m results and the reference method
results were mostly not anomalous, however, during the BC injections, the differences
were often larger than the experimental uncertainty.
Operational Characteristics
During the evaluation of the 9210m, general operational characteristics were observed.
Installation and operation of the 9210m was straight forward and clearly articulated by
the vendor during a one day visit. Operation of the 9210m instrument control software
was simple and intuitive. The 9210m used an aqueous phosphoric acid reagent to
facilitate the electrochemical oxidation and soda lime as a constituent of a gas filter. The
phosphoric acid reagent was changed once during the evaluation when the acid inlet was
no longer submerged in the reagent reservoir. The soda lime was changed on the
recommendation of the vendor only after a change in the response of the 9210m, relative
to the reference measurement, was observed. The 9210m software produced an overflow
error on approximately 10 of the 56 days of testing. When this happened, it required a
restart of the software and re-initialization of data acquisition. During one set of
colchicine injections, this error caused a loss of data. The vendor attributed this error to
viii
-------
the fact that the instrument tested was an "Alpha" development unit. According to OI
Analytical, this problem has been corrected in a newer release of the data collection
software. Additionally, the 9210m is designed for connection to a supervisory control
and data acquisition (SCADA) system rather than using the computer interface as the
primary means of data collection.
IX
-------
1.0 Introduction
The U.S. Environmental Protection Agency's (EPA's) National Homeland Security
Research Center (NHSRC) is protecting human health and the environment from adverse
impacts resulting from intentional acts of terror. With an emphasis on decontamination
and consequence management, water infrastructure protection, and threat and
consequence assessment, NHSRC is developing tools that will detect the intentional
introduction of chemical or biological contaminants in buildings or water systems, and
will provide the information needed for the containment of these contaminants, the
decontamination of buildings and/or water systems, and the disposal of material resulting
from clean-ups.
NHSRC's Technology Testing and Evaluation Program (TTEP) works in partnership
with recognized testing organizations, with stakeholder groups consisting of buyers and
users of homeland security technologies, and with the participation of individual
technology developers in carrying out performance testing on such technologies. The
program evaluates the performance of innovative homeland security technologies by
developing evaluation plans that are responsive to the needs of stakeholders, by
conducting tests, by collecting and analyzing data, and by preparing peer-reviewed
reports. All evaluations are conducted in accordance with rigorous quality assurance
(QA) protocols to ensure that data of known and high quality are generated and that the
results are defensible. TTEP provides high-quality information that is useful to decision
makers in purchasing or applying the evaluated technologies. It provides potential users
with unbiased, third-party information that can supplement vendor-provided information.
Stakeholder involvement ensures that user needs and perspectives are incorporated into
the evaluation design so that useful performance information is produced for each of the
evaluated technologies.
Under TTEP, the performance of the OI Analytical Inc. 9210m Online TOC Monitor
(hereafter referred to as the 9210m) was evaluated. The development schedule of the
9210m did not allow for a final production unit to be ready for the TTEP testing, but
upon request from EPA, OI Analytical Inc. submitted an "Alpha" development unit for
testing. Therefore, we expected testing to reveal performance issues in the "Alpha" unit
that would be corrected in the final production units.
The primary objective of this evaluation was to determine the ability of the 9210m to
detect changes in the TOC concentration in response to the introduction of contaminants
in drinking water. Two other objectives were to evaluate the accuracy of the TOC
measurement and to document deployment and operational characteristics. This
evaluation was conducted according to a peer-reviewed test QA plan that was developed
according to the requirements of the TTEP's quality management plan (QMP) and
amendments'^ to the QMP.
-------
2.0 Technology Description
This report provides results for the evaluation of the "Alpha" development unit of the
9210m. Following is a description of the 9210m based on information provided by the
vendor.
Figure 2-1 shows the 9210m connected to EPA's
portable pipe loop (PPL). The 9210m was mounted
on a freestanding base which supported the
instrument and provided space for the phosphoric acid
reagent required for operation. The 9210m enclosure
was 30 cm wide, 25 cm deep, and 48 cm tall. When
mounted on the stand, the dimensions were 58 cm
wide, 30 cm deep, and approximately 1.8 m tall.
The 9210m measures the non-purgeable organic
carbon in a liquid sample and can operate either in an
unattended stand-alone mode with continuous
analysis of a process stream or a discrete sample
mode. The 9210m draws in sample, removes
purgeable carbon from the sample, oxidizes the
available carbon to CO2 via electrochemical oxidation
and detects the generated CC>2 using a non-dispersive
infrared (NDIR) detector. Process gas required for
operation of the 9210m is generated by an integrated
process gas module.
Figure 2-1. OI 9210m
The 9210m was operated using the RS-232
connection for this work, but a keypad on the front
face of the instrument can also be used to control the instrument. Data collection,
visualization, and storage can be performed using the supplied software, but this software
does not have to be running while the instrument is operating. Data files are stored as
comma delimited (.csv) files while the software is operational and can also be
downloaded from the instrument.
OI Analytical plans to sell the final production units of the 9210m for less than $10,000
and the estimated yearly non-labor operation and maintenance costs, which includes only
the cost of replacing the phosphoric acid reagent and soda lime, are estimated to be less
than $1,200 (vendor estimate).
-------
3.0 Experimental Details
The primary objective of the series of TTEP evaluations was to determine the capability
of online TOC analyzers and UVSs to measure changes in TOC level due to the
introduction of contaminants in drinking water. In all four technologies, two online TOC
analyzers and two UVSs, were evaluated. Two drinking water matrices were used for all
of the testing conducted in this evaluation: (1) finished drinking water from Columbus,
OH (chlorinated and filtered surface water), and (2) water prepared to simulate water
from a utility that uses chloramination as its primary means of disinfection. Eleven
contaminants were injected using the EPA's portable pipe loop (PPL) and four
contaminants were analyzed using a discrete analysis approach.
This OI 9210m evaluation took place between June 3, 2009 and September 24, 2009. QA
oversight of this evaluation was provided by the contractor and EPA. The contractor QA
staff conducted a technical systems audit (TSA) and an audit of data quality.
3.1 Portable Pipe Loop and Experimental Conditions
This evaluation was conducted using EPA's PPL, which is shown in Figure 3-1 and
consists of: (1) an equipment rack that contains a 78 liter (L) stainless steel mixing tank, a
recirculating pump, a peristaltic pump, and three contaminant injection ports and (2) a
piping rack that contains approximately 29 meters (m) of 7.6 cm diameter stainless steel
pipe (316L grade). The two racks were connected for use during this evaluation. All
four TOC analyzers and UVSs evaluated in the series of TTEP evaluations were
connected to the PPL by one of the eight separate sample ports with 6.35 millimeter
(mm) (or one quarter inch) inner diameter tubing.
When evaluating several technologies simultaneously, an adequate flow of water must be
maintained to supply each of the technologies. The 9210m drew in water from a flow
through loop consisting of a small PVC pipe receiving water from the PPL. The 9210m
drew in a 2.5 milliliter sample from the flowing water every 7 minutes and would
generate about 100 ml of waste per hour.
The PPL flow was controlled by the variable flow recirculating pump which allowed the
operator to set flow rates from 44 liters/minute (L/min) to 440 L/min in the PPL. For
testing, the PPL contained approximately 250 L of water (including the mixing tank and
pipe) with a flow rate of approximately 88 L/min (linear velocity of 0.33 meters/second
(m/sec)). Because of the addition of reagents to the water, the water sampled by the
9210m was discharged to a waste container after analysis.
-------
Figure 3-2. EPA's portable pipe loop
3.2 Baseline Conditions
Prior to the start of daily testing, the PPL was filled with drinking water using a hose
(15.9 mm or 5/8" ID) that connected the laboratory water supply to the mixing tank of the
PPL using a hose-thread to sanitary-fitting coupler. During the chlorinated water testing,
this water was used with no alterations after the free chlorine level was measured using
CJ~\
U.S EPA Method 330.5.l; Over the course of the evaluation, free chlorine concentrations
in the chlorinated water ranged from 1.0 to 1.6 milligrams/liter (mg/L) with an average of
1.3 mg/L. In addition, the pH of the water was always between 7.4 and 8.1.
The chloraminated test water matrix was prepared by mixing chlorine and ammonia in
the proper ratio to yield approximately 2 mg/L monochloramine, following an EPA
Testing & Evaluation (T&E) Facility standard operating procedure for preparation of
chloraminated water. The total chlorine concentration was measured and then chlorine
was added to the PPL to increase the total chlorine concentration to 2 mg/L. The total
chlorine concentration was then measured again to confirm the total chlorine
concentration was within 10% of 2 mg/L. Ammonia was then added to the PPL to form
monochloramine at a concentration of 2 mg/L in the PPL. Prior to the injection of a
contaminant, the monochloramine concentration was confirmed by Hach® Method
10200*^, not an EPA approved method for reporting monochloramine concentrations.
-------
Throughout the evaluation, the monochloramine concentrations ranged from 1.8 to 2.3
mg/L with an average of 2.0 mg/L.
Once the applicable chlorine measurement had been completed, a 30 minute baseline
measurement was conducted with the 9210m. During this baseline measurement period,
a reference sample was collected from an unoccupied PPL sampling port. The reference
sample was collected by flushing approximately 500 mL of water from the sampling port
into a waste container and then collecting two separate 40 mL vials of sample. The
reference samples were immediately preserved by acidification with phosphoric acid.
Reference samples were analyzed daily on the TOC reference instrument which was
housed in the same laboratory as the PPL. The reference TOC results were used to
evaluate the accuracy of the TOC measurements from the 9210m.
3.3 Portable Pipe Loop (PPL) Contaminant Injections
The toxic industrial chemicals ( TICs) and biological contaminants (BCs) were injected
into the PPL as concentrated 250 mL solutions were prepared within 8 hours of injection
(with the exception of the discrete samples which were prepared the day before) using the
same water used within the PPL, either chlorinated or chloraminated. The preparation of
contaminant injection solutions in water containing disinfectants could cause degradation
or transformation of the contaminant. For example, it is possible that the interaction of
the cyanide ion with chlorinated water could have formed cyanogen chloride which could
have continued to breakdown to the cyanate ion. These reactions are dependent on the
water quality of the dissolution water so the results presented here should be interpreted
carefully and not extrapolated too broadly. However, the experimental plan was intended
to simulate an actual contamination event during which the use of tap water as the
dissolution solvent would be expected. In addition, degradation of organic carbon in
drinking water is not likely so for the 9210m, the presence of a contaminant could still be
measured. Several surrogates were used in testing which included, Bacillus globigii and
Bacillus thuringiensis (surrogates for Bacillus anthracis), Chlorella (surrogate for
Cryptosporidium), and Ovalbumin (surrogate for botulinum toxin and ricin).
Table 3-1 shows the sources of each TIC and toxin contaminants as well as their purity.
The purity of the TICs varied substantially from 89% to 99%. Information about the
content of the impurities for each contaminant was not available so the impurities might
have contained organic carbon. In addition, aldicarb, carbofuran, and disulfoton were
difficult to get into solution and required gentle heating to encourage them into solution.
Heating these solutions could have favored transformations preventing the stated
contaminant (and instead transformational product) from being injected into the PPL.
Originally, the two Bacillus species were grown in nutrient broth while the Chlorella was
grown in Bold 1NV Medium ^\ The final BC cultures were pelleted by centrifugation
and washed in PBS three times. The washed pellet cake was then resuspended in
phosphate buffered saline (PBS). The BC solution was enumerated by culturing and
counting the colonies and injection solutions were prepared by diluting the BC to the
appropriate concentrations in PBS. The concentration of each injection solution was such
-------
that injection of 250 mL of the solution into 250 L of water in the PPL gave the desired
steady-state concentration in the PPL.
Table 3-1 Source and Purity of Contaminants
Contaminant
Aldicarb
Carbofuran
Colchicine
Diesel
Disulfoton
Mevinphos
Nicotine
Ovalbumin
Potassium Cyanide
Ricin
Sodium Fluoroacetate
Sodium Fluoroacetate
Supplier
Ultra Scientific (North Kingston, RI)
Sigma-Aldrich (St. Louis, MO)
Sigma-Aldrich (St. Louis, MO)
Marathon (Columbus, OH)
Chem Service (West Chester, PA)
Ultra Scientific (North Kingston, RI)
Acros Organics (Geel, Belgium)
MP Biomedicals (Solon, OH)
Sigma-Aldrich (St. Louis, MO)
Vector Laboratories (Burlingame, CA)
Sigma-Aldrich (St. Louis, MO)
Pfaltz & Bauer (Waterbury, CT)
Purity
99%
98%
97%
Retail-grade (from pump)
98.7%
89.4%
98%
98%
96%
5 mg/mL (in phosphate buffered
sodium azide)
99%
95%
Upon establishing a steady response in the PPL and collecting a minimum of 30 minutes
of baseline data, a series of single contaminant injections was made into the circulating
water of the PPL. With each injection, the concentration of the contaminant at any point
within the PPL changed until a steady-state concentration for the contaminant was
reached. As the contaminant stock solution was introduced into the intake side of the
recirculating pump of the PPL, the initial contaminant "slug" made a first pass by the
9210m intake. The contaminant concentration within the slug was higher than the
eventual steady-state contaminant concentration within the PPL. As the contaminant slug
flowed throughout the PPL, it entered the mixing tank, became greatly diluted, and then
continued to recirculate until a steady-state concentration was reached within
approximately 10 minutes. This evaluation focused on the steady-state concentration
reached in the PPL after mixing. The time scale for mixing of the PPL in the context of
the evaluation (approximately 10 minutes) and the frequency of the 9210m TOC
measurements (approximately 7 minutes) prevented evaluation of the 9210m with respect
to the first pass of a contaminant through the PPL. However, in an operational setting the
9210m would be able to measure continuous changes in TOC as long as the changes
being measured lasted for at least 7 minutes.
Starting with the lowest concentration level for each contaminant, the contaminant
injection solution was pumped into the circulating water of the PPL at a rate that made
the concentration of the contaminant 10 times the eventual steady-state concentration in
the water moving past the 9210m. This injection lasted for 15-20 seconds, which at a
flow rate of 88 L/m (linear velocity of 0.33 m/s) corresponds to approximately a 4.5 m
long (-25 L) slug of injected contaminant at the desired concentration.
Following a contaminant injection and subsequent mixing, a steady-state TOC
concentration was reached in the PPL approximately 10 minutes after the contaminant
injection. To ensure that the contaminant concentration had reached steady-state, a 20
minute stabilization period was allowed after a contaminant injection. After this 20
-------
minute stabilization period, 30 minutes of data from each instrument were collected at the
post-injection steady-state concentration. In addition to being used to determine the
steady-state response for each instrument, these 30 minutes of data were also used as the
baseline for the next contaminant injection. The next higher concentration level of
contaminant was introduced using the identical procedure. Therefore, a minimum of 50
minutes of time passed between contaminant injections, which resulted in approximately
a volume change of 4400 liters. One reference sample was collected when the PPL
reached the steady-state concentration following each contaminant injection. A duplicate
sample was taken for one of the reference samples during each day of testing.
Each set of contaminant injections with increasing concentration levels (some
contaminants had three concentration levels, some had four concentration levels)
represented one replicate. Three replicate sets of injections were made for each
contaminant. Between the replicate sets of injections, the system was exchanged at least
five times with contaminant-free water (from the laboratory supply). As this
uncontaminated water filled the PPL, the online measurements returned to the original
baseline. Once five water exchanges had been completed (approximately 30 minutes of
water exchange) and the response from each technology steadied so it deviated from the
average by less than 10% over 30 minutes, testing proceeded with the next replicate set of
injections for that contaminant.
3.4 Discrete Sample Analysis
Three contaminants (diesel fuel, disulfoton, and ricin) were analyzed by the 9210m as
discrete samples rather than injections into the PPL. This experimental approach was
used for ricin must be contained within a biosafety hood preventing use of the PPL for
the ricin injections. Disulfoton was added to the experimental plan late in the testing
through the same test/QA plan amendment as the ricin, so was analyzed in the same
fashion. Diesel fuel was analyzed discretely out of concern of contaminating the PPL
with diesel fuel. Carbofuran was analyzed discretely so that one contaminant would be
analyzed using both experimental approaches.
For the discrete samples, 100 mL containers of chlorinated and chloraminated water each
contaminated at the desired concentration as well as controls with uncontaminated
chlorinated or chloraminated water were prepared. The inlet line for the 9210m was
placed in a container of uncontaminated water and the 9210m sampled for a minimum of
32 minutes (to ensure a minimum of four analyses cycles). The inlet line was then
moved to the first concentration and the process repeated until the highest concentration
level for the contaminant was reached. Three containers of each concentration were
analyzed as discrete samples.
The 0.01 mg/L concentration was not evaluated as this concentration was below the
sensitivity specification of the 9210m. Due to limited solubility, disulfoton and
carbofuran were not analyzed at 10 mg/L as discrete samples. Diesel fuel was not soluble
in water and separated upon mixing with water, but discrete sample analyses at 10 mg/L
were still performed.
-------
3.5 Contaminant Concentrations
Table 3-2 gives the injected contaminants and their corresponding concentrations. As
described in the test/QA plan' TIC injection concentrations were selected based on
previous testing performed at EPA's Testing and Evaluation (T&E) Facility*^. BC
injection concentrations were based on relevant biological data (e.g., infective dose)*-6-* as
well as concentrations recommended by TTEP water security stakeholders. The 0.01
mg/L contaminant solutions were injected into the PPL during the evaluation, but the
results were not reported for the 9210m because that concentration level is lower than
what the 9210m specifications state as a detectable concentration.
3.6 Data Analysis
3.6.1 Response
During the evaluation, the 9210m made TOC measurements approximately once every 7
minutes. The baseline concentration of TOC prior to injection is defined as the average
concentration over the 30 minute baseline measurement period before each injection.
Baseline measurement periods following injections began 20 minutes after the previous
injection. For the purposes of this study, following a contaminant injection, the 9210m
was considered able to detect a change in TOC concentration if the absolute change
(ATOC) in TOC concentration was at least three times the standard deviation of the
baseline TOC concentration prior to the contaminant introduction and three times the
standard deviation of the post-injection TOC concentration measurements. When these
conditions were met, the 9210m was defined to exhibit a response to an injected
contaminant. To simplify the wording within this report, hereafter, this report will refer
to this scenario as the 9210m having "responded" to the injection of a contaminant. The
response threshold of three times the standard deviation was subjectively selected to
show a change from baseline as part of this evaluation. Depending on operational
parameters of a water system, a change in TOC that is at least three times the standard
deviation of the baseline, might be a concern in terms of a contamination event.
In addition to the injection of the 14 contaminants, control injections of both chlorinated
water and chloraminated water were performed to determine if the act of injecting water
into the PPL caused a response for the 9210m. Water for control injections was removed
from the PPL and then injected back into the PPL within 4 hours. Five such injections
resulted in an average response of 0.01 mg/L ± 0.06 mg/L. Therefore, in addition to the
requirement for a response described in this section, a second requirement for a response
was that the response exceeded the average and standard deviation of the blank
injections. Following this criteria (in addition to the change in TOC from the baseline
TOC), the threshold for detection was a change of 0.07 mg/L TOC.
-------
Table 3-2. Contaminant List
Type
Toxic
Industrial
Chemicals
(TICs)
Biological
Contaminants
(BCs)
Toxins
Controls
Agent
AldicarbT
Carbofuran
Colchicine
Diesel fuel
Disulfoton
Mevinphos
Nicotine
Potassium Cyanide
Sodium fluoroacetate
Bacillus globigii
(surrogate for Bacillus
anthracis)
Bacillus thuringiensis
(surrogate for Bacillus
anthracis)
Chlorella (surrogate
for Cryptosporidium)
Ovalbumin (surrogate
for botulinum toxin
and ricin)
Ricin
Water
PBS/nutrient broth
PBS/Bold 1NV
medium
Buffered sodium azide
Analysis
Method
PPL
PPL, discrete
PPL
Discrete
Discrete
PPL
PPL
PPL
PPL
PPL
PPL
PPL
PPL
Discrete
PPL
PPL
PPL
Discrete
Concentrations
0.01*, 0.1, 1 (mg/L)
0.01,0.1, 1, 10 (mg/L)
0.01,0.1, 1, 10 (mg/L)
0.1, 1, 10 (mg/L)
0.1, 1 (mg/L)
0.01,0.1, l(mg/L)
0.01,0.1, 1, 10 (mg/L)
0.01,0.1, 1, 10 (mg/L)
0.01,0.1, 1, 10 (mg/L)
103 ,104, 105, 106, 107
(spore s/L)
103 ,104, 105, 106, 107
(spore s/L)
103, 104, 105 (cells/L)
0.01,0.1, 1, 10 (mg/L)
0.1, 1, 10 (mg/L)
J
t
t
t
Medium
Water
Water
Water
Water
Water
Water
Water
Water
Water
PBS and
Nutrient
Broth
PBS and
Nutrient
Broth
PBS and
Bold 1NV
Medium
Water
Buffered
sodium
azide
Water
Water
Water
Water
PBS- phosphate buffered saline, PPL - portable pipe loop,
•f No 10 mg/L injections due to the response at 1 mg/L and prohibitive cost of 10 mg/L injections.
* 0.01 mg/L results not reported for the 9210m because below instrument specifications.
J Concentrations (or volumes for water injections) equivalent to those present in contaminant injections.
The magnitude of a change in TOC concentration was calculated and expressed as
ATOC. The signal change of the 9210m as ATOC for TOC was calculated using Equation
1:
ATOC = TOC - TOCbasetine (1)
where TOC is the average post-injection TOC concentration measured by the 9210m
(mg/L); and TOCbaseiine is the average baseline TOC concentration as determined by the
9210m (mg/L).
-------
3.6.2 Accuracy
Results from the evaluation of the 9210m were compared to the results obtained from
analysis of reference grab samples collected during the same period. The results for
each sample are expressed as the percent difference (%D) between the 9210m and the
reference measurements calculated from Equation 2:
Q/oD= * xlOO
l/2(OC + OCR)
(2)
where OCRis the concentration determined by the reference method (mg/L) and OC is the
average measurement from the 9210m when the reference grab sample was collected
(mg/L). Ideally, if the results from the 9210m and reference method measurement are the
same, there would be a %D of zero percent.
The combined experimental uncertainty (%U) of the %D was determined using the
method of propagation of errors and is defined by Equation 3:
%U = 100 X Vl6(OC + OCR)-4 X (OC^SOC + OC^SOCR) (3)
where SOCR and SQC are the standard deviations of OCR and OC, respectively.
10
-------
4.0 Quality Assurance/Quality Control
Quality assurance/quality control (QA/QC) procedures were performed in accordance
with the program QMP^ and the test/QA plan for this evaluation.
4.1 Reference Method
EPA Method 415.3(7) was used to provide a reference total organic carbon (TOC)
concentration in the PPL. The reference instrument used was a Teledyne-Tekmar
(Mason, OH) Fusion TOC Analyzer™. Reference samples were collected immediately
after contaminant injections as well as after the contaminant had become well-mixed in
the PPL (steady-state). The analysis method is described in Table 4-1.
Table 4-1. Summary of Total Organic Carbon (TOC) Reference Method
Instrument
Teledyne-
Tekmar Fusion
Analyzer
(Mason, OH)
Method
EPA
415.36
(Standard
Method
53 IOC)
Measurement
Principle
UV/persulfate
oxidation
Detection
Limit
0.0002 mg/L
Maximum
Holding Time
28 days with
acidification to
pH<2
4.2 Instrument Calibration
The vendor connected the 9210m to the PPL and verified the operation of the instrument.
Prior to testing, the 9210m was calibrated using calibration solutions provided by the
vendor. Calibration was performed with the low calibration point in deionized water and
the high calibration point a 25 ppm TOC solution. Following calibration, check
standards provided by the vendor were analyzed on the 9210m at 1 ppm, 5 ppm, and 10
ppm TOC to verify the calibration. The acceptable tolerance for these standards was
20%. All of the check standards analyzed were within 20% of the expected
concentration.
The reference instrument was calibrated prior to the start of the evaluation and once per
month throughout the approximately three month duration of testing. Calibration check
standards were analyzed with every batch of samples analyzed using the TOC reference
instrument to ensure that the reference instrument calibration had not drifted. With the
exception of one day out of 56 days of analysis that occurred during that three month
time period, all calibration check standards analysis results were within the required
tolerance throughout the evaluation. Samples from that day were reanalyzed the next
day, and the calibration check standards for the reanalysis were within the acceptable
tolerance.
11
-------
4.3 Audits
4.3.1 Performance Evaluation (PE) Audit
A performance evaluation (PE) audit was conducted to assess the accuracy of the TOC
reference method. A PE sample containing 5 mg/L organic carbon as potassium
hydrogen phthalate was obtained (Pharmaceutical Resource Associates [Environmental
Resource Associates], Arvada, CO) and analyzed. Accuracy of the TOC measurement
was expressed in terms of the percent error (%E), as calculated from the following
equation:
(3)
CR
where CRwas the standard or reference concentration of the PE sample and Jis the
measurement obtained using the reference method. Ideally, if the reference value and the
measured value are the same, there would be a percent error of zero percent. Table 4-2
shows that the results of the PE audit was below the maximum allowed %E for TOC.
Table 4-2. Performance Evaluation Audit Results
Reference Sample
Total Organic Carbon (TOC)
(Pharmaceutical Resource
Associates, Arvada, CO)
Expected Result
5. 00 mg/L
Reference
Method
Result
5. 30 mg/L
%E
(% error)
6.0
Maximum
Allowed
(%E)
20
4.3.2 Technical Systems Audit (ISA)
The contractor QA manager conducted a technical systems audit (TSA) at the Columbus,
Ohio testing location to ensure that the evaluation was performed in accordance with the
test/QA plan and the TTEP quality management plan (QMP)^. As part of the audit, the
contractor QA manager reviewed the reference sampling and analysis methods used,
compared actual evaluation procedures with those specified in the test/QA plan, and
reviewed data acquisition and handling procedures. No significant adverse findings were
noted in this audit. The records concerning the TSA are permanently stored with the
contractor QA manager.
4.3.3 Amendments and Deviations
One amendment was made to the test/QA plan for this evaluation. To accommodate the
latest needs of EPA's homeland security mission the amendment changed the list of
contaminants to be tested. The amendment removed cesium, as well as the chemical
12
-------
warfare agents VX, soman, and sarin from the contaminant list and added ricin,
disulfoton, mevinphos, and sodium fluoroacetate to the list of injected contaminants. The
amendment stipulated that sodium fluoroacetate and mevinphos should be evaluated in
the PPL and ricin and disulfoton as discrete samples and also changed the tests with
diesel fuel from using the PPL to testing diesel as discrete samples.
Throughout the course of testing, there were a few instances of slight deviation from the
test/QA plan.
• The PE sample for the reference method was not analyzed until after the first tests
were conducted rather than before testing began. The first attempt at the PE audit
was not successful due to the use of PE audit samples that unknowingly contained
chemical constituents that interfered with only the reference TOC instrument.
However, two TOC standards (one provided by the vendor) were measured
accurately by the reference method repeatedly during preparation for the
evaluation. Therefore, the evaluation staff had a high degree of confidence in the
accuracy of the reference method/instrument. Instead of holding up the
evaluation, testing proceeded while obtaining interferent-free PE audit samples.
• An alternate test method for monochloramine was used. Rather than using the
difference between total and free chlorine (EPA method 330.5^), a test for
monochloramine (Hach Method 10200^) was used to determine the
monochloramine level in the water used for the chloraminated tests.
• Injections at 0.01 mg/L were included in the test matrix in addition to the levels
specified in the test/QA plan (0.1, 1, and 10 mg/L) for the TICs. Some of the
participating technologies already tested by TTEP in this series of technology
evaluations were more sensitive than anticipated to the 0.1 mg/L injections. The
injections at 0.01 mg/L were added to better understand the performance of the
analyzers at the low end of their measurement range. In addition, because the 1
mg/L injections of aldicarb and mevinphos were detected by all the technologies
and the injection of 10 mg/L would have been extremely expensive, they were not
performed.
• For the elevated TOC component of the testing, the TOC was elevated by
approximately 1 mg/L rather than 2 mg/L because of the change in background
TOC of the source water on the day of testing.
• To identify detectable levels, concentrations of the BCs were increased to include
samples at 106 and 107 organisms/L.
• Percent difference was used to compare the reference method with the 9210m
instead of percent error.
13
-------
4.3.4 Data Quality Audit
At least 10% of the data acquired during the evaluation were audited. The contractor QA
manager traced the data from the initial acquisition, through reduction and statistical
analysis, to final reporting, to ensure the integrity of the reported results. All calculations
performed on the audited data were checked.
14
-------
5.0 Evaluation Results
This section presents the results of the evaluation including the ability of the 9210m to
measure changes in TOC concentrations in response to the injection of TICs and BCs in
drinking water. Also given are the accuracy of the 9210m TOC measurements and the
operational characteristics of the 9210m that were observed during the evaluation.
5.1 Toxic Industrial Chemicals (TICs) in Drinking Water
5.1.1 9210m Response to Toxic Industrial Chemical (TIC) Injections
A total of seven TICs were individually injected into the PPL at the concentration levels
given in Table 3-2. As described in Section 3.3, contaminant injections were performed
in sets. Each set consisted of sequential injections of increasing concentration to attain
the target concentrations of contaminants in the PPL. Three sets of injections were
performed for each contaminant. Section 3.6.1 thoroughly describes the change in TOC
concentration that was considered a response (or change) due to a contaminant injection.
After each set of TIC injections, the PPL was flushed with uncontaminated drinking
water before the next set of injections was performed. The results presented in this report
reflect the scenarios specific to those defined by the test/QA plan for this evaluation. As
discussed in Section 3.3, it is possible that chemical transformations took place during the
solution preparation prior to contaminant injections into the PPL.
Figure 5-1 shows an example of the 9210m response to one set of injections of colchicine
into chlorinated water. Injections of colchicine are marked on Figure 5-lwith vertical
lines and labeled with the concentration level. The TOC concentration over the time
period prior to the first injection was used as the baseline TOC concentration for the 0.01
mg/L injection. In the set of injections shown, the 9210m did not have a response to the
0.01 (below instrument specification) and 0.1 mg/L injections of colchicine but the
9210m did have a response to the 1 and 10 mg/L injections. During the 1 and 10 mg/L
injections, the 9210m TOC concentration did not increase until the third reported
measurement after the contaminant was injected into the PPL. This is because the 9210m
only draws water in for analysis once every seven minutes. Therefore, the first reported
concentration (represented by the points on the line representing the TOC values in
Figure 5-1) following the injection represented water that had been drawn into the 9210m
before the contaminant injection and the second reported concentration represented water
that was drawn into the 9210m prior to the time when the water in the PPL was well
mixed. Additionally, the response to the 10 mg/L injection shows one increased TOC
concentration data point occurring before a steady-state concentration was achieved.
This result was likely caused because the water sample taken up by the 9210m happened
to coincide with the first pass of the contaminant after the injection and before the water
in the PPL was well-mixed.
15
-------
10
9
8
7
"^j
"SB e
4
(N
at
3
2
1
Baseline
0.01 mg/L
x Injection
0.1 mg/L
>" Injection
lmg/L
Injection
10 mg/L
^Injection
30 60 90 120 150
Time (minutes)
180
210
240
Figure 5-1. Change in 9210m Total Organic Carbon (TOC) in response to injections
of colchicine.
Table 5-1 presents the contaminant injected, the concentration of the injected
contaminant, and the average and standard deviation of the measured change in TOC (as
measured by replicate measurements of the 9210m) for each TIC injection. Those
injections that were determined to produce a response (as defined previously) from the
9210m are highlighted in gray. The average change in TOC measured by the reference
instrument for the three replicates is also included. Changes in TOC measured by the
reference instrument were determined using the standard deviation of the check standard
analyses. Over the course of the evaluation, the standard deviation of reference analyses
of 2 mg/L check standards was 0.04 mg/L for 28 samples. Therefore, changes in TOC
measured by the reference instrument greater than 0.12 mg/L (three times the standard
deviation) are highlighted in gray as changes in response to contaminant injections.
The 9210m and the reference method measure the TOC concentration and not the
contaminant concentration (which would include all non-organic carbon substituents).
Because the contaminant concentrations (and not TOC concentrations) are given, the
TOC concentrations from both the reference method and the 9210m would not be
expected to be the same as the injected contaminant concentration. Therefore, the
reference method and the 9210m results would not be expected to be the same as the
injected nominal contaminant concentration, but the changes in TOC as measured by the
reference method should be similar to the change measured by the 9210m. The reference
16
-------
method and instrument used during this evaluation was selected because of its common
use. However, different TOC instruments and measurement methodologies can respond
differently to varying compounds. Therefore, the comparative results should be
interpreted with the recognition that the 9210m is being compared with a UV-persulfate
oxidation method and that there can be inherent differences in how each method
measures a particular compound and differences between the 9210m and the reference
method could indicate a deficiency in the 9210m.
Table 5-1. Change in 9210m Total Organic Carbon (TOC) from Injections of Toxic
Industrial Chemicals (TICs) into Chlorinated Water
Contaminant
Aldicarb
Carbofuran
Colchicine
Mevinphos
Nicotine
Potassium
Cyanide
Sodium
Fluoroacetate
Water
Controls
Injected
Contaminant
Cone. (mg/L)
0.1
1
0.1
1
10
0.1
1
10
0.1
1
0.1
1
10
0.1
1
10
0.1
1
10
None
Average
Reference
Method
ATOC
(mg/L)
0.08
0.46
0.04
0.52
4.05
0.07
0.66
8.02
0.04
0.36
0.03
0.67
7.01
0.02
0.02
0.34
0.01
0.25
1.98
0.01
Injection 1
ATOC
(mg/L)
0.05
0.54
0.03
0.48
2.85
0.06
0.49
5.41
0.10
0.46
0.08
0.79
7.91
0.00
0.01
0.38
0.00
0.29
1.73
0.05
-0.01
Std.
Dev.
(mg/L)
0.07
0.09
0.04
0.05
0.20
0.09
0.09
0.35
0.06
0.06
0.05
0.02
0.19
0.03
0.07
0.07
0.07
0.05
0.13
0.03
0.11
Injection 2
ATOC
(mg/L)
0.03
0.51
0.10
0.70
3.16
0.06
0.60
5.06
0.07
0.35
0.09
0.75
7.26
0.04
0.05
0.38
0.06
0.22
1.70
-0.09
0.04
Std.
Dev.
(mg/L)
0.04
0.03
0.06
0.06
0.71
0.07
0.07
0.30
0.07
0.05
0.07
0.07
0.15
0.04
0.06
0.07
0.09
0.09
0.05
0.07
0.07
Injection 3
ATOC
(mg/L)
0.10
0.48
0.01
0.61
5.61
0.08
0.62
5.16
0.07
0.32
0.03
0.79
8.15
-0.08
-0.01
0.47
-0.03
0.25
1.81
0.05
o.oit
Std.
Dev.
(mg/L)
0.06
0.07
0.06
0.06
0.22
0.04
0.08
0.10
0.11
0.11
0.10
0.10
0.18
0.06
0.07
0.07
0.06
0.02
0.05
0.09
0.06t
Responses (indicated by shading) must be at least three times the baseline standard deviation and also exceed the
average and standard deviation of the background injection result [0.07 mg/L].
f Average and standard deviation of water controls.
In chlorinated water, across all of the TICs, the 9210m did not respond to any
contaminants at the 0.1 mg/L concentration level. The change in TOC due to the
injection of these concentrations was also not able to be determined by the reference
method. With the exception of potassium cyanide, the 9210m responded to all of the
TICs at 1 mg/L. The reference method also was able to determine a change in TOC at 1
mg/L for all contaminants except potassium cyanide. Only mevinphos did not produce a
response across the three replicate 9210m injections. Two of the three mevinphos
injections at 1 mg/L resulted in a response while the third did not. The 9210m measured
a change in TOC of 0.32 mg/L during the third injection with a standard deviation of 0.11
mg/L. Therefore, the change in the TOC measurement was only 0.01 mg/L less than the
17
-------
concentration that would have been considered a response so the data was not
inconsistent. The 9210m and the reference method measured a change in TOC due to the
injection of 10 mg/L of each contaminant injected at that concentration.
Table 5-2 presents the same information for injections made into the chloraminated water
matrix. In general, the results are similar for chlorinated and chloraminated water. The
only exceptions are that, with chloraminated water, the 9210m responded to all three of
the 1 mg/L injections of mevinphos while one of the 1 mg/L injections of aldicarb into
Table 5-2. Change in 9210m Total Organic Carbon (TOC) from Injections of Toxic
Industrial Chemicals (TICs) into Chloraminated Water
Contaminant
Aldicarb
Carbofuran
Colchicine
Mevinphos
Nicotine
Potassium
Cyanide
Sodium
Fluoroacetate
Water
Controls
Injected
Contaminant
Cone. (mg/L)
0.1
1
0.1
1
10
0.1
1
10
0.1
1
0.1
1
10
0.1
1
10
0.1
1
10
None
Average
Reference
Method
ATOC
(mg/L)
0.15
0.39
0.03
0.63
4.88
0.08
0.69
6.41
0.05
0.36
0.08
0.68
6.86
0.01
0.03
0.37
0.03
0.30
2.17
0.01
Injection 1
ATOC
(mg/L)
0.25
0.54
0.05
0.71
6.01
0.12
0.64
5.97
0.12
0.39
0.19
0.60
7.54
0.05
0.04
0.59
0.08
0.25
2.63
0.05
-0.01
Std.
Dev.
(mg/L)
0.19
0.19
0.03
0.03
0.34
0.12
0.12
0.04
0.06
0.06
0.20
0.20
0.06
0.07
0.05
0.19
0.07
0.06
0.12
0.03
0.11
Injection 2
ATOC
(mg/L)
0.04
0.87
0.15
0.74
6.11
0.07
0.67
6.39
-0.03
0.45
0.08
0.89
7.93
0.03
0.03
0.36
0.11
0.29
1.76
-0.09
0.04
Std.
Dev.
(mg/L)
0.05
0.08
0.08
0.14
0.27
0.05
0.04
0.09
0.04
0.08
0.06
0.07
0.43
0.06
0.06
0.06
0.07
0.07
0.04
0.07
0.07
Injection 3
ATOC
(mg/L)
0.07
0.43
0.16
0.68
5.15
Std.
Dev.
(mg/L)
0.12
0.12
0.09
0.12
0.26
t
0.06
0.39
0.08
0.76
8.69
0.00
0.06
0.38
0.00
0.27
1.78
0.05
0.01*
0.11
0.11
0.04
0.06
0.21
0.09
0.09
0.04
0.07
0.06
0.08
0.09
0.06*
Responses (indicated by shading) must be at least three times the baseline standard deviation and also exceed the
average and standard deviation of the background injection result [0.07 mg/L].
f Data from injection 3 were lost due to a 9210m software error.
} Average and standard deviation of water controls.
chloraminated water did not produce a response from the 9210m. In that instance, the
9210m standard deviation (0.19 mg/L) was approximately three times larger than the
average standard deviation throughout the rest of the evaluation (0.07 mg/L). This large
standard deviation prevented the change in TOC concentration from being considered a
response. One other aberration from the chlorinated testing was that the reference
method was able to measure the change in TOC due to the injection of 0.1 mg/L aldicarb
18
-------
while the 9210m was not. Other than those instances, the 9210m responses matched that
of the reference method for both chlorinated and chloraminated water.
5.1.2 9210m Accuracy during Toxic Industrial Chemical (TIC) Testing
Table 5-3 and Table 5-4 compare the steady-state TOC measurements from the 9210m to
the measurements from the reference method through analysis of reference samples from
the PPL. In addition to the steady-state contaminant concentrations in the second
column, the actual TOC concentrations and standard deviations measured by the
reference method and the 9210m are shown in the middle columns. These measurements
include the background TOC concentration of the drinking water used for the testing in
addition to any contribution of TOC from the injected contaminant. The last column
includes the %D between the average 9210m TOC measurement and the average
reference TOC measurement. This %D can be used to assess the accuracy of the 9210m
measurements compared to the reference method. As mentioned before, different TOC
instruments and methods can provide different responses to contaminants so the result
should be interpreted accordingly. In most cases, the average and standard deviation
reported are based on three replicate measurements but there are a few cases where they
represent four replicates. The propagated experimental uncertainty of the %D is shown
in the same column as %D in Tables 5-3 and 5-4.
The average %D between the 9210m and the reference method was typically less than
15% with the only exceptions being 0.1 mg/L of mevinphos in chlorinated water (16%),
10 mg/L of colchicine in chlorinated water (-29%), and 10 mg/L carbofuran in
chloraminated water (15%). Overall, the average of the absolute value of the %Ds for all
the comparisons across the evaluation was 6% plus or minus (±) a standard deviation
(SD) of 5%. For TICs in chlorinated and chloraminated water, the absolute value of the
%Ds averaged 5% ± 6% and 6% ± 4%, respectively. During the TIC injections, the
differences between the 9210m results and the reference method results are mostly not
significant, indicating that the 9210m and the reference method were generating similar
results most of the time.
In addition to the accuracy data presented in Tables 5-3 and 5-4, the standard deviation
data presented represents not only the precision of the baseline measurements, but the
precision of the resulting TOC concentration due to the injections themselves. This
measure of precision could be used to calculate the amount of change that would be
required for detection of a contamination event. For example, if a set of replicate
injections is precise (e.g., 1 mg/L mevinphos) and therefore, has a low standard deviation,
a small change would be detectable. Conversely, a set of replicates (e.g. 10 mg/L
carbofuran) that were somewhat imprecise and therefore, have a higher standard
deviation would require a relatively large change in TOC to attain a detectable
concentration. These data are also useful for evaluating the precision of the 9120m
versus the reference method. In many cases, the precision data are rather similar. In this
instance, the precision is not only dependent on the instrument, but also on experimental
factors having to do with the PPL operation. However, similar variables can exist in the
context in an operational setting.
19
-------
Table 5-3. 9210m Accuracy for Total Organic Carbon (TOC) from Injections of
Toxic Industrial Chemicals (TICs) in Chlorinated Water
Contaminant
Aldicarb
Carbofuran
Colchicine
Mevinphos
Nicotine
Potassium
Cyanide
Sodium
Fluoroacetate
Injected
Contaminant
Concentration
(mg/L)
Pre-Injection
0.01
0.1
1
Pre-Injection
0.01
0.1
1
10
Pre-Injection
0.01
0.1
1
10
Pre-Injection
0.01
0.1
1
Pre-Injection
0.01
0.1
1
10
Pre-Injection
0.01
0.1
1
10
Pre-Injection
0.01
0.1
1
10
Reference TOC
Average
(mg/L)
2.52
2.47
2.55
3.01
2.45
2.51
2.55
3.07
7.12
2.41
2.43
2.50
3.16
11.19
1.38
1.37
1.42
1.77
2.54
2.55
2.63
3.29
10.05
2.37
2.41
2.43
2.45
2.79
1.55
1.55
1.55
1.81
3.66
Std.f
Dev.
(mg/L)
0.14
0.15
0.20
0.24
0.07
0.16
0.15
0.24
1.54
0.05
0.05
0.04
0.06
0.16
0.01
0.01
0.03
0.04
0.10
0.03
0.03
0.06
2.06
0.06
0.06
0.06
0.08
0.06
0.08
0.12
0.10
0.08
0.21
9210m TOC
Average
(mg/L)
2.52
2.55
2.61
3.12
2.59
2.54
2.58
3.18
7.05
2.56
2.50
2.56
3.13
8.34
1.59
1.58
1.66
2.04
2.62
2.66
2.71
3.49
10.94
2.48
2.44
2.43
2.44
2.85
1.49
1.54
1.58
1.83
3.24
Std.f
Dev.
(mg/L)
0.20
0.19
0.18
0.20
0.07
0.08
0.06
0.16
1.59
0.14
0.12
0.11
0.05
0.21
0.02
0.03
0.01
0.06
0.13
0.14
0.07
0.09
1.10
0.11
0.06
0.09
0.10
0.06
0.14
0.11
0.16
0.18
0.23
Percent
Difference
(%D)
0.0 ±9.7
3.2 ±9.5
2.3 ±10
3.6 ±10
5.6 ±3.8
1.2 ±7.0
1.2 ±6.3
3.5 ±9.1
-1.0 ±31
6.0 ±5.8
2.8 ±5.2
2.4 ±4.6
-1.0 ±2.5
-29 ±3.1
14 ±1.4
14 ±2.0
16 ±1.9
14 ±3.5
3.1 ±6.3
4.2 ±5.4
3.0 ±2.8
5.9 ±3.1
8.5 ±21
4.5 ±5.0
1.2 ±3.5
0.0 ±4.5
-0.4 ±5.2
2.1 ±3.0
-3.9 ±11
-0.6 ±11
1.9 ±12
l.lill
-12 ±9.6
f Standard deviation is of the replicates on which the average TOC concentration is based.
20
-------
Table 5-4. 9210m Accuracy for Total Organic Carbon (TOC) from Injections of
Toxic Industrial Chemicals (TICs) in Chloraminated Water
Contaminant
Aldicarb
Carbofuran
Colchicine
Mevinphos
Nicotine
Potassium
Cyanide
Sodium
Fluoroacetate
Injected
Contaminant
Concentration
(mg/L)
Pre-Injection
0.01
0.1
1
Pre-Injection
0.01
0.1
1
10
Pre-Injection
0.01
0.1
1
10
Pre-Injection
0.01
0.1
1
Pre-Injection
0.01
0.1
1
10
Pre-Injection
0.01
0.1
1
10
Pre-Injection
0.01
0.1
1
10
Reference TOC
Average
(mg/L)
2.01
1.92
2.07
2.46
1.95
2.06
2.09
2.72
7.60
1.80
1.79
1.88
2.59
9.04
.43
.44
.49
.85
.92
.95
2.03
2.71
9.56
.91
.93
.94
.97
2.34
.56
.50
.53
.83
3.86
Std.f
Dev.
(mg/L)
0.36
0.32
0.45
0.40
0.18
0.30
0.24
0.30
0.16
0.09
0.05
0.05
0.11
0.09
0.02
0.03
0.02
0.03
0.12
0.11
0.09
0.08
0.35
0.05
0.05
0.03
0.04
0.03
0.10
0.05
0.05
0.14
0.45
9210m TOC
Average
(mg/L)
2.14
2.10
2.22
2.83
2.18
2.22
2.34
3.04
8.80
1.84
1.83
1.93
2.58
8.76
1.57
1.58
1.63
2.04
1.96
1.98
2.10
2.85
10.91
1.99
1.95
1.98
2.02
2.46
1.60
1.67
1.74
2.00
3.68
Std.f
Dev.
(mg/L)
0.28
0.45
0.55
0.59
0.20
0.22
0.27
0.29
0.60
0.07
0.03
0.07
0.05
0.25
0.06
0.11
0.12
0.13
0.07
0.05
0.12
0.06
0.56
0.01
0.03
0.06
0.05
0.17
0.14
0.03
0.06
0.07
0.79
Percent
Difference
(%D)
6.3 ±21
9.0 ± 26
7.0 ±32
14 ±25
11 ±12
7.5 ±17
11 ±15
11 ±14
15 ±7.0
2.2 ±6.2
2.2 ±3.2
2.6 ±4.5
-0.4 ± 4.7
-3.1 ±3.0
9.3 ±4.0
9.3 ±7.2
9.0 ±7.4
9.8 ±6.5
2.1±7.1
1.5±6.1
3.4 ±7.1
5.0 ±3.5
13 ±6.0
4.1 ±2.6
1.0 ±3.0
2.0 ±3.4
2.5 ±3.2
5.0 ±7.0
2.5 ±11
11±3.5
13 ±4.5
8.9 ±7.8
-4.8 ± 25
f Standard deviation is of the replicates on which the average TOC concentration is based.
As is evident from the pre-injection TOC levels, the water in the laboratory did show
some variation over the course of the evaluation. The average pre-injection baseline
TOC in the water ranged from 1.38 mg/L to 2.54 mg/L during the evaluation period.
However, during the course of each set of replicate contaminant injections, the
background TOC was steady because the same water was used for each contaminant
injection.
21
-------
5.2 Biological Contaminants (BCs) in Drinking Water
5.2.1 9210m Response to Biological Contaminants (BCs) Injections
Three BCs and one toxin surrogate were injected into the PPL. These injections were
performed in the same manner as the TIC injections with a concentrated solution injected
into the PPL over approximately 15-20 seconds. The same injection and flush procedures
were used and response determination (as defined in Section 3.6.1) was performed in the
same way. Initially, concentrations of 103, 104, and 105 organisms/L were to be injected
into the PPL for each organism. After the results of the initial injections showed no
response, the concentrations were increased to include a maximum concentration of 107
organisms/L for Bacillus globigii and Bacillus thuringiensis. Limitations on the amount
of stock solution available prevented increasing the concentrations of Chlorella in a
similar manner. Figure 5-2 shows one set of injections for Bacillus globigii into
chlorinated water. Once again, the 9210m and the reference method did not exhibit a
response to any of the injected concentrations (105, 106, and 107 organisms/L) of Bacillus
globigii shown in Figure 5-2.
10
9
8
"SB 6
(J
o
iH
fM
01
Baseline
105 organism/L
^ Injection
106 organism/L
Injection
107 organism/L
// Injection
30 60 90 120
Time (minutes)
150
180
210
Figure 5-2. Change in 9210m Total Organic Carbon (TOC) in response to injections
of Bacillus globigii.
22
-------
Table 5-5 presents the contaminant injected, the concentration of the injected
contaminant, and the average and standard deviation of the measured change in TOC (as
measured by replicate measurements of the 9210m) for each BC injection. Those
injections which the 9210m measured as causing a change are highlighted in gray.
None of the injections of BCs resulted in a response from the 9210m or reference method
at the concentrations injected. Ovalbumin, a protein surrogate for biological toxins such
as ricin or botulinum, did produce a response from the 9210m for all three 1 mg/L and 10
mg/L injections.
Table 5-5 presents the response to injection of the BCs and ovalbumin into chlorinated
water. The 9210m and reference method detected a change in TOC for only the two
highest concentration injections of ovalbumin. None of the injected concentrations of
organisms resulted in a response from the 9210m or the reference method. The 9210m
responses to BCs and the toxin surrogate in chloraminated water were the same as for
chlorinated water. The 9210m detected a change in TOC for all three 10 mg/L injections
of ovalbumin. Ovalbumin was also the only contaminant that caused a change in TOC as
determined by the reference method.
Table 5-5. Change in 9210m Total Organic Carbon (TOC) from Injections of
Biological Contaminants (BCs) into Chlorinated Water
Contaminant
Bacillus
globigii
Bacillus
thuringiensis
Chlorella
Ovalbumin
Water
controls
PBS/nutrient
broth
PBS/Bold
1NV medium
Injected
Contaminant
Cone.
(organisms/L
or mg/L)
105
106
107
105
106
10'
103
104
10'
0.1
1
10
None
Equivalent to
107 Bacillus
Equivalent to
107 Chlorella
Average
Reference
Method
ATOC
(mg/L)
-0.09
0.01
-0.03
-0.03
-0.02
-0.02
-0.04
-0.04
-0.04
0.02
0.09
1.10
0.01
-0.02
-0.01
Injection 1
ATOC
(mg/L)
0.11
-0.02
0.02
0.03
0.00
0.08
-0.01
0.00
-0.04
0.04
0.13
1.22
0.05
-0.01
-0.05
-0.02
0.03
0.03
Std.
Dev.
(mg/L)
0.12
0.12
0.05
0.06
0.06
0.08
0.08
0.05
0.07
0.09
0.05
0.08
0.03
0.11
0.05
0.06
0.07
0.13
Injection 2
ATOC
(mg/L)
-0.03
0.05
0.04
0.00
0.03
-0.01
0.07
-0.01
-0.01
0.07
0.14
1.04
-0.09
0.04
-0.05
0.05
0.00
0.07
Std.
Dev.
(mg/L)
0.05
0.04
0.04
0.09
0.05
0.07
0.07
0.03
0.06
0.04
0.05
0.05
0.07
0.07
0.04
0.05
0.06
0.03
Injection 3
ATOC
(mg/L)
-0.02
0.06
-0.02
0.01
0.03
0.03
0.01
0.03
0.02
-0.05
0.17
1.15
0.05
o.oit
-0.01
-0.03
0.15
0.02
Std.
Dev.
(mg/L)
0.04
0.05
0.05
0.08
0.08
0.06
0.07
0.06
0.09
0.09
0.02
0.03
0.09
0.06t
0.07
0.05
0.09
0.06
Responses (indicated by shading) must be at least three times the baseline standard deviation and also exceed the
average and standard deviation of the background injection result [0.07 mg/L].
f Average and standard deviation of water controls.
23
-------
Table 5-6. Change in 9210m Total Organic Carbon (TOC) from Injections of BCs
into Chloraminated Water
Contaminant
Bacillus
globigii
Bacillus
thuringiensis
Chlorella
Ovalbumin
Water
controls
PBS/nutrient
broth
PBS/Bold
1NV medium
Injected
Contaminant
Cone.
(organisms/L
or mg/L)
103
104
105
10'
103
104
105
106
10'
103
104
103
0.1
1
10
None
Equivalent to
107 Bacillus
Equivalent to
107 Chlorella
Average
Reference
Method
ATOC
(mg/L)
-0.01
-0.03
-0.01
-0.05
-0.03
-0.03
-0.02
-0.04
-0.01
-0.05
-0.04
-0.02
0.04
0.29
1.34
0.01
-0.02
-0.01
Injection 1
ATOC
(mg/L)
0.04
0.04
-0.01
t
-0.01
0.01
0.07
t
t
-0.01
0.01
0.00
0.03
0.13
1.66
0.05
-0.01
-0.05
-0.02
0.03
0.03
Std.
Dev.
(mg/L)
0.06
0.06
0.05
t
0.12
0.06
0.07
t
t
0.05
0.06
0.06
0.04
0.07
0.18
0.03
0.11
0.05
0.06
0.07
0.13
Injection 2
ATOC
(mg/L)
-0.04
0.08
-0.03
t
0.02
-0.01
0.05
t
t
-0.04
0.14
0.06
-0.04
0.11
1.92
-0.09
0.04
-0.05
0.05
0.00
0.07
Std.
Dev.
(mg/L)
0.04
0.12
0.12
t
0.04
0.04
0.02
t
t
0.04
0.06
0.09
0.04
0.03
0.14
0.07
0.07
0.04
0.05
0.06
0.03
Injection 3
ATOC
(mg/L)
-0.02
0.05
0.00
-0.02
t
t
0.06
0.06
0.03
-0.03
0.10
-0.04
0.01
0.11
2.10
0.05
0.01*
-0.01
-0.03
0.15
0.02
Std.
Dev.
(mg/L)
0.09
0.08
0.06
0.09
t
t
0.05
0.05
0.07
0.06
0.06
0.06
0.04
0.03
0.13
0.09
0.06*
0.07
0.05
0.09
0.06
Responses (indicated by shading) must be at least three times the baseline standard deviation and also exceed the
average and standard deviation of the background injection result [0.07 mg/L].
f Fewer than three replicates performed at this concentration.
} Average and standard deviation of water controls.
Table 5-6 presents the response due to injection of the BCs and ovalbumin into
chloraminated water. Initial injections of Bacillus globigii were performed at 103, 104,
and 105 organisms/L. One injection of 107 organisms/L was included with the final
replicate set of injections to confirm the results determined during the chlorinated water
injections. The final set of injections of Bacillus thuringiensis was performed at 105, 106,
and 107 organisms/L. Injections which produced a response in the 9210m TOC
concentration and the reference method are highlighted in gray. In addition to the
injections of BCs, control injections freshly prepared growth media, without organisms
added, which were handled and washed in the same manner as the stock solutions of the
BCs were injected into the PPL. Triplicate sets of injections of the washed growth media
for both the Bacillus and Chlorella were made into chlorinated and chloraminated water.
None of the 12 injections produced a change in TOC that was detectable by the 9210m or
the reference method.
24
-------
5.2.2 9210m Accuracy during BC Testing
Table 5-7 and Table 5-8 summarize the comparison of the steady-state TOC
measurements from the 9210m to the results obtained from the reference method through
analysis of reference samples from the PPL for chlorinated and chloraminated water,
respectively. In addition to the steady-state contaminant concentrations in the second
column, the reference method TOC concentrations and standard deviations measured by
the 9210m and the reference method are shown in the middle columns. These
measurements include the background TOC concentration of the drinking water used for
the testing in addition to any contribution of TOC from the injected contaminant. The
last column includes the %D between the average 9210m TOC measurement and the
average reference TOC measurement. This %D can be used to assess the accuracy of the
9210m measurements. In most cases, the average and standard deviation reported are
based on three replicate measurements, but in some instances were based on four
replicates.
The percent difference between the 9210m and the reference method was typically less
than 25% with the only exceptions being 105 organisms/L oiChlorella in chlorinated
water (25%), and Bacillus thuringiensis in chloraminated water (26-28%). Overall, the
average absolute value of %D for the chlorinated water was 16% ± 8%, for
chloraminated water 13% ± 7%, for an overall average absolute %D of 14% ± 7%. For
each individual comparison, the experimental error was propagated using the uncertainty
Table 5-7. 9210m Accuracy for Total Organic Carbon (TOC) from Injections of
Biological Contaminants (BCs) in Chlorinated Water
Contaminant
Bacillus
globigii
Bacillus
thuringiensis
Chlorella
Ovalbumin
Contaminant
Concentration
(mg/L or
organisms/L)
Pre-Injection
l(f
106
10'
Pre-Injection
l(f
106
10'
Pre-Injection
103
104
l(f
Pre-Injection
0.01
0.1
1
10
Reference TOC
Average
(mg/L)
.61
.52
.54
.51
.54
.50
.48
.49
.66
.62
.58
.54
2.38
2.39
2.41
2.50
3.59
Std.f
Dev.
(mg/L)
0.09
0.10
0.06
0.06
0.08
0.09
0.11
0.07
0.02
0.02
0.01
0.03
0.02
0.01
0.02
0.04
0.31
9210m TOC
Average
(mg/L)
1.90
1.92
1.95
1.93
1.78
1.80
1.82
1.83
1.96
1.98
1.99
1.98
2.60
2.48
2.50
2.65
3.78
Std.f
Dev.
(mg/L)
0.07
0.06
0.05
0.04
0.07
0.06
0.05
0.10
0.12
0.12
0.10
0.08
0.11
0.09
0.02
0.04
0.11
Percent
Difference
(%D)
16 ±6.0
23 ±6.0
24 ±3. 9
24 ±3.7
14 ±5.9
18 ±6.0
21 ±6.6
20 ± 6.6
17 ±6.2
20 ±6.1
23 ±5.0
25 ±4.2
8.8 ±4.3
3.7 ±3.7
3.7±1.1
5.8±2.1
5.2 ±8.7
Standard deviation is of the replicates on which the average TOC concentration is based.
TOC - total organic carbon
25
-------
Table 5-8. 9210m Accuracy for Total Organic Carbon (TOC) from Injections of
Biological Contaminants (BCs) in Chloraminated Water
Contaminant
Bacillus
globign
Bacillus
thuringiensis
Chlorella
Ovalbumin
Contaminant
Concentration
(mg/L or
organisms/L)
Pre-Injection
103
104
10*
10'
Pre-Injection
103
104
10*
106
10'
Pre-Injection
103
104
10'
Pre-Injection
0.01
0.1
1
10
Reference TOC
Average
(mg/L)
.76
.75
.72
.71
.62
.63
.68
.65
.56
.38
.38
.46
.41
.37
.36
.80
.78
.83
2.12
3.46
Std.f
Dev.
(mg/L)
0.02
0.01
0.03
0.04
}
0.14
0.01
0.01
0.12
}
}
0.08
0.07
0.07
0.08
0.04
0.03
0.04
0.28
0.20
9210m TOC
Average
(mg/L)
1.84
1.84
1.89
1.88
1.91
1.82
1.90
1.89
1.88
1.79
1.82
1.67
1.64
1.72
1.69
1.92
1.89
1.89
2.01
3.90
Std.f
Dev.
(mg/L)
0.06
0.04
0.03
0.04
}
0.13
0.03
0.05
0.13
}
}
0.11
0.10
0.12
0.11
0.12
0.11
0.08
0.07
0.29
Percent
Difference
(%D)
4.4 ±3.4
5.0 ±2.2
9.4 ±2.2
9.5 ±3.0
16
11 ±10
12 ±1.7
14 ±2.7
19 ±9.3
26
28
13 ±8.1
15 ±7.4
23 ± 8.0
22 ± 8.0
6.5 ±6.6
6.0 ±6.0
3.2 ±4.7
-5.3 ± 14
12 ±9.0
f Standard deviation is of the replicates on which the average TOC concentration is based.
J Only one replicate conducted at these concentrations.
TOC - total organic carbon
of both measurements. For the analyses performed during the BC injections, the
differences between the 9210m results and the reference method results are larger than
the experimental uncertainty, indicating that the 9210m and the reference method were
not performing equally.
As is evident from the pre-injection TOC levels, the water in the laboratory did show
some variation over the course of the testing. The average pre-injection baseline TOC in
the water ranged from 1.46 mg/L to 2.38 mg/L during the test period. However, during
the course of each set of replicate contaminant injections, the background TOC was
steady because the same water was used for each contaminant injection.
5.3 Discrete Sample Analyses
5.3.1 921 Om Response for Discrete Sample Analyses
Three contaminants were analyzed as discrete samples. Carbofuran was analyzed as
discrete samples as well as in the PPL. For the PPL tests, water was continuously
flowing past the 9210m inlet and during the discrete tests the inlet was placed in a static
container containing the solution being tested. The 9210m was allowed to draw several
26
-------
(at least four) samples from a reservoir and analyze the samples for each contaminant
concentration. These tests were carried out with containers containing 100 mL of
solutions at the desired contaminant concentration. The reported concentration is the
average of the measurements collected from one container. The discrete samples were
analyzed from uncontaminated drinking water to the highest contaminant concentration.
For each contaminant, three containers of each concentration were analyzed in both
chlorinated and chloraminated water.
Table 5-9 and Table 5-10 give the contaminant tested, the concentration of the
contaminant, the measured change in TOC, and the standard deviation for each
contaminant concentration in chlorinated and chloraminated water, respectively. Those
concentrations which exhibited a response from the 9210m are highlighted in gray.
Results of the ricin and sodium azide phosphate buffer samples are included in both
tables. The ricin was stored in the sodium azide phosphate buffer so along with each
measurement of ricin, the sodium azide phosphate buffer was analyzed at the
concentration at which it was present in the ricin solutions. The diesel fuel, disulfoton,
and carbofuran were insoluble or only partially soluble in water making interpretation of
the 9210m and reference method results for these discrete samples difficult.
Table 5-9. Change in 9210m Total Organic Carbon (TOC) for Discrete Sample
Analyses in Chlorinated Water
Contaminant
Carbofuran
Diesel fuel
Disulfoton
Ricin
Sodium
azide/phosphate
buffer (ricin
blank)
Solution
Contaminant
Cone. (mg/L)
0.1
1
0.1
1
10
0.1
1
0.1
1
10
0.1
1
10
Average
Reference
Method
ATOC
(mg/L)
t
t
0.12
0.00
-0.02
-0.01
-0.23
*
Container 1
ATOC
(mg/L)
0.04
0.20
0.06
0.01
0.20
-0.03
-0.23
0.37
0.68
2.64
-0.08
-0.06
-0.07
Std.
Dev.
(mg/L)
0.12
0.12
0.05
0.05
0.05
0.11
0.07
0.08
0.13
0.21
0.12
0.12
0.12
Container 2
ATOC
(mg/L)
0.03
0.27
0.17
0.02
0.19
-0.10
-0.19
-0.02
0.17
2.11
Std.
Dev.
(mg/L)
0.05
0.08
0.09
0.09
0.09
0.09
0.09
0.05
0.07
0.12
Container 3
ATOC
(mg/L)
0.10
0.38
0.14
0.01
0.07
-0.06
-0.18
0.06
0.14
2.36
Std.
Dev.
(mg/L)
0.04
0.03
0.08
0.08
0.08
0.05
0.06
0.14
0.14
0.18
#
Responses (indicated by shading) must be at least three times standard deviation.
Reference samples not analyzed for discrete analyses of carbofuran in chlorinated water.
*
Reference instrument could not be placed within the biosafety hood so ricin and corresponding azide blanks reference
analyses were not performed.
^
One set of blank replicate samples was analyzed.
27
-------
Table 5-10. Change in 9210m Total Organic Carbon (TOC) for Discrete Sample
Analyses in Chloraminated Water
Contaminant
Carbofuran
Diesel fuel
Disulfoton
Ricin
Sodium
azide/phosphate
buffer (ricin
blank)
Solution
Contaminant
Cone. (mg/L)
0.1
1
0.1
1
10
0.1
1
0.1
1
10
0.1
1
10
Average
Reference
Method
ATOC
(mg/L)
0.07
0.30
0.00
0.03
0.05
-0.01
0.29
}
Container 1
ATOC
(mg/L)
0.00
0.06
-0.04
0.02
0.09
0.07
0.41
-0.08
0.08
2.09
-0.02
-0.04
-0.04
Std.
Dev.
(mg/L)
0.08
0.04
0.07
0.07
0.07
0.03
0.10
0.13
0.13
0.13
0.04
0.05
0.04
Container 2
ATOC
(mg/L)
0.20
0.34
0.01
0.04
0.04
0.04
0.26
0.02
0.18
1.97
Std.
Dev.
(mg/L)
0.11
0.11
0.07
0.07
0.08
0.05
0.05
0.07
0.10
0.11
Container 3
ATOC
(mg/L)
0.12
0.21
-0.07
0.00
0.12
-0.04
0.27
-0.02
-0.03
1.69
Std.
Dev.
(mg/L)
0.07
0.05
0.06
0.06
0.06
0.08
0.08
0.05
0.11
0.17
#
Responses (indicated by shading) must be at least three times standard deviation.
Reference instrument could not be placed within the biosafety hood so ricin and corresponding azide blanks reference
analyses were not performed.
One set of blank replicate samples was analyzed.
5.3.1.1 Carbofuran
Discrete analyses of carbofuran at 0.1 and 1 mg/L were conducted. No analysis of
discrete 10 mg/L carbofuran samples were conducted due to the limited solubility of
carbofuran at that concentration. During the discrete analyses, the 9210m measured a
change in TOC for the 1 mg/L concentration level in both chlorinated and chloraminated
water. This concentration had also exhibited a change in 9210m and reference method
TOC concentration when injected into the PPL. The discrete samples resulted in about
half the level of TOC that had been determined during the PPL analyses. The reason for
this difference in results between the experimental setups was not obvious.
5.3.1.2 Diesel fuel
Discrete analysis of diesel fuel samples was conducted at 0.1, 1, and 10 mg/L. The
solubility of diesel was such that diesel was not soluble in water at any of the
concentrations evaluated. For each sample, two phases were visible with diesel on top of
water. Prior to each analysis, the samples were mixed, but over the course of the analysis
period the samples began to separate into a distinct diesel (organic) phase and an aqueous
phase. Therefore, during the first analysis of the 10 mg/L diesel sample in chlorinated
water, the 9210m did measure a change in TOC, however, there were no other changes in
either water matrix. With similar inconsistency, only the lowest concentration diesel was
detected by the reference method. These very inconsistent results can be explained by
28
-------
the lack of solubility of diesel fuel. Without a homogeneous solution, results from any
chemical measurement, including TOC, is problematic.
5.3.1.3 Disulfoton
Discrete analysis of disulfoton samples was conducted at 0.1 and 1 mg/L. No analysis of
discrete 10 mg/L disulfoton samples was conducted due to the limited solubility of
disulfoton. None of the disulfoton samples produced an increased 9210m TOC response
in chlorinated water. However, for both the reference method and the 9210m the 1 mg/L
samples caused a decreased response. The 9210m and the reference method responded
with an increased TOC to all three 1 mg/L samples in chloraminated water. There was
not a clear reason for the notably different responses between the two water matrices.
5.3.1.4 Ricin
Ricin tests were carried out inside a hood in a biological safety level 2 laboratory.
Discrete analysis of ricin samples was conducted at 0.1, 1, and 10 mg/L. The 9210m
measured a TOC response at 0.1 and 1 mg/L for only one of three sets of tests in
chlorinated water; at 10 mg/L test in chlorinated and chloraminated water, all replicates
(three out of three) precipitated a TOC response. As mentioned above, the sodium azide
phosphate buffer solution was analyzed at the concentrations at which it was present in
the ricin solutions to determine whether the 9210m measured a change in TOC due to the
sodium azide phosphate buffer. The phosphate buffered sodium azide solution did not
cause any change in TOC in either water matrix, therefore, the change in TOC
concentration due to ricin was considered to be anomalous. Analysis of ricin by the
reference method would have required relocation of the reference instrument into a BL-2
hood, which was not logistically possible.
5.3.2 9210m Accuracy during Discrete Sample Analyses
Table 5-11 and Table 5-12 summarize the comparison of the TOC measurements from
the 9210m to the results obtained from the reference method through analysis of samples
prepared in the same manner as the test samples analyzed by the 9210m in chlorinated
and chloraminated water, respectively. The %D of the 9210m compared to the reference
method was typically less than 10% for discrete sample analyses. The only exceptions
were the carbofuran samples in chloraminated water (10-20%). Overall, the average
absolute value of %D for the chlorinated water was 5% ± 2%, for chloraminated water
6% ± 7%, for an overall average %D of 6% ± 6%.
29
-------
Table 5-11. 9210m Accuracy for Total Organic Carbon (TOC) from Discrete
Sample Analyses of Contaminant Injections in Chlorinated Water
Contaminant
Carbofuran
Diesel fuel
Disulfoton
Contaminant
Concentration
(mg/L)
Water
0.1
1
Water
0.1
1
10
Water
0.1
1
Reference TOC
Average (mg/L)
NAf
1.64
1.76
1.64
1.62
1.97
1.96
1.73
9210m TOC
Average
(mg/L)
1.63
1.68
1.91
1.51
1.64
1.53
1.67
1.90
1.84
1.70
Std.
Dev.
(mg/L)
0.04
0.03
0.03
0.07
0.03
0.07
0.13
0.02
0.05
0.01
Percent
Difference
(%D)
NA
NA
NA
-8.3 ±4.8
-7.1 ±2.2
-6.9 ±4.8
3.0 ±7.9
-3.6 ± 1.5
-6.3 ±2.9
-1.7 ± 1.3
Reference samples not analyzed for this concentration.
Table 5-12. 9210m Accuracy for Total Organic Carbon (TOC) from Discrete
Sample Analyses of Contaminant Injections in Chloraminated Water
Contaminant
Carbofuran
Diesel fuel
Disulfoton
Contaminant
Concentration
(mg/L)
Water
0.1
1
Water
0.1
1
10
Water
0.1
1
Reference TOC
Average (mg/L)
1.41
1.48
1.71
1.89
1.88
1.92
1.94
1.90
1.89
2.19
9210m TOC
Average
(mg/L)
1.69
1.80
1.90
1.89
1.86
1.91
1.98
1.84
1.86
2.15
Std.
Dev.
(mg/L)
0.14
0.05
0.04
0.04
0.01
0.02
0.07
0.02
0.06
0.07
Percent
Difference
(%D)
18 ±8.3
20 ±3.0
10 ±2.3
0.0 ±2.4
-1.1± 1.2
-0.5 ± 1.5
2.0 ±3.7
-3.2 ±1.5
-1.6 ±3.4
-1.8 ±3.4
5.4 Additional Tests
5.4.1 Effect of Elevated Total Organic Carbon (TOC) Concentration on 9210m
Response
A minor component of this evaluation was undertaken to determine if the background
TOC level had any effect on the ability of the 9210m to detect a change in response to a
contaminant injection. This control included three replicate injections of nicotine at 0.1
and 1 mg/L into chlorinated water fortified with quinine to raise the background TOC
level. Nicotine was chosen as it is a relatively easy contaminant for TOC analyzers to
detect. Quinine was added to the PPL to increase the background TOC concentration by
30
-------
Table 5-13. Change in 9210m Total Organic Carbon (TOC) with Elevated TOC
Concentrations for chlorinated water
Contaminant
Nicotine
(background
TOC)
Nicotine
(elevated
TOC)
Injected
Contaminant
Cone. (mg/L)
0.1
1
0.1
1
Average
Reference
Method
ATOC
(mg/L)
0.03
0.67
0.11
0.65
Injection 1
ATOC
(mg/L)
0.08
0.79
0.08
0.54
Std.
Dev.
(mg/L)
0.05
0.02
0.07
0.06
Injection 2
ATOC
(mg/L)
0.09
0.75
0.13
0.65
Std.
Dev.
(mg/L)
0.07
0.07
0.05
0.03
Injection 3
ATOC
(mg/L)
0.03
0.79
0.02
0.61
Std
Dev.
(mg/L)
0.10
0.10
0.06
0.05
Responses (indicated by shading) must be at least three times the baseline standard deviation and also exceed the
average and standard deviation of the background injection result [0.07 mg/L].
approximately 1 mg/L. Additional experiments would needed to obtain more definitive
results, but Table 5-13 presents the response from the elevated TOC injections as well as
those from the injections of nicotine at the same levels without elevated background
TOC.
The addition of quinine was not observed to have an effect on the ability of the 9210m to
detect a change in TOC. The 9210m did not detect a response at 0.1 mg/L in the
background TOC water or the water with elevated TOC, but a response was detected for
the 1 mg/L injections in both the background TOC water and the elevated TOC water.
The reference method also measured a change in TOC at 1 mg/L, but not at 0.1 mg/L.
5.4.2 Effect of Elevated Ionic Strength on 9210m Response
The effect of elevated ion strength on 9210m response was also evaluated. This control
included three replicate injections of nicotine at 0.1 and 1 mg/L into chlorinated water
which had been fortified with calcium chloride to raise the calcium cation concentration
from approximately 42 mg/L to 126 mg/L. The response from these elevated ionic
strength injections was used to determine if the ionic strength of the water had any effect
on the ability of the 9210m to detect a change in response to a contaminant injection.
Grab samples collected before and after the addition of the calcium chloride were
analyzed to confirm that the calcium chloride addition tripled the background calcium
concentration. Table 5-14 presents the results from the elevated ionic strength injections
as well as those from the injections of nicotine at background ionic strength. An increase
in ionic strength was not observed to have an effect on the ability of the 9210m to detect
changes in TOC concentration due to nicotine injections. The same responses were
observed in the background ionic strength water and the elevated ionic strength water.
However, high concentrations of chloride interfered with the reference measurements so
a decreased reference result was obtained. The 9210m was apparently less affected by
the increased chloride concentration while the reference method was inhibited. However,
additional experiments would be needed to obtain more definitive results.
31
-------
Table 5-14. Change in 9210m Total Organic Carbon (TOC) with Elevated Ionic
Strength
Contaminant
Nicotine
(background
ionic
strength)
Nicotine
(elevated
ionic
strength)
Injected
Contaminant
Cone. (mg/L)
0.1
1
0.1
1
Average
Reference
Method
ATOC
(mg/L)
0.03
0.67
0.01
0.23
Injection 1
ATOC
(mg/L)
0.08
0.79
0.02
0.71
Std.
Dev.
(mg/L)
0.05
0.02
0.05
0.04
Injection 2
ATOC
(mg/L)
0.09
0.75
0.04
0.60
Std.
Dev.
(mg/L)
0.07
0.07
0.09
0.05
Injection 3
ATOC
(mg/L)
0.03
0.79
0.05
0.59
Std.
Dev.
(mg/L)
0.10
0.10
0.06
0.06
Responses (indicated by shading) must be at least three times the baseline standard deviation and also exceed the
average and standard deviation of the background injection result [ 0.07 mg/L].
5.4.3 Effect of Monochloramine Level on 9210m Response
Lastly, we conducted a preliminary experimental to determine whether the level of
monochloramine had an effect on the TOC measurement from the 9210m.
Monochloramine levels of 1.92, 5.72, and 7.54 mg/L were prepared in the PPL following
the procedure described above. Table 5-15 presents the 9210m and reference TOC
measurements for the tests conducted at different monochloramine concentrations. The
column at the left of the table shows the different monochloramine concentrations. At the
lowest monochloramine concentration of 1.92 mg/L, the baseline 9210m TOC
concentration was 1.90 mg/L. For the reference method, the TOC baseline was 1.74
mg/L. Neither the 9210m nor the reference method responded with changes in
monochloramine concentration from 1.92 mg/L to 7.54 mg/L.
Table 5-15. Change in 9210m Total Organic Carbon (TOC) with Varying
Monochloramine Concentrations
Monochloramine
Concentration
(mg/L)
1.92
5.72
7.54
Reference
Method
TOC
(mg/L)
1.74
1.69
1.69
Reference
Method
ATOC
(mg/L)
Baseline
-0.05
-0.00
9210m
TOC
(mg/L)
1.90
1.92
1.95
9210m
Std.
Dev.
(mg/L)
0.05
0.07
0.08
9210m
ATOC
(mg/L)
Baseline
0.02
0.05
Responses (indicated by shading) must be at least three times the baseline standard deviation and also exceed the
average and standard deviation of the background injection result [0.07 mg/L].
32
-------
5.5 Operational Characteristics
Operational characteristics of the 9210m for this evaluation are organized into the
following categories:
• Training/Education Material
• Installation
• Operation
• Maintenance/Consumables/Waste
• Software/Data Collection.
As noted, the development schedule of the 9210m did not allow for a final production
unit to be ready for the TTEP testing. Therefore, some operational factors found for the
"Alpha" unit will be modified in the final production units.
5.5.1 Training/Educational Material
The training for operation and maintenance of the 9210m was a combination of vendor
provided in-person training and printed instructional material. A vendor representative
set up the 9210m and gave an overview of the operation, calibration, and maintenance of
the instrument. Instruction in operation of the operating software was also provided.
The printed material was provided on a CD along with the operating software. The
installation and start-up guide included information for installing and configuring the
9210m, calibration and calibration verification instructions, and operating instructions.
Also included were requirements for user provided components and instructions for the
preparation of the phosphoric acid reagent. The printed material was easy to read and
understand but did not cover all of the maintenance tasks required over the course of the
evaluation. The 9210m tested was a prototype, and as such, the printed material was
generated specifically for this testing. The printed material focused on installation and
routine operation but did not contain information on troubleshooting or routine
maintenance. Information required to complete routine maintenance tasks was obtained
via telephone conversations with the vendor.
5.5.2 Installation
Installation of the 9210m was straight-forward. The vendor had shipped the 9210m
ahead of their arrival. Upon arrival, the vendor assembled the stand and mounted the
9210m enclosure. The flow-through inlet was plumbed into the PPL and the recycle
reservoir. The process waste, which had reagent added to it, was directed to a waste
container in the laboratory, but would normally have been directed to a drain. The
vendor installed the operation software and performed the instrument start-up procedure.
The contractor calibrated the 9210m using vendor supplied calibration solutions and
following the instructions provided in the installation and start-up guide.
33
-------
5.5.3 Operation
After the evaluation staff had become familiar with using the 9210m, operation was
straight-forward. At the start of each day of testing, the operating software was started
and a new data file created. At the end of each day of testing, analysis was stopped and
the data file from the day was transferred to a dedicated computer. The 9210m remained
powered up during periods of inactivity (overnight) to minimize start-up time each day.
These activities were required by the experimental design and would not be required in
an operational setting.
5.5.4 Maintenance/Consumables/Waste
The 9210m required a 5% phosphoric acid reagent solution for operation. This reagent
was provided by the vendor in 4 L containers. One container of phosphoric acid that was
installed upon start-up lasted for approximately 6 weeks. The second container was
adequate for the balance of the evaluation. Upon replacement of the phosphoric acid
reagent, the 9210m power was cycled to restore proper operation.
The only other maintenance performed on the instrument during testing was the
replacement of soda lime in the process gas module. This replacement was necessary
after approximately 45 days of operation over about 3 months and took around 30
minutes to complete. This maintenance is only required with the integrated Process Gas
Module. Without the Process Gas Module the 9210m can be run on bottled/compressed
gas.
Waste from the 9210m was collected in a dedicated waste container and then disposed of.
Waste was generated at a rate of approximately 100 mL an hour. During normal
operation the waste from the 9210m would typically be routed directly to a drain without
collection in a waste container.
5.5.5 Software/Data Collection
The vendor supplied software (TOM Host) was used to operate the 9210m. TOM Host
was installed on a laptop computer which was connected to the 9210m via RS-232
connection. The software provided a real-time view of the non-dispersive infrared
(NDIR) response as well as the last measurement reading. The software was used to start
new daily data files as well as to periodically download data from the 9210m memory.
The software generally worked well, but an overflow error message was displayed on the
screen on approximately 10 of the 56 days of testing. When this error occurred, the
software had to be restarted. Typically, the error message was seen immediately and the
software restarted quickly resulting in no loss of data. On one occasion, the error
message was not detected immediately and data from one set of injections were lost. The
error was caused by a problem in the software that the vendor indicated has been fixed in
the newest revision.
34
-------
6.0 Performance Summary
Summary results from evaluation of the 9210m are presented below for each performance
parameter evaluated. Discussion of the observed performance can be found in Section 5
of this report.
6.1 9210m Response to Contaminant Injections
Contaminant injections were performed for aldicarb, carbofuran, colchicine, diesel fuel,
disulfoton, mevinphos, nicotine, potassium cyanide, sodium fluoroacetate, Bacillus
globigii, Bacillus thuringiensis, Chlorella, ovalbumin, and ricin. The contaminant
injection solutions were prepared within 24 hours (most within 8 hours) in the same water
that was within the PPL. Since this water contained disinfectants it could cause
degradation or transformation of the injected contaminants prior to injection. The 9210m
measured a change in TOC in response to aldicarb, carbofuran, colchicine, mevinphos,
nicotine, and sodium fluoroacetate at a minimum concentration of 1 mg/L in both
chlorinated and chloraminated water. The 9210m measured a change in TOC in response
to potassium cyanide only at 10 mg/L for both chlorinated and chloraminated water.
The 9210m did not measure a change in TOC in response to injections of Bacillus
globigii, Bacillus thuringiensis, or Chlorella at any injected concentration. Ovalbumin
injected at 1 mg/L and 10 mg/L in both chlorinated and chloraminated water was
measured by the 9210m as an increase in TOC. Disulfoton, diesel fuel, and ricin, were
analyzed only as discrete samples and carbofuran was analyzed both using the PPL and
discretely. For disulfoton, the 9210m measured a change in TOC in response to 1 mg/L
in chloraminated water, but not in chlorinated water. Diesel fuel was not soluble in water
so the results were inconsistent and difficult to interpret. For ricin, the 9210m measured
a change in TOC in response to 10 mg/L in chlorinated and chloraminated water. In
addition to these measurements, limited experiments were performed to examine the
effect of elevated TOC, ionic strength, and monochloramine concentrations on the 9210m
TOC measurements.
6.2 Accuracy of 9210m Measurements
The TOC measurements from the 9210m were compared with those from a commonly
used reference method during all of the contaminant injections performed during the
evaluation. These comparisons should be interpreted with the awareness the different
TOC instruments and oxidation methods can respond differently to various contaminants.
Overall, the average absolute value of the %Ds for all the comparisons across the
evaluation was 6% ± a SD of 5%. For TICs in chlorinated and chloraminated water, the
%Ds averaged 5% ± 6% and 6% ± 4%, respectively. For the biological contaminants
(BC), the %D averaged 16% ± 8% and 13% ± 7%, for chlorinated and chloraminated
water, respectively. For each individual comparison, the experimental error was
propagated using the uncertainty of both measurements. During the TIC injections, the
differences between the 9210m results and the reference method results are mostly not
35
-------
significant, however, during the BC injections, the differences are often larger than the
experimental uncertainty.
6.3 Operational Characteristics
During the evaluation of the 9210m, general operational characteristics were observed
over the four months that the unit was tested. Installation and operation of the 9210m
was straight forward and clearly articulated by the vendor during a one day visit.
Operation of the 9210m instrument control software was simple and intuitive. The
9210m used an aqueous phosphoric acid reagent to facilitate the electrochemical
oxidation and soda lime as a constituent of a gas filter. The phosphoric acid reagent was
changed once during the evaluation when the acid inlet was no longer submerged in the
reagent reservoir. The soda lime was changed on the recommendation of the vendor only
after a change in the response of the 9210m, relative to the reference measurement, was
observed. The 9210m software produced an overflow error on approximately 10 of the
days of testing. When this happened, it required a restart of the software and re-
initialization of data acquisition. During one set of colchicine injections, this error
caused the loss of data. The vendor attributed this error to the fact that the instrument
tested was an "Alpha" development unit. According to OI Analytical, this problem has
been corrected in a newer release of the data collection software. Additionally, the
9210m is designed for connection to a SCADA for data collection rather than using the
computer interface as the primary means of data collection. This evaluation did not
consider other possible data retrieval methods (e.g. SCADA) that could be utilized with
the 9210m.
36
-------
7.0 References
1. Quality Management Plan for the Technology Testing and Evaluation Program,
Version 3.0, Battelle, Columbus, Ohio, January 2008.
2. U.S. EPA, EPA Method 330.5, Chlorine, Total Residual (Spectrophotometric, DPD) in
Methods for Chemical Analysis of Water and Wastes, EPA/600/4-79/020, March 1983.
3. Hach Method 10200, Nitrogen, Free Ammonia and Chloramine (Mono), Indophenol
Method (0-4.50 mg/L C12 and 0-0.50 mg/L NH3-N) for finished chloraminated
drinking water. April 2004. U.S. Patent 6,315,950, filed September 4, 1998 and issued
November 13, 2001.
4. R.C. Statt and J. A. Zeikus. 1993. Utex—the culture collection of algae at the
University of Texas at Austin. J. Phycol. 29: 1-106.
5. Hall, John; Zaffiro, Alan D.; Marx, Randall B.; Kefauver, Paul C., Krishnan, E.
Radha; Haught, Roy C.; Herrmann, Jonathan G., "On-line water quality parameters as
indicators of distribution system contamination," Journal American Water Works
Association (AWWA), 99(l):66-77, January 2007.
6. Burrows, W. Dickinson; Renner, Sara E., "Biological warfare agents as threats to
potable water" Environmental Health Perspectives, 107 (12):975-984, December
1999.
7. B. B. Potter and Wimsatt, J. C. Method 415.3, Determination of Total Organic Carbon
and Specific UV Absorbance at 254 nm in Source Water and Drinking Water. U.S.
Environmental Protection Agency, National Exposure Research Lab: Cincinnati, Ohio.
EPA/600/R-05/055, February 2005.
37
-------
United States
Environmental Protection
Agency
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT NO. G-35
Office of Research and Development (8101R)
Washington, DC 20460
Official Business
Penalty for Private Use
$300
------- |