May 2006
Environmental Technology
Verification Report
CONSTELLATION TECHNOLOGY
CORPORATION
CT-1128 PORTABLE GAS
CHROMATOGRAPH-MASS SPECTROMETER
Prepared by
Battelle
Batteiie
Ine Business of Innovation
Under a cooperative agreement with
U.S. Environmental Protection Agency
ET1/ET1/ET1/
-------
THE ENVIRONMENTAL TECHNOLOGY VERIFICATION
PROGRAM
oEPA Baiteiie
U.S. Environmental Protection Agency ^ BtlsillCSS of Innovation
ETV Joint Verification Statement
TECHNOLOGY TYPE: Mobile Mass Spectrometers
APPLICATION: Monitoring volatile organic compounds, pesticides, and
chemical agents in water
TECHNOLOGY
NAME: CT-1128 Portable Gas Chromatograph-Mass
Spectrometer (GC-MS)
COMPANY: Constellation Technology Corporation
ADDRESS: 7887 Bryan Dairy Road PHONE: (727) 547-0600
Suite 100 FAX: (727) 545-6150
Largo, FL 33777-1498
WEB SITE: www.contech.com
E-MAIL: info@contech.com
The U.S. Environmental Protection Agency (EPA) has established the Environmental Technology Verification
(ETV) Program to facilitate the deployment of innovative or improved environmental technologies through
performance verification and dissemination of information. The goal of the ETV Program is to further
environmental protection by accelerating the acceptance and use of improved and cost-effective technologies.
ETV seeks to achieve this goal by providing high-quality, peer-reviewed data on technology performance to
those involved in the design, distribution, financing, permitting, purchase, and use of environmental
technologies. Information and ETV documents are available at www.epa.gov/etv.
ETV works in partnership with recognized standards and testing organizations, with stakeholder groups
(consisting of buyers, vendor organizations, and permitters), and with individual technology developers. The
program evaluates the performance of innovative technologies by developing test plans that are responsive to
the needs of stakeholders, conducting field or laboratory tests (as appropriate), collecting and analyzing data,
and preparing peer-reviewed reports. All evaluations are conducted in accordance with rigorous quality
assurance (QA) protocols to ensure that data of known and adequate quality are generated and that the results
are defensible.
The Advanced Monitoring Systems (AMS) Center, one of six technology areas under ETV, is operated by
Battelle in cooperation with EPA's National Exposure Research Laboratory. The AMS Center evaluated the
performance of the Constellation Technology Corporation's CT-1128 Portable Gas Chromatograph-Mass
Spectrometer (GC-MS). This verification statement provides a summary of the test results.
-------
VERIFICATION TEST DESCRIPTION
Many volatile and semivolatile contaminants in water are detected using bench-top mass spectrometers in a
traditional laboratory setting. However, the CT-1128 verified in this test was a portable unit designed to be
taken outside the laboratory setting for field analysis. This portability offers an advantage to first-responders
and other users who need chemical information when time, sampling, and other limitations preclude analysis
in a laboratory.
The ability of the CT-1128 to identify and quantify target contaminants was tested in various water matrices.
The CT-1128 was evaluated for the following performance parameters:
• Accuracy
• Precision
• Linearity
• Instrument stability
• Potential matrix and interference effects
• Sensitivity
• Field portability
• Operational factors.
Three classes of contaminants were used for testing: volatile organic compounds (benzene, toluene, ethyl
benzene, total xylenes [BTEX]), pesticides (2,4-D and dicrotophos), and chemical warfare agents (VX, GB,
and GD). The contaminants were selected based on recommendations from the AMS Center stakeholders.
Performance test (PT) samples were prepared in American Society for Testing and Materials (ASTM) Type II
water. The target contaminant concentrations were constructed to bracket the concentrations of interest, which
were calculated using LD50 values assuming a 70-kilogram individual consuming 250 milliliters of the
contaminated water. When LD50 data were not available or feasible for testing, maximum contaminant levels
(MCLs), as defined by EPA National Primary Drinking Water Regulations, were used. Reference
measurements were conducted on PT samples only, to confirm the accuracy of sample preparation. EPA
methods 524.2 and 515.1 were used for analyzing BTEX and 2,4-D, respectively. Internally developed
methods were used for the remainder of the contaminant reference methods since no external methods were
available from commercial laboratories at the time of testing.
The PT samples were used to determine the accuracy of the CT-1128; one set of which was used to establish a
calibration curve. Subsequent analyses of the PT samples on multiple testing days were then used to calculate
the accuracy of the CT-1128 measurements. To measure the potential matrix effects on the CT-1128 in
selected real-world applications, it was challenged by analyzing samples fortified with the target contaminant
in various matrices including drinking water (DW) samples (which varied in source and treatment), a weakly
buffered water sample, a strongly buffered water sample, and atrihalomethanes (THMs)-fortified water
sample. The concentration of a mid-level PT sample was used to fortify the matrix samples. This
concentration provided a convenient level that was approximate to or below the concentration of interest for
the target contaminants.
In addition to the PT, DW, buffered waters, and THMs-fortified water samples, blanks and unfortified matrix
samples were analyzed to confirm negative responses in the absence of target contaminants and also to ensure
that no sources of contamination were introduced during the analysis.
Experienced GC-MS operators were used for testing since the vendor suggests that a new user obtain training
in the use of a GC-MS prior to operating the CT-1128. The vendor identified solid phase microextraction
(SPME) as the technique for preparing the water samples for subsequent GC-MS analysis by the CT-1128. It
is very important to note that the methodology provided by the vendor was not optimized for any one specific
target chemical. The same SPME fiber type and GC column were used throughout the test for all analytes.
-------
QA oversight of verification testing was provided by Battelle and EPA. Battelle QA staff conducted a
technical systems audit, a performance evaluation audit, and a data quality audit of 10% of the test data. This
verification statement, the full report on which it is based, and the test/QA plan for this verification test are all
available at www.epa.gov/etv/centers/centerl.html.
TECHNOLOGY DESCRIPTION
The following description of the CT-1128 is based on information provided by the vendor. This technology
description was not verified in this test.
The CT-1128 analyzes, on-site, known and unknown chemicals. The CT-1128 is a lightweight, ruggedized,
field deployable GC-MS system that can accommodate the applications of traditional laboratory based GC-
MS systems. With the appropriate extraction techniques, analysis may be performed on a variety of matrices
including DW, which can be prepared using SPME.
The CT-1128 weighs approximately 75 pounds (34 kilograms) and is 15 inches (38.1 cm) by 23 inches (58.4
cm) by 15 inches (38.1 cm). It is contained in a carrying case housing the entire system. The CT-1128 has a
range of 1.6 to 800 atomic mass units with unit resolution throughout the mass range. In selected ion mode
(SIM), the CT-1128 can scan for 50 groups of masses with 30 masses per group. For identification of
chemicals, the CT-1128 is equipped with an automated mass spectral data base searching function that can
use a range of commercial mass spectral libraries (e.g., National Institute of Standards and Technology Mass
Spectral Library) as well as user-defined libraries. The system is controlled with a laptop computer that uses a
program for GC control and MSB Chemstation (Agilent Technologies) for MS control and data analysis.
The CT-1128, which requires ultra-high purity hydrogen or helium (or nitrogen if desired) for the carrier gas,
can use either an external gas tank or its on-board hydrogen storage bottle. The metal hydride storage bottle
can be charged with hydrogen to provide a source of carrier gas that is convenient for mobile operation. The
mass spectrometer can be tuned using an internal calibrant such as perfluorotributylamine (PFTBA) to
perform a standard spectra tune or autotune (for maximum sensitivity over the entire scanning range)
depending on the user's needs. At the time of testing, the cost of the CT-1128 GC-MS system, with optional
SPME stirrer/heater, was $140,000.
VERIFICATION RESULTS
Summary of Accuracy, Precision, Linearity, and Stability
Contaminant
benzene
toluene
ethylbenzene
xylenes (total)
2,4-D
dicrotophos
GB
GD
VX
Accuracy
Mean
Percent
Recovery
(R)
172%
440%
104%
103%
62%
143%
108%
75%
109%
Precision
Mean
Relative
Standard
Deviation
(RSD)
10%
43%
16%
10%
21%
42%
24%
14%
15%
Linearity
Coefficient of
Determination
of Curve (r2)
1.000
1.000
1.000
1.000
0.921
0.999
1.000
1.000
0.959
Stability
Mean
Relative
Percent
Difference
(RPD)
27%
52%
9%
12%
35%
92%
48%
27%
27%
-------
Benzene accuracy was considerably higher than ideal (100%) at 172% recovery (R) because of a change in
response several days after establishing the calibration curve. Toluene exhibited significant over-recoveries,
with an overall mean R of 440%, though accuracy for ethylbenzene and xylenes (total) was close to 100%.
The mean R for 2,4-D and dicrotophos was 62% and 143%, respectively. For GB, the accuracy was close to
ideal at 108%. For GD, R was acceptable at 75. The mean R for VX was 109%, though the concentrations
tested were significantly higher than the LD50 for this agent. Precision, as measured by relative standard
deviation (RSD) of replicate samples, ranged from 10% for benzene to 43% for toluene.
In regard to linearity, the calibration curves of seven of the nine contaminants had coefficients of
determination (r2) of 0.999 or greater. The exceptions were those contaminants for which the provided method
lacked sensitivity—2,4-D and VX (r2 of 0.921 and 0.959, respectively), though for all contaminants, r2 values
were greater than 0.920. Instrument stability was evaluated by comparing the results of mid-level PT samples
at the beginning and end of the testing day and determining relative percent difference (RPD) of the PT
samples (ideal RPD is 0%). Stability results ranged from 9% RPD for ethylbenzene to 92% for dicrotophos.
Only two contaminants, ethylbenzene and xylenes (average RPD of 9% and 12%, respectively), had average
RPDs less than 20%, while toluene and dicrotophos had RPDs significantly greater than 20% (52% and 92%,
respectively).
Summary of Matrix Effects Observed
Contaminant
benzene
toluene
ethylbenzene
xylenes
2,4-D
dicrotophos
GB
GD
VX
Matrix Effect(a) from Potential Interferents ^ = observed
DW1
v'
•/
ND
DW2
v'
v'
v'
ND
DW3
v'
v'
v'
v'
v'
ND
DW4
v'
v'
•/
ND
Weakly
Buffered
Water
•/
v'
v'
ND
Strongly
Buffered
Water
•/
v'
v'
v'
•/
v'
v'
v'
ND
THMs
Spiked
Water
v'
v'
ND
{?L} matrix effect defined as recovery ± 30% of average response of daily PT samples fortified at the same concentration
ND = no data; VX matrix testing was not performed due to lack of sensitivity for this contaminant using the vendor-provided method
A matrix effect was present with the strong buffer matrix, which gave R values outside the 70% to 130%
range for all eight of the target contaminants tested. DW3 also showed matrix effects for five of the eight
contaminants, which may be due to its origin as a groundwater sample.
-------
With the exception of VX and 2,4-D (for which the provided methods lacked sensitivity), the sensitivity of the
CT-1128 was sufficient to detect the target contaminants at the concentrations of interest (i.e., LD50 or MCL
concentrations).
Results of CT-1128 Sensitivity Testing for Target Contaminants
Contaminant
benzene
toluene
ethylbenzene
xylenes (total)
2,4-D
dicrotophos
GB
GD
VX
Concentration
of Interest
(mg/L)
0.005
1
0.7
10
0.07
1400
20
1.4
2.1
Sufficient Sensitivity
to Detect Cone, of
Interest
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
No
Field portability and operational factors. Because the CT-1128 requires time for thermal equilibration once
electrical power and gas have been supplied, it should be kept on standby (under vacuum and thermally
equilibrated) as long as possible when time is a critical factor for analyzing field samples. Mobilization in the
field is straightforward, and the CT-1128 requires only a source of electrical power for several hours of field
deployment when used with its on-board hydrogen canister for a source of carrier gas. Typical extraction and
sample run times ranged from 22 minutes to 32 minutes. Average sample throughput during verification
testing was 11 samples per 10-hour working day, or approximately one sample per hour. For 100 samples, the
total cost for supplies was approximately $914, not including the GC column and standard chemicals.
Original signed by Gregory A. Mack
Gregory A. Mack Date
Vice President
Energy, Transportation, and Environment Division
Battelle
5/24/06 Original signed by Andrew P. Avel
Andrew P. Avel
Acting Director
National Homeland Security Research Center
Office of Research and Development
U.S. Environmental Protection Agency
7/3/06
Date
NOTICE: ETV verifications are based on an evaluation of technology performance under specific,
predetermined criteria and the appropriate quality assurance procedures. EPA and Battelle make no expressed or
implied warranties as to the performance of the technology and do not certify that a technology will always
operate as verified. The end user is solely responsible for complying with any and all applicable federal, state,
and local requirements. Mention of commercial product names does not imply endorsement.
-------
May 2006
Environmental Technology Verification
Report
ETV Advanced Monitoring Systems Center
Constellation Technology Corporation
CT-1128 Portable
Gas Chromatograph-Mass Spectrometer
by
Raj Mangaraj
Amy Dindal
Zachary Willenberg
Karen Riggs
Battelle
Columbus, Ohio 43201
-------
Notice
The U.S. Environmental Protection Agency (EPA), through its Office of Research and
Development, has financially supported and collaborated in the extramural program described
here. This document has been peer reviewed by the Agency. Mention of trade names or
commercial products does not constitute endorsement or recommendation by the EPA for use.
11
-------
Foreword
The U.S. Environmental Protection Agency (EPA) is charged by Congress with protecting the
nation's air, water, and land resources. Under a mandate of national environmental laws, the
Agency strives to formulate and implement actions leading to a compatible balance between
human activities and the ability of natural systems to support and nurture life. To meet this
mandate, the EPA's Office of Research and Development provides data and science support that
can be used to solve environmental problems and to build the scientific knowledge base needed
to manage our ecological resources wisely, to understand how pollutants affect our health, and to
prevent or reduce environmental risks.
The Environmental Technology Verification (ETV) Program has been established by the EPA to
verify the performance characteristics of innovative environmental technology across all media
and to report this objective information to permitters, buyers, and users of the technology, thus
substantially accelerating the entrance of new environmental technologies into the marketplace.
Verification organizations oversee and report verification activities based on testing and quality
assurance protocols developed with input from major stakeholders and customer groups
associated with the technology area. ETV consists of six environmental technology centers.
Information about each of these centers can be found on the Internet at http://www.epa.gov/etv/.
Effective verifications of monitoring technologies are needed to assess environmental quality
and to supply cost and performance data to select the most appropriate technology for that
assessment. Under a cooperative agreement, Battelle has received EPA funding to plan,
coordinate, and conduct such verification tests for "Advanced Monitoring Systems for Air,
Water, and Soil" and report the results to the community at large. Information concerning this
specific environmental technology area can be found on the Internet at http://www.epa.gov/etv/
centers/center 1 .html.
in
-------
Acknowledgments
The authors wish to acknowledge the support of all those who helped plan and conduct the
verification test, analyze the data, and prepare this report. Many thanks go to Battelle's
Hazardous Materials Research Center for providing the facilities for and personnel capable of
working with chemical warfare agents. We sincerely appreciate the contribution of drinking
water samples from the Metropolitan Water District of Southern California (Paul Rochelle and
Melinda Stalvey), The New York Department of Environmental Protection (Virginia Murray),
and Orange County Utilities, Orlando, Florida (Theresa Slifko and Liza Robles). We would also
like to thank Matthew Magnuson (US EPA, National Homeland Security Research Center), Lisa
Olsen (U.S. Geological Survey), Victor Silvestri (New York City Department of Environmental
Protection) and Lynn Wright (US EPA, National Exposure Research Laboratory) for their
careful review of the test/quality assurance plan and this verification report.
IV
-------
Contents
Notice ii
Foreword iii
Acknowledgments iv
List of Abbreviations viii
Chapter 1 Background 1
Chapter 2 Technology Description 2
Chapters Test Design 5
3.1 Test Samples 6
3.1.1 Performance Test (PT) Samples 6
3.1.2 Potential Matrix and Interference Effects Samples 7
3.1.3 Quality Control Samples 10
3.2 Testing Procedure 10
3.2.1 Laboratory Testing 10
3.2.2 Non-Laboratory Testing 12
3.3 Reference Methods 12
3.3.1 CWAReference Method 13
3.3.2 Pesticide Reference Method 13
3.3.3 BTEX Reference Method 13
Chapter4 Quality Assurance Quality Control 14
4.1 Audits 15
4.1.1 Performance Evaluation Audit 15
4.1.2 Technical Systems Audit 16
4.1.3 Audit of Data Quality 16
4.2 QA/QC Reporting 16
4.3 Data Review 16
Chapter 5 Statistical Methods and Reported Parameters 18
5.1 Accuracy 18
5.2 Precision 18
5.3 Linearity 19
5.4 Sensitivity 19
5.5 Instrument Stability 19
5.6 Potential Matrix and Interference Effects 20
5.7 Field Portability 20
5.8 Operational Factors 20
Chapter 6 Test Results 21
6.1 Accuracy 21
6.2 Precision 26
6.3 Linearity 27
6.4 Sensitivity 32
-------
6.5 Instrument Stability 32
6.6 Potential Matrix and Interference Effects 33
6.7 Field Portability 34
6.8 Operational Factors 36
Chapter 7 Performance Summary 38
Chapters References 41
Figures
Figure 2-1. Constellation Technology Corporation CT-1128 Portable GC-MS 4
Figure 6-la. Library Match for GB without Background Subtraction 24
Figure 6-lb. Library Match for GB with Background Subtraction 25
Figure 6-2a. Calibration Curve for Benzene 28
Figure 6-2b. Calibration Curve for Toluene 28
Figure 6-2c. Calibration Curve for Ethylbenzene 29
Figure 6-2d. Calibration Curve for Xylenes (Total) 29
Figure 6-2e. Calibration Curve for 2,4-D 30
Figure 6-2f. Calibration Curve for Dicrotophos 30
Figure 6-2g. Calibration Curve for GB 31
Figure 6-2h. Calibration Curve for GD 31
Figure 6-2i. Calibration Curve for VX 32
Figure 6-3. Use and Installation of CT-1128 Outside of the Laboratory 35
Tables
Table 2-1. CT-1128 Technical Specifications (from Constellation Technology Corporation)(a) 4
Table 3-1. Target Contaminants and Concentrations of Interest 6
Table 3-2. Performance Test Sample Solution Concentrations for Target Contaminants 7
Table 3-3. Potential Matrix and Interference Spiking Concentrations 8
Table 3-4. Treatment and Source Characteristics of Drinking Water Samples 8
Table 3-5. Physio-Chemical Characterization of Drinking Water Samples 9
Table 3-6. CT-1128 Analytical Methods for Target Contaminants 11
Table 3-7. Reference Methods for Target Contaminants 12
Table 4-1. Changes to Test/QA Plan for Verification of Mobile Mass Spectrometers 14
Table 4-2. Summary of Reference Measurement Percent Recovery (R) 15
VI
-------
Table 4-3. Summary of Data Recording Process 17
Table 6-1 a. Accuracy (Percent Recovery) Results - Benzene 22
Table 6-lb. Accuracy (Percent Recovery) Results - Toluene 22
Table 6-lc. Accuracy (Percent Recovery) Results - Ethylbenzene 22
Table 6-ld. Accuracy (Percent Recovery) Results - Xylenes (Total) 23
Table 6-le. Accuracy (Percent Recovery) Results - 2,4-D and Dicrotophos 23
Table 6-lf. Accuracy (Percent Recovery) Results - GB and GD 23
Table 6-lg. Accuracy (Percent Recovery) Results - VX 26
Table 6-2. Summary of RSD of the Various Sample Types(a) 26
Table 6-3. Summary of Calibration Curve Data 27
Table 6-4. Sensitivity of CT-1128 32
Table 6-5. Results of Stability Testing for CT-1128 33
Table 6-6. Results of Potential Matrix and Interference Effects 34
Table 6-7. Field Portability Results 36
Table 7-1. Summary of Accuracy, Precision, Linearity, and Stability 38
Table 7-2. Summary of Matrix Effects Observed 39
Table 7-3. Results of CT-1128 Sensitivity Testing for Target Contaminants 39
vn
-------
List of Abbreviations
AMS Advanced Monitoring Systems
ASTM American Society for Testing and Materials
BTEX benzene, toluene, ethylbenzene, and xylenes
CWA chemical warfare agent
DW drinking water
El electron impact
EPA U.S. Environmental Protection Agency
ETV Environmental Technology Verification
GC-MS gas chromatograph-mass spectrometer
kg kilogram
m meter
MCL maximum contaminant level
mg milligram
uMHO micromhos
mL milliliter
NaCl sodium chloride
R recovery (percent)
PE performance evaluation
PFTBA perfluorotributylamine
ppb part per billion
ppm part per million
psi pounds per square inch
PT performance test
QA quality assurance
QC quality control
QMP quality management plan
RPD relative percent difference
RSD relative standard deviation
SEVI selected ion mode
SPME solid phase microextraction
THMs trihalomethanes
TIC total ion chromatogram
TSA technical systems audit
Vlll
-------
Chapter 1
Background
The U.S. Environmental Protection Agency (EPA) supports the Environmental Technology
Verification (ETV) Program to facilitate the deployment of innovative environmental
technologies through performance verification and dissemination of information. The goal of the
ETV Program is to further environmental protection by accelerating the acceptance and use of
improved and cost-effective technologies. ETV seeks to achieve this goal by providing high-
quality, peer-reviewed data on technology performance to those involved in the design,
distribution, financing, permitting, purchase, and use of environmental technologies.
ETV works in partnership with recognized testing organizations; with stakeholder groups
consisting of buyers, vendor organizations, and permitters; and with the full participation of
individual technology developers. The program evaluates the performance of innovative
technologies by developing test plans that are responsive to the needs of stakeholders,
conducting field or laboratory tests (as appropriate), collecting and analyzing data, and preparing
peer-reviewed reports. All evaluations are conducted in accordance with rigorous quality
assurance (QA) protocols to ensure that data of known and adequate quality are generated and
that the results are defensible.
The EPA's National Exposure Research Laboratory and its verification organization partner,
Battelle, operate the Advanced Monitoring Systems (AMS) Center under ETV. The AMS Center
recently evaluated the performance of the Constellation Technology Corporation CT-1128
Portable Gas Chromatograph-Mass Spectrometer (GC-MS). Mobile mass spectrometers were
identified as a priority technology verification category through the AMS Center stakeholder
process.
-------
Chapter 2
Technology Description
The objective of the ETV AMS Center is to verify the performance characteristics of
environmental monitoring technologies for air, water, and soil. This verification report provides
results for the verification testing of the CT-1128 Portable GC-MS. Following is a description of
CT-1128, based on information provided by the vendor, Constellation Technology Corporation.
The information provided below was not verified in this test.
The CT-1128 (Figure 2-1)
analyzes, on-site, known
and unknown chemicals.
The CT-1128 is a
lightweight, ruggedized,
field deployable GC-MS
system that can
accommodate the
applications of traditional
laboratory based GC-MS
systems. With the appro-
priate extraction tech-
niques, analysis may be
performed on a variety of
matrices including drinking
water (DW), which can be
prepared using solid phase
microextraction (SPME).
Figure 2-1. Constellation Technology Corporation CT-1128
Portable GC-MS
Technical specifications of the CT-1128 are presented in Table 2-1. The CT-1128 weighs
approximately 75 pounds (34 kg) and is 15 inches (38.1 cm) by 23 inches (58.4 cm) by 15 inches
(38.1 cm). It is contained in a carrying case housing the entire system. The CT-1128 has a range
of 1.6 to 800 atomic mass units with unit resolution throughout the mass range. In selected ion
mode (SIM), the CT-1128 can scan for 50 groups of masses with 30 masses per group. For
identification of chemicals, the CT-1128 is equipped with an automated mass spectral data base
searching function that can use a range of commercial mass spectral libraries (e.g., National
Institute of Standards and Technology Mass Spectral Library) as well as user-defined libraries.
-------
The system is controlled with a laptop computer that uses a program for GC control and MSD
Chemstation (Agilent Technologies) for MS control and data analysis.
The CT-1128, which requires ultra-high purity hydrogen or helium (or nitrogen if desired) for
the carrier gas, can use either an external gas tank or its on-board hydrogen storage bottle. The
metal hydride storage bottle can be charged with hydrogen to provide a source of carrier gas that
is convenient for mobile operation. The mass spectrometer can be tuned using an internal
calibrant such as perfluorotributylamine (PFTBA) to perform a standard spectra tune or autotune
(for maximum sensitivity over the entire scanning range) depending on the user's needs. At the
time of testing, the cost of the CT-1128 GC-MS system, with optional SPME stirrer/heater, was
$140,000.
-------
Table 2-1. CT-1128 Technical Specifications (from Constellation Technology
Corporation)(a)
INSTRUMENT PARAMETER
SPECIFICATION
Gas Chromatograph
GC Injection Port
GC Oven Ramp Speed
GC Oven / Injection Port Temperature Range
Carrier Gas Pressure Capacity
Carrier Gas Compatibility
Carrier Gas Source
Injection Mode
GC Column
Varian, model CP-1177; direct injection
Upto60°C/minto325°C
Ambient to 325°C, oven with double ramp capability
< 200 pounds per square inch (psi) at gas inlet, 0-40 psi variable
head pressure to control carrier gas flow through column
Helium, hydrogen, or nitrogen
Internal hydrogen bottle (-160 operating hours) or external gas
tank (He, H2, N2 (although N2 is rarely used))
Split or splitless
Available from a variety of GC column vendors including
Quadrex Corp.
Mass Spectrometer
Quadrupole Mass Spectrometer
Mass Range Capability
Mass Resolution
Scan Speed
Detector Dynamic Range
Mass Axis Stability
SIM Capability
SIM Dwell Time
lonization Mode/Energy/Current
Ion Source Temperature
Quadrupole Temperature
El SIM Sensitivity
Software
MS Data Base
MS Tuning Capability
Agilent Technologies, model 5973N MSD
1.6-800m/z, O.lm/z steps
Unit mass resolution throughout mass range
Up to 10,400 amu per second
Total analog/digital converter (ADC) = 106
+0.15 amu over 12 hours
Up to 50 groups of masses, with 30 masses per group, can be
varied throughout a run
10 - 9,999 msec/mass
El (electron impact), voltage/current user-selectable (5-240 eV),
dual filaments
Upto250°CforEI
Up to 200°C, independent of ion source
RMS signal/noise ratio at 272.0 m/z > 10: 1 for 20 fg
octafluoronaphthalene
MSD Chemstation
National Institute of Standards and Technology /EPA/National
Institutes of Health comprehensive library
Multiple autotune algorithms available, compound-specific
autotunes available forPFTBA, BFB, DFTPP, etc
CT-1128 GC-MS System
Dimensions (L x W x H)
Weight
Operating Conditions
Power Requirements
Vacuum System
Operating System
Detection Limits
38. 1 cm x 58.4 cm x 38.1 cm (15" x 23" x 15")
~ 34 kg (75 pounds)
10-35°C, 5-95% humidity
99-127 volts, single phase; A/C; 48-66 Hz
1 Diaphragm rough pump, 2 turbomolecular pumps
Windows 2000
-100 pg hexachlorobenzene; most compounds have detection
limits in the picogram range assuming scanning ion mode
(a) these specifications were not verified during testing; analytical conditions used during testing are provided in
Chapter 3
-------
Chapter 3
Test Design
Many volatile and semivolatile contaminants in water are detected using bench-top mass
spectrometers in a traditional laboratory setting. However, the CT-1128 verified in this test was a
portable unit designed to be taken outside the laboratory setting for field analysis. This
portability offers an advantage to first-responders and other users who need chemical
information when time, sampling, and other limitations preclude analysis in a laboratory.
This verification was conducted from September through December 2005 according to
procedures specified in the Test/QA Plan for Verification of Mobile Mass Spectrometer. (1-) The
ability of the CT-1128 to identify and quantify target contaminants was tested in various water
matrices. The CT-1128 was evaluated for the following performance parameters:
• Accuracy
• Precision
• Linearity
• Instrument stability
• Potential matrix and interference effects
• Sensitivity
• Field portability
• Operational factors.
The testing was conducted on the contaminants listed in Table 3-1. Three classes of
contaminants were tested: volatile organic compounds (benzene, toluene, ethyl benzene, total
xylenes (BTEX)), pesticides (2,4-D and dicrotophos), and chemical warfare agents (VX, GB,
and GD). The contaminants were selected based on recommendations from the AMS Center
stakeholders.
-------
Table 3-1. Target Contaminants and Concentrations of Interest
Contaminant
VX
GB (sarin)
GD (soman)
dicrotophos
2,4-D
benzene
toluene
ethylbenzene
xylenes (total)
CAS#(a)
50782-69-9
107-44-8
96-64-0
141-66-2
94-75-7
71-43-2
108-88-3
100-41-4
1330-20-7
LD50 (mg/L)(b)
2.1
20
1.4
1,400
NA
NA
NA
NA
NA
Maximum Contaminant Level
(mg/L)(c)
NA
NA
NA
NA
0.07
0.005
1
0.7
10
(a) Chemical Abstracts Number
(b) LD50 values assume a 70-kilogram individual consuming 250 milliliters (mL) of the contaminated water
(c) Maximum Contaminant Level (MCL) as defined by EPA National Primary Drinking Water Regulations
NA = Not applicable
3.1 Test Samples
Samples were prepared daily from stock solutions to minimize loss of target contaminants due to
hydrolysis. For chemical warfare agent (CWA) testing, a stock solution containing both GB and
GD and another containing VX only were prepared in acetone. Stock solutions for the other
contaminants, one containing all BTEX compounds and another containing both 2,4-D and
dicrotophos, were prepared in methanol to minimize degradation. Each of the test sample types
are described in detail in this section.
3.1.1 Performance Test (PT) Samples
The performance test (PT) samples were prepared in American Society for Testing and Materials
(ASTM) Type II water. Table 3-2 shows the concentrations of the PT samples that were analyzed
for the target contaminants. The target contaminant concentrations were constructed to cover or
approximate the concentrations of interest presented in Table 3-1. The concentrations of interest
were calculated using LD50 values assuming a 70-kilogram individual consuming 250 milliliters
(mL) of the contaminated water. When LD50 data were not available or feasible for testing, MCL
values, as defined by EPA National Primary Drinking Water Regulations, were used. For 2,4-D
and VX, the levels of interest in Table 3-1 could not be reached due to the sensitivity of the
analytical methods provided by the vendor. For these contaminants, concentrations with
adequate instrumental response were determined and used for verification testing. Reference
measurements were conducted on PT samples only, to confirm the accuracy of sample
preparation. To avoid discrepancies due to contaminant degradation, reference measurements
were made as close as possible in time with the measurements made with the CT-1128.
-------
Table 3-2. Performance Test Sample Solution Concentrations for Target Contaminants
Class
CWA
Pesticide
BTEX
Contaminant
VX
GB
GD
dicrotophos
2,4-D
benzene
toluene
ethylbenzene
xylenes (total)
PT Sample
1
(mg/L)
10
0.5
0.5
5
0.5
0.01
0.001
0.001
0.001
PT Sample
2
(mg/L)
15
1
1
50
5
0.1
0.01
0.01
0.01
PT Sample
3
(mg/L)
30
10
10
500
50
1
0.1
0.1
0.1
PT Sample
4
(mg/L)
NA
NA
NA
1,000
75
10
1
1
1
PT Sample
5
(mg/L)
NA
NA
NA
NA
100
NA
NA
NA
NA
NA = Not applicable.
The PT samples were used to determine the accuracy of the CT-1128. One set of these PT
samples was used to establish a calibration curve. Subsequent analyses of the PT samples on
multiple testing days were then used to calculate the accuracy of the CT-1128 measurements.
Three replicate measurements were made for each PT sample for the accuracy testing, with the
exception of 2,4-D and dicrotophos. As shown in Tables 3-1 and 3-2, the tested concentrations
were considerably higher and lower than the levels of interest for 2,4-D and dicrotophos,
respectively, due to the analytical sensitivity of the CT-1128. Since a great deal of effort was
invested in determining the appropriate concentration levels for 2,4-D and dicrotophos, only one
replicate of each pesticide PT sample concentration was analyzed in order to analyze the rest of
the test samples prior to CWA testing.
3.1.2 Potential Matrix and Interference Effects Samples
To measure the potential matrix effects on the CT-1128 in selected real-world applications, it
was challenged by analyzing samples fortified with the target contaminant in various matrices
including regional DW samples, a weakly buffered water sample, a strongly buffered water
sample, and a trihalomethanes (THMs)-fortified water sample as shown in Table 3-3. The
concentration of a mid-level PT sample was used to fortify the matrix samples. This
concentration provided a convenient level that was at or below the concentration of interest for
most of the target contaminants. As shown in Table 3-3, matrix and interference effects were not
evaluated for VX, due to the relatively low sensitivity for this contaminant with the testing
methodology.
-------
Table 3-3. Potential Matrix and Interference Spiking Concentrations
Sample Type
DWl(b)
DW2(b)
DW3(b)
DW4(b)
Weakly Buffered
Water (442-3 0)(c)
Strongly Buffered
Water (442-3000)(c)
THMs Spiked Water
(ASTM Type II
Water)(c)
Fortified Concentration in Solution (mg/L)
Benzene
1
1
1
1
1
1
1
Toluene,
Ethylbenzene,
Xylenes
0.1
0.1
0.1
0.1
0.1
0.1
0.1
2,4-0
50
50
50
50
50
50
50
Dicrotophos
500
500
500
500
500
500
500
GB,
GD
1
1
1
1
1
1
1
VX(a)
NA
NA
NA
NA
NA
NA
NA
(a) VX testing was limited due to lack of sensitivity for this contaminant using the vendor-provided method.
NA = Not applicable.
(b) See Section 3.1.2.1 for identification of DW sources.
(c) See Section 3.1.2.2 for a description of these samples.
3.1.2.1 Drinking Water Samples
DW samples were collected from four geographically distributed municipal sources to evaluate
the performance of the CT-1128 with various sample matrices. These finished DW samples
varied in their source, treatment, and disinfection process. All samples underwent either
chlorination or chloramination disinfection prior to receipt. Samples were collected from water
utility systems with the treatment and source characteristics listed in Table 3-4.
Table 3-4. Treatment and Source Characteristics of Drinking Water Samples
Drinking Water
ID
DW1
DW2
DW3
DW4
Water Utility
Columbus, Ohio
(OH)
New York City, New York (NY)
Orlando, Florida
(FL)
Metropolitan Water District of Southern California
(CA)
Water
Treatment
chlorinated filtered
chlorinated
unfiltered
chlorinated filtered
chloraminated
filtered
Source
Type
surface
surface
ground
surface
All samples were collected in pre-cleaned high density polyethylene containers. After sample
collection, to characterize the DW matrix, an aliquot of each DW sample was sent to a
subcontract laboratory to determine the following water quality parameters: concentration of
THM, haloacetic acids, total organic halides, pH, conductivity, alkalinity, turbidity, organic
carbon, and hardness (see Table 3-5 for results of these analyses).
-------
Table 3-5. Physio-Chemical Characterization of Drinking Water Samples
Parameter
Turbidity
Dissolved Organic
Carbon
Total Organic
Carbon
Specific
Conductivity
Alkalinity
PH
Calcium
Magnesium
Hardness
Total Organic
Halides
Trihalomethanes
Haloacetic Acids
Unit
NTU(a)
mg/L
mg/L
uMHO
mg/L
mg/L
mg/L
mg/L
ug/L
ug/L/analyte
ug/L/analyte
Method
EPA 180.1(2)
SM5310(3)
SM5310(3)
SM2510(3)
SM 2320(3)
EPA 150.1(4)
EPA200.8(5)
EPA200.8(5)
EPA 130.2(6)
SM5320(3)
EPA 524.2(7)
EPA 552.2(8)
Sources of Drinking Water Samples
Columbus,
OH
(DW1)
0.1
2.1
2.1
572
40
7.6
33
7.7
118
220
74.9
32.8
New York
City, NY
(DW2)
1.1
1.1
1.6
84
14
6.9
5.6
1.3
20
82
39.0
39.0
Orlando,
FL
(DW3)
0.5
1.6
1.7
322
142
8.5
8.8
43
143
300
56.4
34.6
MWD00,
CA
(DW4)
0.1
2.9
2.5
807
71
8.0
45
20
192
170
39.2
17.4
(a) NTU = Nephelometric turbidity unit.
(b) MWD = Metropolitan Water District of Southern California
Because free chlorine degrades many of the target contaminants and interferents during storage,
the DW samples were immediately dechlorinated with sodium thiosulfate pentahydrate upon
arrival at Battelle. The dechlorination of the DW was qualitatively confirmed by adding a
diethyl-p-phenylene diamine tablet to an aliquot of the DW. If the water did not turn pink, the
dechlorination process was determined to be successful. If the water did turn pink, an additional
dechlorinating reagent was added and the dechlorination confirmation procedure was repeated.
The bulk DW samples were dechlorinated upon arrival and did not require any additional
dechlorination during testing.
3.1.2.2 Weakly Buffered, Strongly Buffered, and Trihalomethanes Matrices
The effect of ionic strength on the response of the CT-1128 was examined. Since natural water
salt type and concentration can vary greatly by location, two sample types were fortified at a mid
level PT sample concentration in 442 Natural Water™ Standard Solution (Myron L Instruments,
Carlsbad, CA). Two 442 solutions, 442-30 and 442-3000, containing 21.8 parts-per-million
(ppm) sodium chloride (NaCl) and 2,027 ppm NaCl, respectively, were used for this purpose.
The CT-1128 was also challenged by the presence of potential interferents. THMs are typically
observed at low levels in DW as by-products of the disinfection process. Four THMs
(chloroform, bromoform, bromodichloromethane, and dibromochloromethane) were spiked into
a midlevel PT sample at 80 parts per billion (ppb) total, which is the MCL for total THMs as
-------
defined in EPA's National Primary Drinking Water Regulations.(12) Chloroform, bromoform,
bromodichloromethane, and dibromochloromethane were spiked so that their concentrations in
solution were 50, 5, 15, and 10 ppb, respectively, to represent typical ratios of THMs in finished
DW.
3.1.3 Quality Control Samples
In addition to the PT, DW, buffered water, and THMs-fortified water samples, blanks and
unfortified matrix samples were analyzed to confirm negative responses in the absence of target
contaminants and also to ensure that no sources of contamination were introduced during the
analysis.
3.2 Testing Procedure
3.2.1 Laboratory Testing
Experienced GC-MS operators were used for testing since the vendor suggests that a new user
obtain training in the use of a GC-MS prior to operating the CT-1128. All of the operators were
trained laboratory staff and were experienced in the proper use and handling of laboratory
chemicals. Additionally, analyses of CWAs were handled by staff trained with specific skills
necessary for working in a CWA laboratory.
The vendor provided a sample preparation technique as well as analytical methodology for the
three groups of contaminants to be analyzed during the verification test: BTEX, pesticides, and
CWAs. Ultra-high purity (UHP) helium was used for the carrier gas. The vendor identified
SPME as the technique for preparing the water samples for subsequent GC-MS analysis by the
CT-1128. SPME is a relatively simple adsorption/desorption technique for extracting volatile and
semivolatile chemicals from a liquid sample or headspace. The technique employs the use of a
fused silica fiber that is coated with a polymer. The specific SPME fiber should be chosen based
on the application of interest. It is very important to note that the methodology provided by the
vendor was not optimized for any one specific target chemical. The same SPME fiber type and
GC column were used throughout the test for all analytes. The SPME fiber was replaced
approximately after every 50 extractions or at the discretion of the operator. Therefore, further
optimization of the extraction and separation/detection process may be possible.
The provided methodology specified daily check procedures that were followed each day before
the analysis of samples. These procedures included diagnostic checks for system operation (e.g.,
check of MS vacuum and temperature of heated zones), tuning of the MS, and column/SPME
fiber bakeout. The CT-1128 was mass tuned using PFTBA (as the internal calibrant) and the
autotune function. It should be noted that the method provided by the vendor originally specified
that the operator should perform MS autotune on a daily basis. According to the vendor, their
typical users, who are seeking a qualitative rather than quantitative result, start up and shut down
the CT-1128 on a regular basis as they transport the instrument from site to site. Since the
quantity of a particular contaminant may be a critical factor in determining the appropriate
corrective action(s), this verification test focused on the quantitative abilities of the mobile mass
spectrometer. For the verification test, the CT-1128 ran continuously once it was installed in a
10
-------
particular location. Since the autotune function will change the MS settings, any quantitative
data obtained after an autotune may not compare well to such data obtained prior to the autotune.
Since the operator intended to obtain one calibration curve and use it to quantify the results of
several days of subsequent testing, the vendor suggested performing an autotune only if the CT-
1128 had been shut down and restarted after a period of two hours or more. This change in the
methodology was requested by the vendor when the Verification Test Coordinator inquired about
the variation in instrument response over the course of several days of testing.
Table 3-6 summarizes the instrumental parameters for each of the three methods provided.
Testing solutions were prepared daily from stock solutions. To an empty vial, 15 mL of a test
water sample and a disposable magnetic stirbar were added before capping and crimping a seal
onto the vial. For all of the target contaminants analyzed in the verification test, the vendor
provided Supelco SPME fibers that were coated with 100 jim polydimethylsiloxane (Supelco
Catalog # 57300-U). The operator inserted the SPME fiber, which is housed in a syringe-type
assembly, into the liquid. After exposing the fiber for a specific amount of time, the operator
pulled the fiber back into the assembly, and then removed it from the sample. After inserting the
assembly into the injector port of the CT-1128, the operator exposed the fiber for desorption and
began the analysis. A SPME mixer heater compartment, which is built into the CT-1128 unit,
allowed the operator to keep the sample properly mixed and thermally equilibrated.
Table 3-6. CT-1128 Analytical Methods for Target Contaminants
Analytes
benzene
toluene
ethylbenzene
(total)
2,4-D
dicrotophos
GB
GD
VX
Quantitation
Ion
78m/z
91m/z
91m/z
91m/z
162 m/z
127 m/z
99 m/z
126 m/z
114 m/z
SPME
Extraction
luu um
polydimethylsilo
xane (Supelco
57300-U)
exposed for 20
minutes at 25°C
with mixer speed
— 3
GC Column
007-5MS,
Silpheylene
Silo xane, 30m x
0 25 mm ID x
0.25 urn film
(Quadrex
Corporation
007-5MS-30-
0.25F)
GC Oven
Program
35°C hold 2
minutes, ramp
to 200°C at
10°C/minute,
ramp to 300°C
at 20°C/minute
50°C hold 0.5
minutes, ramp
to 300°C at
15°C/minute,
hold 5 minutes
50°C hold 0.5
minutes, ramp
to 300°C at
15°C/minute,
hold 15 minutes
MS
Scan
Range
35 to
120 m/z
35 to
250 m/z
35 to
275 m/z
MS
Sampling
Rate
5
(2.2
scans/
second)
4(1.75
scans/
second)
4(1.58
scans/
second)
Total Run
Time
23.5
minutes
22.2
minutes
32 2
minutes
Like the choice of SPME fiber, the vendor provided a single GC capillary column (30 m x 0.25
mm ID x 0.25 um film, type 007-5MS; Quadrex Corporation, Woodbridge, CT) that was to be
11
-------
used for the target contaminants analyzed during the verification test. The operator exposed the
SPME fiber to the sample for a total absorption time of 20 minutes while the mixer speed was set
to "3." After removing the SPME assembly from the sample and inserting it into the injector
port of the CT-1128, the operator exposed the fiber, and selected the "Start" button in the GC
control software (which works in tandem with the MS control software). The operator retracted
and removed the SPME fiber after the end of the 10 minute desorption time. After data
acquisition, the operator processed the raw data by extracting the appropriate quantitation ion
from the total ion chromatogram (TIC) and then integrating the corresponding peak using the
Chemstation software. Using Microsoft Excel, the peak area (response) was plotted against
expected analyte concentration to generate calibration curves which were used to quantify the
subsequent sample analyses. Though a single calibration curve was acquired, a midlevel PT
sample was analyzed at the beginning and end of each day of testing to assess stability of the CT-
1128 response.
3.2.2 Non-Laboratory Testing
The CT-1128 was deployed into a non-laboratory setting to verify its performance in a relatively
uncontrolled environment. Using the operator's manual for guidance, the operator set up the CT-
1128 for analysis in a warehouse located on Battelle's Columbus campus. The site had electricity
though temperature and humidity were not as controlled as in a laboratory. For the non-
laboratory analysis, the operator followed the same procedures used in the laboratory and
analyzed BTEX calibration standards and a water sample that was fortified with BTEX (at
concentrations unknown to the operator).
3.3 Reference Methods
Laboratory reference methods (Table 3-7) were used to determine the accuracy of sample
preparation and demonstrate the stability of the target contaminants in the PT matrix. The
reference laboratories were asked to follow the sample preservation/handling and quality control
(QC) requirements specified in Chapter 4, in addition to any QC requirements specified in each
reference method.
Table 3-7. Reference Methods for Target Contaminants
Contaminant
VX
GB
GD
dicrotophos
2,4-D
benzene
toluene
ethylbenzene
xylenes (total)
Reference Method
Battelle Internally Developed Method(a)
HMRC-IV-118-05(9)
HMRC-IV-118-05(9)
Battelle Internally Developed Method(b)
U.S.EPA515.1(10)
U.S. EPA 524.2(7)
U.S. EPA 524.2(7)
U.S. EPA 524.2(7)
U.S. EPA 524.2(7)
(a) VX analysis by Liquid Chromatography-Tandem Mass Spectrometry (LC-MS/MS).
(b) Dicrotophos analysis by GC-MS.
HMRC = Hazardous Materials Research Center
12
-------
3.3.1 CWA Reference Method
Samples submitted for GB and GD determination were analyzed by Battelle according to
procedures outlined in HMRC-IV-118-05. This procedure involves a liquid-liquid extraction of
the water sample and subsequent analysis of the extract by GC-MS. The samples containing VX
were analyzed directly in water using an internally developed electrospray ionization liquid-
chromatography tandem mass spectrometry (LC-MS/MS) method.
3.3.2 Pesticide Reference Method
Samples for 2,4-D determination were analyzed according to EPA Method 515.1 in which
chlorinated acids are extracted and derivatized. The derivatives are then determined by GC using
an electron capture detector. Samples for dicrotophos determination were analyzed according to
an internal Battelle method. This procedure involves solid phase extraction followed by GC-MS
analysis. Battelle performed this method when it was learned that the subcontract laboratory that
was going to perform EPA Method 8141, as proposed in test/QA plan for this test,(1) no longer
offered this analysis.
3.3.3 BTEXReference Method
Samples submitted to a subcontract laboratory for BTEX determination were analyzed according
to EPA 524.2. This "purge and trap" method involves the purging of volatile organic compounds
with low water solubility and trapping these compounds onto a sorbent tube. The compounds are
then thermally desorbed and determined by GC-MS.
13
-------
Chapter 4
Quality Assurance Quality Control
QA/QC procedures were performed in accordance with the quality management plan (QMP) for
the AMS Center (11) and the test/QA plan for this verification test(1) except as noted in Table 4-1.
Table 4-1. Changes to Test/QA Plan for Verification of Mobile Mass Spectrometers
Change
Proposed PT sample concentrations, as listed in test/QA plan, (1) were changed after
running preliminary solutions to determine analytical sensitivity.
Three replicate measurements of the PT samples were not made for two analytes: 2,4-D
and dicrotophos.
Accuracy was determined using the theoretical concentration of target analyte instead of
the concentration of the target analyte determined by the reference measurement.
Research trailer was not available during time of field testing and one PT sample was
analyzed (instead of triplicate analysis of a raw water sample).
The reference methods for VX and dicrotophos were changed.
Reference samples were submitted for analysis to demonstrate that the PT samples can
be prepared accurately. In particular, BTEX and 2,4-D reference methods were most
problematic.
Only PT samples were analyzed for VX.
Relative percent difference (RPD) was used to measure instrument stability instead of
percent recovery (R) of PT samples analyzed on each day of testing.
Further
Discussion
Section 6.8
Section 3.1.1
Section 4. 1.1
Section 6.7
Section 3.3
Section 4.1.1
Section 6.1
Section 5.5
14
-------
A summary of the recovery values (R) for the reference measurements is presented in Table 4-2
(a definition of R is included in Section 5.1). R values for benzene, dicrotophos, and the CWAs
were all 80% or greater. Sample degradation was suspected as the cause for lower recoveries
observed with the other chemicals.
Table 4-2. Summary of Reference Measurement Percent Recovery (R)
Contaminant
VX
GB
GD
dicrotophos
2,4-D
benzene
toluene
ethylbenzene
xylenes (total)
R of Reference
Measurements
124%
80%
87%
99%
64%
86%
52%
46%
52%
Reference analyses for BTEX and 2,4-D did not meet acceptance criteria for all samples. The
EPA-specified holding time for laboratories using EPA method 524.2 for BTEX analysis is 14
days. Several of the samples could not be confirmed to have been analyzed within this specified
holding time due to discrepancies in the chain of custodies. Also, a number of results were
received with low surrogate recoveries. In addition, as shown in Table 4-2, sample recovery was
near 50% for all BTEX analytes except benzene which was 86%. Of 11 samples submitted for
2,4-D analysis, 7 samples were analyzed after the recommended 14 day holding time per EPA
method 515.1.
Based on these deficiencies in the reference analyses and since test samples were prepared daily,
these reference results were not used to determine accuracy of the CT-1128 GC-MS analytical
measurement, but rather to confirm that sample preparation was being performed properly.
Since sample preparation procedures were standardized and no systematic errors were observed
in the reference measurements, the results provide confidence in the sample preparation. There
is no impact on the data because it is a common practice to use the theoretical fortified value for
comparison.
4.1 Audits
A performance evaluation audit, a technical systems audit, and an audit of data quality were
performed for this verification test.
4.1.1 Performance Evaluation Audit
For all contaminants, "blind" samples were submitted to analysts performing the reference
measurements. Since the methods are specific to the contaminant, only the concentration of
target contaminants was not disclosed when submitting samples for reference measurements.
15
-------
These performance evaluation (PE) samples were used to assess the accuracy of the reference
measurements and were prepared in accordance with the stated detection limits of the reference
laboratories. At least one PE sample was submitted per reference method prior to the start of the
verification test and once during the verification test. PE samples that were not within ± 20% of
the expected result were repeated.
4.1.2 Technical Systems Audit
The Battelle Quality Manager conducted a technical systems audit (TSA) to ensure that the
verification test was performed in accordance with the test/QA plan(1) and the AMS Center
QMP.(11) As part of the audit, the Battelle Quality Manager reviewed the reference methods,
compared actual test procedures to those specified in the test/QA plan, and reviewed data
acquisition and handling procedures. Observations and findings from this audit were documented
and submitted to the Battelle Verification Test Coordinator for response. No findings were
documented that required any significant action. The records concerning the TSA are
permanently stored with the Battelle Quality Manager.
4.1.3 Audit of Data Quality
At least 10% of the data acquired during the verification test was audited. Battelle's Quality
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
data undergoing the audit were checked.
4.2 QA/QC Reporting
Each internal assessment and audit was documented in accordance with Sections 3.3.4 and 3.3.5
of the QMP for the ETV AMS Center.(11) Once the assessment report was prepared, the Battelle
Verification Test Coordinator ensured that a response was provided for each adverse finding or
potential problem and implemented any necessary follow-up corrective action. The Battelle
Quality Manager ensured that follow-up corrective action was taken. The results of the TSA
were sent to EPA.
4.3 Data Review
Records generated in the verification test were reviewed before these records were used to
calculate, evaluate, or report verification results. Table 4-3 summarizes the types of data
recorded. The review was performed by a technical staff member involved in the verification
test, but not the staff member who originally generated the record. The person performing the
review added his/her initials and the date to a hard copy of the record being reviewed.
16
-------
Table 4-3. Summary of Data Recording Process
Data to be Recorded
Dates, times, and
details of test events,
mass spectrometer
maintenance, down
time, etc.
Mass spectrometer
calibration
information
Mass spectrometer
readings
Reference method
sample preparation
Reference method
procedures,
calibrations, QA, etc.
Reference method
analysis results
Responsible
Party
Battelle
Battelle
Battelle
Battelle
Battelle or
subcontract
laboratory
Battelle or
subcontract
laboratory
Where Recorded
ETV laboratory
record books or
data recording
forms
ETV laboratory
record books or
electronically
Recorded
electronically by
each instrument
and then
downloaded daily
Laboratory record
books
Laboratory record
books, or data
recording forms
Electronically
from reference
analytical method
How Often
Recorded
Start/end of test
procedure, and at
each change of a test
parameter or change
of instrument status
At instrument
calibration or
recalibration
For each sample
Throughout sample
preparation
Throughout sampling
and analysis
processes
Every sample
analysis
Disposition of Data
Used to organize/check test
results; manually incorporated in
data spreadsheets as necessary
Incorporated in verification report
as necessary
Converted to spreadsheet for
statistical analysis and
comparisons
Used to demonstrate validity of
samples submitted for reference
measurements
Retained as documentation of
reference method performance
Converted to spreadsheets for
calculations
17
-------
Chapter 5
Statistical Methods and Reported Parameters
The statistical methods presented in this chapter were used to verify the performance parameters
listed in Chapter 3.
5.1 Accuracy
Determination of accuracy of the PT samples was based on the concentration of the prepared
standards. A minimum of three concentrations of PT samples was used to establish the
calibration curve. All subsequent analyses of the PT samples were used to determine the
accuracy of the CT-1128 relative to the prepared concentrations. Ideally, independent reference
measurements would be used to confirm the accuracy of the technology. However, as discussed
in Chapter 4, the reference laboratory measurements (particularly in the case of BTEX and 2,4-
D) were of questionable quality due to analytical holding times that were exceeded and
inadequate QC results. Accuracy of the CT-1128 was therefore determined for all target
contaminants using the prepared concentrations rather than the reference laboratory
concentrations. The subsequent analyses of the PT samples for determination of accuracy
included two types: 1) analysis of an additional two replicate sets of PT samples (except for 2,4-
D and dicrotophos where, as described in Section 3.1.1, the calibration curve PT samples were
not repeated) and 2) daily verifications of the mid-level PT samples. The accuracy of the CT-
1128 was assessed as recovery (R), using Equation 1:
R = (Y, + X^xlOO (1)
where Yt is the concentration of target contaminant /' as measured by the CT-1128 and Xt is the
prepared concentration of the contaminant /'. R values are presented for PT samples. The ideal R
value is 100%. The mean, median, maximum, and minimum R values are presented, as well as
the standard deviation of the R values.
5.2 Precision
When possible, the relative standard deviation (RSD) of three replicate measurements of the PT
samples across three or four concentrations was determined to assess the precision of the CT-
18
-------
1128 measurements. Precision was also evaluated on replicate measurements of mid-level PT
samples that were used as daily calibration verifications. Lastly, the precision of replicate matrix-
fortified samples was calculated. Regardless of the type of sample analyzed, the precision was
defined as the RSD of replicate measurements, using Equation 2:
RSD = (SD + Yi)xlOO (2)
where Yt is the average concentration of target contaminant /' as measured by the CT-1128, and
SD, the standard deviation of the results. Ideal RSD values are near 0%.
5.3 Linearity
Linearity was determined by performing a linear regression analysis of the instrument response
for one set of the PT samples versus the theoretical target contaminant concentrations across a
minimum of three concentrations. The linear regression analysis used the target contaminant
concentration as the independent variable and the response (integrated peak area for the
contaminant of interest) from the CT-1128 as the dependent variable. Linearity was expressed in
terms of slope, intercept, and coefficient of determination (r2), in addition to presenting a plot of
the data. The ideal value for r2 is 1.
5.4 Sensitivity
Sensitivity was assessed by the ability of the CT-1128 to measure the target contaminant at or
below the concentration of interest. This evaluation did not focus on determining instrumental
detection limits since the purpose was to ascertain whether the CT-1128 was sensitive enough to
measure the target contaminants at the concentrations of interest in the various water matrices.
Serial ten-fold dilutions of the lowest concentration PT sample were performed until no change
in instrument response for the target contaminant was observed. It was also noted whether the
instrument response changed accordingly (e.g., ten-fold decrease for each ten-fold dilution).
5.5 Instrument Stability
Instrument stability (S) was determined by analyzing a PT sample at the beginning and end of
the analytical sequence (usually over the course of a day) to assess the degree of calibration
stability. The result of the PT sample at the end of the sequence was compared to that of the PT
sample analyzed at the start of the analytical sequence using Equation 3:
RPD = [ (Yt2 - El) •*• average (Ytl, Y,2)] x 1 00 (3 )
where Y-2 and Y-l are the results for the last and first PT sample, respectively, for target
contaminant /'. The ideal value for relative percent difference (RPD) is 0%.
19
-------
5.6 Potential Matrix and Interference Effects
Potential matrix and interference effects were assessed by comparing the response of the various
test matrices fortified with the target contaminants to the average response of the PT samples
analyzed on the day of testing. In the absence of instrument drift and target contaminant
degradation, it was assumed that a matrix or interferent effect was responsible for any change in
performance for the mid-level PT sample. The matrix effects were calculated using the percent
recovery calculation. R values between 70% and 130% indicated that a matrix effect was not
observed, while R values outside of that range were considered to indicate a matrix effect.
5.7 Field Portability
For this verification test, field portability was defined as the ability for a user to operate the CT-
1128 in a non-laboratory environment for sample analysis. Observations related to field
portability were observed and reported. These considerations included weight and dimensions of
unit, impact on user mobility, start-up time, and power requirements.
5.8 Operational Factors
Operational factors such as maintenance needs, data output, consumables used, ease of use,
repair requirements, and sample throughput were evaluated based on observations recorded by
Battelle staff. A separate laboratory record book was maintained for the CT-1128 and was used
to enter daily observations on these factors. Examples of information to be recorded in the record
book include the status of diagnostic indicators for the CT-1128, use or replacement of any
consumables, the effort or cost associated with maintenance or repair, vendor effort (e.g., time on
site) for repair or maintenance, the duration and causes of any instrument down time or data
acquisition failure, and operator observations about ease of use of the CT-1128. These
observations were summarized to aid in describing CT-1128 performance.
The time required for each sample from the start of sample preparation to reporting of results
defined sample time. The number of samples that could be analyzed per unit time defined sample
throughput. The sample throughput was noted for laboratory and field portions of testing.
20
-------
Chapter 6
Test Results
6.1 Accuracy
Accuracy (Tables 6-la-g), expressed as percent recovery (R), was determined using the equation
given in Section 5.1. This calculation was only performed on PT samples since these samples
represent the matrix free of any potential interference (ASTM Type IIDI water). In all cases,
with the exception of toluene, the integrated peak areas of the first replicate set of three or four
PT sample concentrations including the blank when possible (all cases except for CWAs where
no blanks were analyzed) were used to generate a linearly regressed calibration curve on which
quantitation was based. For toluene, the third replicate set was used to generate the curve due to
non-linear response of the lower standards on the first two replicates. In addition to presenting
the mean R for each PT level (when more than one PT level was tested) the median, maximum,
and minimum R are reported along with the standard deviation of the R values in Tables 6-la-g.
For observations of less than three, standard deviation of R was not determined (as indicated by
"nd"). Overall statistics that combine all R values across all concentrations are also presented to
summarize the accuracy results for each target analyte. Note that the number of observations for
the mid-level PT samples (i.e., PT 2 or PT 3) is considerably higher than that of the other
concentration levels due to the periodic calibration checks that were performed daily.
21
-------
Table 6-1 a. Accuracy (Percent Recovery) Results - Benzene
Benzene PT Level
PT1 (O.Olmg/L)
PT2(0.10mg/L)
PT3(1.00mg/L)
PT4(10.00mg/L)
benzene ALL
Observations
4
3
25
3
35
Mean
168%
116%
189%
94%
172%
Median
126%
115%
135%
91%
125%
Maximum
305%
124%
399%
103%
399%
Minimum
115%
109%
81%
87%
81%
SD
91%
7%
111%
8%
103%
For benzene, accuracy is considerably higher than ideal (100%) at 172% overall mean R. This is
primarily due to a change in response for benzene observed several days into testing after
establishing the calibration curve. It could not be determined whether the change was due to the
SPME extraction of benzene or the instrumental response for benzene. The response for a
particular analyte may change over the course of several days and may be due to a number of
factors including the extraction technique or environmental conditions; in such a case, it is
usually left to the operator's discretion whether to reacquire a calibration curve. It is important to
note that this procedure was followed throughout testing (i.e., one calibration curve for each
analyte was used to quantify results from subsequent testing). The high degree of variability in
the benzene response throughout testing using the single curve is reflected by the SD of the R.
Table 6-1 b. Accuracy (Percent Recovery) Results - Toluene
Toluene PT Level
PT 1 (0.001 mg/L)
PT2(0.01mg/L)
PT 3 (0.10 mg/L)
PT 4 (1.00 mg/L)
toluene ALL
Observations
4
3
25
3
35
Mean
299%
261%
508%
243%
440%
Median
251%
94%
611%
126%
489%
Maximum
533%
599%
770%
489%
770%
Minimum
160%
91%
67%
114%
67%
SD
162%
292%
237%
213%
248%
Toluene was the least accurate of the BTEX compounds tested, exhibiting significant over-
recoveries with an overall mean R of 440%. The third set of replicates was used to generate the
calibration curve since it was the most linear of the replicate sets of PT samples. The high degree
of variability in the toluene response is reflected by the SD of the R.
Table 6-1 c. Accuracy (Percent Recovery) Results - Ethylbenzene
Ethylbenzene PT Level
PT 1 (0.001 mg/L)
PT 2 (0.01 mg/L)
PT 3 (0.10 mg/L)
PT4(1.00mg/L)(a)
ethylbenzene ALL
Observations
4
3
25
3
35
Mean
102%
99%
110%
58%
104%
Median
94%
89%
121%
55%
102%
Maximum
140%
130%
143%
65%
143%
Minimum
82%
79%
66%
54%
54%
SD
26%
27%
27%
6%
29%
(a) PT 4 not used in calibration curve due to suspected detector saturation
The CT-1128 exhibited accuracy for ethylbenzene that was usually in the range of 70-130%,
with a mean of 104%, and relatively small standard deviations of the R values. Ethylbenzene
demonstrated reduced accuracy at the PT 4 concentration level (1.00 mg/L), which may have
22
-------
been due to detector saturation. Therefore, PT 4 was not included in the calibration curve for this
analyte.
Table 6-1 d. Accuracy (Percent Recovery) Results - Xylenes (Total)
Xylenes (total) PT Level
PT 1 (0.001 mg/L)
PT2(0.01mg/L)
PT 3 (0.10 mg/L)
PT 4 (1.00 mg/L)
xylenes ALL
Observations
4
3
25
3
35
Mean
124%
89%
104%
77%
103%
Median
117%
80%
113%
74%
109%
Maximum
158%
113%
136%
84%
158%
Minimum
105%
73%
62%
73%
62%
SD
23%
21%
24%
6%
25%
To produce the total xylenes result, the operator added the response for both m- and p-xylene
(which were not resolved from each other) to that of o-xylene to produce a total response.
Though the PT 4 sample was excluded from the calibration curve for xylenes as it was in
ethylbenzene, accuracy was not affected to the same extent. Total xylenes R was close to ideal,
with a mean R value of 103% and a relatively low SD.
Table 6-1 e. Accuracy (Percent Recovery) Results - 2,4-D and Dicrotophos
2,4-D PT Level
PT 3 (50.0 mg/L)
Dicrotophos PT Level
PT 3 (500.0 mg/L)
Observations
8
Observations
8
Mean
62%
Mean
143%
Median
66%
Median
104%
Maximum
79%
Maximum
520%
Minimum
29%
Minimum
7%
SD
16%
SD
160%
Since only one replicate of each PT sample was analyzed for 2,4-D and dicrotophos, accuracy
was determined using the PT 3 check samples that were analyzed each day of testing. Accuracy
for 2,4-D was less than ideal as the method was not sensitive for this contaminant. For
dicrotophos, the PT 4 sample was excluded from the calibration curve as it did not give a linear
response with respect to the other PT samples. The dicrotophos response showed high variability
as reflected in the SD of the R.
Table 6-lf. Accuracy (Percent Recovery) Results - GB and GD
GB PT Level
PT 1 (0.5 mg/L)
PT 2 (1.0 mg/L)
PT 3 (10.0 mg/L)
GBALL
GD PT Level
PT 1 (0.5 mg/L)
PT 2 (1.0 mg/L)
PT 3 (10.0 mg/L)
GDALL
Observations
5
18
4
27
Observations
5
18
4
27
Mean
120%
106%
101%
108%
Mean
74%
72%
92%
75%
Median
130%
119%
1 14%
127%
Median
70%
70%
95%
74%
Maximum
185%
185%
140%
185%
Maximum
126%
148%
115%
148%
Minimum
51%
35%
36%
35%
Minimum
47%
15%
65%
15%
SD
53%
51%
46%
49%
SD
32%
33%
21%
31%
For GB, the mean accuracy (R) was near ideal at 108%. For GD, the mean R was 75%, though
the variability was less than that of the GB response (as reflected in SD of R). The spiking
23
-------
solution used to fortify the water samples with these two CWAs contains several other chemicals
which affected the library matching for GB (resulting in a lower quality match for this analyte).
As shown in Figures 6-la and 6-lb, the operator was able to use the background subtracting
feature of Chemstation to improve the quality of the library match (and thus, the confidence in
the identification) of GB in the sample from a quality of 50 before the background subtraction to
a quality of 83 after the background subtraction.
need Data Analysis - CTC TEST.M / 111805D.D (MS Data: Quantitated Mulli Pt., QT Reviewed) [Window ...
|JJl|j File Method
r~
if
h
Abundance Scan 1 547 (2.540 min): 1 1 1 805D.D\data.ms
8000-
GOOD-
4000-
2000-
g
c
81
35 ^i/l'.. 59 67 75 |.|,| | . 91,.
9
12
118
5
31
| 138 160166 173 208
m/z»> 20 40 6!0 JXJ idfl 120 140 160 180 '200
Abundance ttl 7484: Sam
8000-
6000-
4000-
2000-
0-
9
43 81
21 3fl.lll , 58 6.7
3
125
m/z-> 20 40 60 80 1 00 1 20 1 40 1 60 1 80 200
PBM Search Results: C:\Database\NISTQ5.L
start
"g Enhanced Data Analy,
Rank | Name
1
2
3
4
r
Sarin
2-B utyl methvlphosphonofluoridate
2-M ethylcyclopenljJl methylphosphonof luoridoate
2-Pentyl methylphosphonofluoridate
Difference Statistics | Text
I Ref No.
17494
26068
43425
35208
Drint Done
MW Qual
140 50
154 50
180 40
168 40
Help |
Figure 6-la. Library Match for GB without Background Subtraction
24
-------
anced Data Analysis - CTC TEST.M / 111805D.D (MS Data: Quantitated Multi Pt., QT Reviewed) [Window ... I- ,|n||X
*siii'^ae:
© H [B Hi H H ^^J^^fir^L^yyiS ri % 1 H^B^J
Abundance
8000-
6000-
4000-
2000-
0-
rn/'z-->
Abundance
8000-
6000-
4000-
2000-
0-
m/z->
Scan 1 546 (
8
41 81
37, M| .50 59 67 75 , |. 37 92
2.539 min): 1 1 1 805D.DSdata.ms (-1 541] (-)
3
125
118 . 166 212
20 4!] eb 80 ido lii i4o 1^0 ido 2iio '
c
43 81
§ 39 III , 59 67 I
tt17494:Sarin
3
125
2b 4b eb sb ido lii i4o 1& 1^0 2do
PBM Search Results: C:\Database\NIST05.L
Rank Name
1 Sarin
2 Sarin
| 3 2-Heptyl methylphosphonofluoridate
n-H exyl methylphosphonof luoridate
o
Ref No. I MWI Qual 11
Difference I Statistics
Text
17484 140
17486 140
54756 196
44855 182
Print
Done
Help
MS Control/Enhance,
j Enhanced Data Analy... I '-• 111805D GB w-o Bac.
Figure 6-lb. Library Match for GB with Background Subtraction
25
-------
Table 6-1 g. Accuracy (Percent Recovery) Results - VX
VX PT Level
PT1 (10.0 mg/L)
PT2(15.0mg/L)
PT 3 (30.0 mg/L)
VXALL
Observations
2
2
2
6
Mean
113%
104%
111%
109%
Median
113%
104%
111%
111%
Maximum
126%
119%
111%
126%
Minimum
100%
89%
111%
89%
SD
nd
nd
nd
13%
nd = SD of R not calculated for observations less than three
Due to the lack of sensitivity for the originally proposed testing levels for VX, data were limited
to three sets of replicates, including the set of replicates from which data were plotted to generate
the calibration curves that were used to quantify the subsequently acquired data. Since there were
only two replicates, SD was not calculated for each PT sample level (though SD is provided for
all PT sample levels combined). Mean recovery for VX was 109%.
6.2 Precision
Precision (Table 6-2), expressed as RSD, was calculated for the three replicate measurements of
the PT samples and matrix-fortified samples (except for 2,4-D and dicrotophos, in which cases
only matrix-fortified samples were used for the calculation, and VX, in which case no matrix
samples were analyzed) by the method listed in Section 5.2. RSD values ranged from 1% to
100% with several trends observed. When compared to benzene and xylenes, where only one
sample type showed a RSD greater than 20%, toluene and ethylbenzene, 2,4-D, dicrotophos, GB,
and GD had multiple samples types with a RSD greater than 20%. This variable response may be
due to either the SPME or the CT-1128 (or contributions from both sources). Additionally, DW3,
a groundwater sample, exhibited a RSD greater than 20% for six of the eight analytes tested.
Table 6-2. Summary of RSD of the Various Sample Types(a)
Sample Type
PT1
PT2
PT3
PT4
DW1
DW2
DW3
DW4
Weakly Buffered
Water (442-30)
Strongly Buffered
Water (442-3000)
THMs Spiked
Water (ASTM
Type II Water)
Benzene
4%
4%
5%
7%
3%
7%
50%
13%
9%
3%
5%
Toluene
198%
1%
61%
12%
6%
7%
66%
14%
100%
3%
3%
Ethylbenzene
12%
11%
6%
1%
36%
44%
52%
3%
11%
4%
0%
Xylenes
6%
13%
8%
2%
7%
14%
3%
14%
13%
5%
30%
2,4-D
NA
NA
NA
NA
36%
17%
25%
15%
9%
32%
16%
Dicrotophos
NA
NA
NA
NA
30%
19%
68%
77%
24%
58%
17%
GB
26%
48%
48%
NA
10%
4%
25%
20%
9%
27%
27%
GD
26%
52%
24%
NA
4%
2%
18%
13%
1%
14%
7%
VX
22%
15%
7%
NA
NA
NA
NA
NA
NA
NA
NA
NA = Not applicable.
(a) Entries in bold italics indicate RSD > 20%.
26
-------
6.3 Linearity
For all target contaminants, a minimum of three PT samples were used to generate a calibration
curve as described in Section 5.3. The curves (Figures 6-2a-i) were constructed from the same
set of replicates if multiple replicates were analyzed (e.g., the first set of PT sample replicates
was used to generate the calibration curve for benzene calibration). When a blank sample was
analyzed as part of the calibration curve, the blank response was included in the construction of
the corresponding calibration curve. It should be noted that not every concentration level of PT
samples was used to construct the curves. For some (ethylbenzene, xylenes, and dicrotophos),
the highest PT sample level was removed to yield better calibration curves. Table 6-3
summarizes the calibration curve results. Calibration for seven of the nine analytes yielded
curves with coefficient of determination, r2, of 0.999 or greater. The notable exceptions are those
analytes for which the provided method lacked sensitivity, 2,4-D and VX (r2 of 0.921 and 0.959,
respectively), though all of the r2 values were greater than 0.920. The number of days for which
the calibration curve was used to quantitate data is also indicated in Table 6-3.
Table 6-3. Summary of Calibration Curve Data
Contaminant
benzene
toluene
ethylbenzene
xylenes
2,4-D
dicrotophos
GB
GD
VX
Number of
Points in Curve
5
5
4
4
5
4
3
3
3
Range
(mg/L)
Oto 10
Otol
Oto 0.1 (1 not linear)
Oto 0.1 (1 not linear)
Oto 100
Oto 500 (1000 not
linear)
0.5 to 10
0.5 to 10
10 to 30
2
r
1.000
1.000
1.000
1.000
0.921
0.999
1.000
1.000
0.959
Number of Days
Curve was Used
8
8
8
8
5
5
9
9
1
27
-------
CT-1128 9-13 Calibration Benzene SPME
7000000
£ 6000000
| 5000000
Q_
3! 4000000
a
to
HI
OL
3000000
2000000
1000000
y = 578244X - 1 098
468
Expected Concentration (mg/L)
Figure 6-2a. Calibration Curve for Benzene
CT-1128 9-13 Calibration Toluene SPME
3 1400000
* 1200000
^ 1000000
OL
CO
800000
400000
200000 -
= 1471470x+3319
n = 5
2 _
10
12
n = 5
r2 = 1.000
0.2
0.4 0.6 0.8
Expected Concentration (mg/L)
1.2
Figure 6-2b. Calibration Curve for Toluene
28
-------
2500000 i
CT-1128 9-13 Calibration Ethylbenzene SPME
0.02
0.04 0.06 0.08
Expected Concentration (mg/L)
Figure 6-2c. Calibration Curve for Ethylbenzene
n=4
r2 = 1.000
0.1
0.12
CT-1128 9-13 Calibration Xylenes SPME
2500000 i
< 2000000
1500000
1000000
03
|
oo
CM
0 f-1^ ,
0.04 0.06 0.08
Expected Concentration (mg/L)
Figure 6-2d. Calibration Curve for Xylenes (Total)
n=4
r2= 1.000
0.12
29
-------
"(?
G)
Si
<
03
-------
9- 100000
2 90000
* 80000
J> 70000
-------
CT-1128 11-01 Calibration VXSPME
n"3
r2 = 0.959
zc
ro
£
01
I
700000
600000
500000
•S- 400000
in £
S<
oo
tM
I-
O
300000
200000
100000
y=26300x- 165810
10 15 20 25
Expected Concentration (mg/L)
30
35
Figure 6-2i. Calibration Curve for VX
6.4 Sensitivity
Sensitivity, as defined in Section 5.4, is the ability to determine the target contaminant at the
concentration of interest in the various water matrices. Results of the sensitivity evaluation are
shown in Table 6-4. For VX and 2,4-D, the concentrations of interest were not achieved due to
lack of sensitivity using the provided method. With the exception of benzene, all other
contaminants were spiked lower than the concentration of interest in the various matrices (see
Table 3-3). For benzene, determining whether the CT-1128 was sensitive enough to detect the
concentration of interest was determined using dilutions of the lowest PT sample. Serial ten-fold
dilutions of PT 1 were performed for each contaminant, and the dilutions were made and tested
until the instrument response did not change by the approximate corresponding dilution factor.
For many of the target contaminants, the response at these levels, while low, may be sufficient
for a qualitative result or identification.
Table 6-4. Sensitivity of CT-1128
Contaminant
benzene
toluene
ethylbenzene
xylenes
2,4-D
dicrotophos
GB
GD
VX
Concentration
of Interest
(mg/L)(a)
0.005
1
0.7
10
0.07
1400
20
1.4
2.1
Sufficient Sensitivity
to Detect Cone, of
Interest
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
No
(a) See Table 3-1 for Target Contaminants and Concentrations of Interest.
32
-------
6.5 Instrument Stability
Stability (Table 6-5), determined by the method described in Section 5.5, was based on the first
and last PT sample of a testing day, expressed as an RPD of the two. The ideal RPD is zero with
those results greater than 20% flagged (those results listed in bold italics). In addition to
presenting the mean percent recovery for each PT level, median, maximum, minimum, and
standard deviation of RPD were reported. Though the verification test involved acquiring a
single calibration which was used for the duration of testing (the number of days over which
stability was tested is indicated in Table 6-5), it is recommended that an operator acquire a new
calibration curve when the response has changed. This frequency will depend on the operator's
need for accuracy and precision. Alternatively, the operator may use the average response of
standards analyzed throughout an analytical run to normalize the sample results for instrumental
drift. This approach was used by the operator in determining potential matrix effects and
interferents. For example: PT 3 samples for benzene, expected to be 1.0 mg/L gave 3.8, 3.9, and
3.7 mg/L for an average response of 3.8 mg/L throughout the testing day when THMs fortified
water samples were analyzed. When determining the R of benzene in the matrix, the operator
normalized the THMs-matrix response, which was 3.9, 3.9, and 4.2 for an average response of
4.01 mg/L to the PT 3 response, giving average R as 106%.
Table 6-5. Results of Stability Testing for CT-1128
Contaminant
benzene
toluene
ethylbenzene
xylenes
2,4-D
dicrotophos
GB
GD
PT Level
(mg/L)
1.00
0.10
0.10
0.10
50.0
500
1.00
1.00
Number of
Days
8
8
8
8
4
4
7
7
Stability (RPD)(a)
Mean
27%
52%
9%
12%
55%
92%
48%
27%
Median
10%
14%
10%
10%
23%
81%
50%
33%
Maximum
100%
133%
17%
33%
83%
195%
83%
63%
Minimum
1%
3%
0%
1%
10%
11%
7%
3%
SD
39%
61%
5%
10%
33%
80%
31%
22%
(a) The ideal RPD is zero with those results greater than 20% flagged (those results listed in bold italics)
The average RPD ranged from that for ethylbenzene (average RPD of 9% with a SD of 5%) to
that for dicrotophos (average RPD of 92% with a SD of 80%). Only two analytes, ethylbenzene
and xylenes (average RPD of 9% and 12%, respectively) had average RPD less than 20%, while
results for toluene and dicrotophos were much greater than 20% (average RPD of 52% and 92%,
respectively).
6.6 Potential Matrix and Interference Effects
The ability of the CT-1128 to detect the target contaminants was challenged by analysis of
fortified samples in different matrices. The results are shown in Table 6-6. To determine
percent recovery (R) of the fortified samples, the average response of the PT samples throughout
the day of testing was used. A matrix effect was defined as recoveries outside of the range of
70% to 130% R value. A trend was observed with the strongly buffered water matrix (442-3000)
which gave R outside of that range, indicating that a matrix effect may be present. This result
33
-------
may be due to an enhancement of the SPME efficiency, which has been observed with the
extraction of analytes in high salt matrices. Results for 2,4-D may be due to sensitivity issues.
For dicrotophos, all but two of the sample types exhibited matrix effects which may also be due
to a less than optimal method for the contaminant. DW3 also showed matrix effects for five of
the eight contaminants tested which may be due to the fact that it is the sole groundwater sample
of the four regional DWs. Potential matrix and interference effects were not tested for VX due to
the lack of sensitivity for this contaminant using the provided method.
Table 6-6. Results of Potential Matrix and Interference Effects
Matrix Effects Observed(a)
benzene
toluene
ethylbenzene
xylenes
2,4-D
dicrotophos
GB
GD
VX
DW1
120%
107%
96%
102%
77%
103%
80%
759%
ND
DW2
82%
193%
102%
129%
30%
142%
123%
92%
ND
DW3
88%
180%
88%
115%
22%
349%
63%
61%
ND
DW4
108%
95%
109%
99%
13%
143%
79%
133%
ND
Weakly
Buffer ed
Water
(442-30)
58%
39%
94%
99%
90%
762%
91%
90%
ND
Strongly
Buffered
Water
(442-3000)
218%
158%
139%
135%
16%
236%
38%
66%
ND
THMs
Spiked
Water
106%
99%
101%
87%
110%
121%
140%
134%
ND
(a) bold italic font indicates values outside 70-130% ranj
ND = no data; VX matrix testing was not performed due
provided method
;e of average response of daily PT samples
to lack of sensitivity for this contaminant using the vendor
6.7 Field Portability
The CT-1128 was removed from the laboratory to be installed and operated within a warehouse
facility. To begin the move, the CT-1128 was shut down after a day of laboratory analysis. The
heating zones were cooled and the turbo pumps were subsequently allowed to spin down. After
venting the instrument, the instrument was packed into the hard transport case for mobilization
into the field. This aspect of transportability was straightforward. The CT-1128 unit weighs
approximately 70 pounds so it is possible to load the unit into a specially designed hard case
using two people. The hard case held the CT-1128 securely and protected the instrument during
mobilization. The packed CT-1128 (which, with its hard case, weighed approximately 225
pounds (102 kg) and measured 36" (91.4 cm) x 36" (91.4 cm) x 39" (99.1 cm)) was loaded onto
the back of a pickup truck.
34
-------
Battelle has a warehouse facility in Columbus, Ohio that is located along the banks of the
Olentangy River. This location was selected for its proximity to the river, as a possible scenario
could involve analyzing grab water samples from the river using the CT-1128. To shield the
instrument from the rain, the truck was moved inside the warehouse (Figure 6-3). The instrument
was powered and supplied with UHP helium from a gas tank (alternatively, the CT-1128's
internal metal hydride storage bottle may be charged to supply hydrogen as the carrier gas but
this was not tested during this verification test). After establishing the gas flow, the source and
quad temperatures were heated. The system was packed and transported at 8:00 A.M. and by
9:30 A.M., the CT-1128 was installed and under vacuum, though it was necessary to wait for the
instrument to reach thermal equilibrium prior to performing the autotune procedure. Though a
dialog box in the MSD Chemstation software tells the operator to wait for two hours, the vendor
claims that the instrument may be operated to obtain data well before this time (this was not
verified as part of the test). At approximately 11:30 A.M., an autotune was performed, followed
by an air and water check. It was at this time that the results of the check indicated an air leak.
After spending a significant amount of time leak checking the system, a stress fracture in the
helium gas line was discovered and replaced. After replacing the gas line, several air and water
checks were performed and showed the N2 and 62 peaks to be gradually decreasing. After
obtaining a satisfactory air and water check (abundance of H^O, N2, and C>2 should all be less
than 10% of the internal calibrant PFTBA), the instrument was left on standby (as the workday
had come to an end). This type of effort highlights the advantage of operating the CT-1128 in a
mobile laboratory environment (e.g., van with source of electrical power) so that downtime
associated with re-establishing vacuum and tuning, which can be several hours, may be
minimized dramatically, allowing the operator to analyze field samples sooner.
On the following day, the
operator tuned the MS and
performed a four-point
calibration for BTEX
analytes. A separate
individual prepared the
standards and provided a
water sample fortified with
BTEX at concentrations
unknown to the operator.
The operator analyzed the
sample and determined
concentrations of the
BTEX analytes based on
the acquired calibration
curve. The results are
presented in Table 6-7.
The recoveries were
Figure 6-3. Use and Installation of CT-1128 Outside of the comparable to what was
Laboratory observed in the laboratory
with recoveries for
benzene, ethylbenzene, and
xylenes at 108%, 135%, and 121% respectively. However, the accuracy for toluene was low,
35
-------
which is opposite to the behavior observed in the laboratory testing (during which a mean
recovery of 440% was observed - see Table 6-lb). This may be due to a non-optimal method for
this contaminant. Overall, in terms of analytical performance, no major differences were
observed due to field deployment.
Table 6-7. Field Portability Results
Contaminant
benzene
toluene
ethylbenzene
xylenes
PT Cone. Level (mg/L)
PT1
0.01
0.001
0.001
0.001
PT2
0.10
0.01
0.01
0.01
PT3
1.00
0.10
0.10
0.10
PT4
10.0
1.00
1.00
1.00
Linearity
r2
1.000
0.996
0.998
1.000
Cone, of Blind
Sample (mg/L)
1.00
0.10
0.10
0.10
Accuracy (R of
Blind Sample)
108
23
135
121
Since the CT-1128 requires approximately two hours to equilibrate once electrical power and gas
have been supplied, it is recommended that the instrument be kept on standby as long as possible
when time for analyzing field samples is a critical factor. Mobilization into the field is
straightforward and is due in large part to the "button up" design of the CT-1128 and its
transport-ready case. When the on-board metal hydride canister is used to supply carrier gas, the
CT-1128 requires only a source of electrical power for several hours of field deployment. The
advantage of on-site analysis of chemicals makes the CT-1128 a powerful tool for the
identification and analysis of chemicals that have been traditionally analyzed by a fixed-
laboratory GC-MS system.
6.8 Operational Factors
In general, GC-MS systems are sophisticated analytical instruments that require proper method
development and an experienced operator to yield optimal results. The verification test was
conducted using experienced GC-MS operators running the analytical methods provided by the
vendor. It is important to note that these methods were not explicitly confirmed by the vendor
using the levels proposed prior to the verification test. For all analytes, the PT sample starting
concentrations had to be determined experimentally in order to determine appropriate levels (as
opposed to the levels listed in test/QA plan for this test(1)). In the case of CWAs, this was not
possible due to the restrictions on the use and availability of these chemicals. The methods
provided for this verification test were not optimized or modified for sensitivity and therefore
may not represent the most optimal performance of the CT-1128 for a particular target
contaminant. It is also important to note that the results of the verification test reflect the
performance of the SPME sample preparation as well as the analysis by the CT-1128. Operator
experience with SPME will therefore be beneficial to using the technique in tandem with the CT-
1128.
36
-------
Though a recurring GC oven temperature error was observed during testing, the CT-1128 did not
require any extensive maintenance during the verification test. The operator replaced the
injection port septum each day as recommended by the vendor. On several occasions, leaks were
suspected. Detection and remedial actions can take significant time. During the field test,
detecting a small leak that seemed to contribute to a large amount of air and water took several
hours. After a suspect copper gas line for the external helium gas cylinder was replaced, it took
additional time to remove the water that had been introduced from the humid environment. It
was noted that the SPME use resulted in great wear and tear on the septum, often resulting in a
cored septum which could serve as a leak. It is therefore advisable to replace the septum at the
end of each testing day. Such situations coupled with daily preparation and sample run times,
which ranged from 22 minutes to 32 minutes, can have a significant impact on sample
throughput. Typically, the daily procedures recommended for operation lasted approximately one
hour. These activities, performed on a system that has reached thermal equilibrium, include a
bakeout procedure, daily mass tuning, and an air/water check. If the system is not at equilibrium
such as when it has been first deployed in the field, it is necessary to allow for two hours for
thermal equilibration after achieving vacuum. Though it is possible to sequence the analytical
steps to maximize sample throughput (e.g., exposing the SPME to absorb analytes for the next
sample while the GC oven temperature is ramping down ), average sample throughput during
verification testing was 11 samples per each ten hour working day, translating to approximately
one sample per hour. Operational costs for the verification test included UHP helium, SPME
vials, crimp seals, SPME fibers, and disposable stir bars. For 100 samples, the total cost for these
supplies was approximately $914. The GC column and standard chemicals were not included in
this cost.
During the verification test, the operator occasionally observed errors for the column temperature
readback. The problem seemed to be intermittent and erratic (alternating from low to correct to
high reading). Inconsistent retention times supported that the problem was not merely a faulty
readback, but actual variation in the GC oven temperature. Based on the technical support
provided by the vendor, a faulty thermocouple was suspected. In addition to providing
troubleshooting support, the vendor provided application support for target contaminants.
37
-------
Chapter 7
Performance Summary
As shown in Table 7-1, benzene accuracy was considerably higher than ideal (100%) at 172%
because of a change in response several days after establishing the calibration curve. Toluene
(which had the lowest accuracy of the BTEX chemicals) exhibited significant over-recoveries,
with an overall mean R of 440%, though accuracy for ethylbenzene and xylenes (total) was close
to 100%. The mean R for 2,4-D and dicrotophos was 62% and 143%, respectively. For GB, the
accuracy was close to ideal at 108%. For GD, the R was acceptable at 75%. The mean R for VX
was 109%, though the concentrations tested were significantly higher than the LDso for this
CWA.
Table 7-1. Summary of Accuracy, Precision, Linearity, and Stability
Contaminant
benzene
toluene
ethylbenzene
xylenes (total)
2,4-D
dicrotophos
GB
GD
VX
Accuracy
Mean R
172%
440%
104%
103%
62%
143%
108%
75%
109%
Precision
Mean RSD
10%
43%
16%
10%
21%
42%
24%
14%
15%
Linearity
r2 of curve
1.000
1.000
1.000
1.000
0.921
0.999
1.000
1.000
0.959
Stability
Mean RPD
27%
52%
9%
12%
35%
92%
48%
27%
27%
Table 7-1 also shows that across all sample matrices, precision, as measured by RSD of
replicates, ranged from 10% RSD for benzene to 43% RSD for toluene.
The calibration curves of seven of the nine contaminants had coefficients of determination (r2) of
0.999 or greater. The exceptions were those contaminants for which the provided methods lacked
sensitivity—2,4-D and VX (r2 of 0.921 and 0.959, respectively), though for all contaminants, r2
values were greater than 0.920.
Instrument stability, in terms of the mean RPD of prepared sample replicates, ranged from 9%
average RPD for ethylbenzene (with a SD of 5%) to 92% for dicrotophos (with a SD of 80%).
38
-------
Only two contaminants, ethylbenzene and xylenes (average RPD of 9% and 12%, respectively),
had average RPDs less than 20%, while toluene and dicrotophos had RPDs significantly greater
than 20% (52% and 92%, respectively). Table 7-2 summarizes the results of testing for matrix
effects on the analyses performed on the CT-1128.
Table 7-2. Summary of Matrix Effects Observed
Contaminant
benzene
toluene
ethylbenzene
xylenes
2,4-D
dicrotophos
GB
GD
VX
Matrix Effect(a) from Potential Interferents ^ = observed
DW1
^
^
ND
DW2
^
^
S
ND
DW3
^
^
S
S
S
ND
DW4
^
S
s
ND
Weakly
Buffered
Water
S
S
S
ND
Strongly
Buffered
Water
S
S
S
S
S
S
S
S
ND
THMs
Spiked
Water
^
S
ND
(a) matrix effect defined as recovery outside range of 70-130% of average response of daily PT samples
ND = no data; VX matrix testing was not performed due to lack of sensitivity for this contaminant using the vendor
provided method
Table 7-2 shows that a matrix effect was present with the strong buffer matrix, which gave Rs
outside the 70% to 130% range for all eight target contaminants tested. DW3 also showed matrix
effects for five of the eight contaminants, which may be due to its origin as a groundwater
sample. VX was not tested for potential matrix and interference effects because of the lack of
sensitivity of the provided method. Table 7-3 shows the results of sensitivity testing of the CT-
1128.
Table 7-3. Results of CT-1128 Sensitivity Testing for Target Contaminants
Contaminant
benzene
toluene
ethylbenzene
xylenes
2,4-D
dicrotophos
GB
GD
VX
Concentration
of Interest
(mg/L)
0.005
1
0.7
10
0.07
1400
20
1.4
2.1
Sufficient Sensitivity
to Detect Cone, of
Interest
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
No
39
-------
Table 7-2 shows that with the exception of VX and 2,4-D (for which the provided methods
lacked sensitivity), the sensitivity of the CT-1128 was sufficient to detect the target contaminants
at the concentrations of interest, i.e., at LD50 or MCL concentrations.
Because the CT-1128 requires approximately two hours to equilibrate once electrical power and
gas have been supplied, it should be kept on standby (under vacuum and thermally equilibrated)
as long as possible when time is a critical factor for analyzing field samples. Mobilization in the
field is straightforward, and the CT-1128 requires only a source of electrical power for several
hours of field deployment when used with its on-board hydrogen canister for a source of carrier
gas. During the non-laboratory testing, the coefficients of determination (r2) achieved in the field
for benzene, toluene, ethylbenzene, and xylenes were 1.000, 0.996, 0.998, and 1.000,
respectively, and the recoveries of a blind test water sample were 108%, 23%, 135%, and 121%,
respectively.
The CT-1128 did not require extensive maintenance during the verification test. On several
occasions, leaks were suspected, and detection and remediation took a significant amount of time
(as it typically does with any GC-MS system). Typical extraction and sample run times ranged
from 22 minutes to 32 minutes. The daily procedures recommended for operation during
verification testing (e.g., mass tuning, air/water check, and SPME fiber/GC column bakeout)
lasted approximately one hour. Average sample throughput during verification testing was 11
samples per 10-hour working day, or approximately one sample per hour. For 100 samples, the
total cost for supplies was approximately $914, not including the GC column and standard
chemicals. At the time of testing, the cost of the CT-1128 GC-MS system, with optional SPME
stirrer/heater, was $140,000.
It is important to note that the results of the verification test reflect the performance of the SPME
sample preparation as well as the analytical results generated by the CT-1128. The methods
provided for this verification test were not optimized or modified for sensitivity by the vendor
prior to use in the verification test and therefore may not represent the most optimal performance
of the CT-1128 for a particular target contaminant. Operator experience with SPME would
therefore be beneficial to using the technique in conjunction with the CT-1128.
40
-------
Chapter 8
References
1. Test/QA Plan for Verification of Mobile Mass Spectrometers, Battelle, Columbus, Ohio,
July 2005.
2. EPA-600-R-93/100. EPA Method 180.1. Turbidity (Nephelometric), Methods for the
Determination of Inorganic Substances in Environmental Samples. 1993.
3. American Public Health Association, et al. Standard Methods for Examination of Water and
Wastewater. 19th Edition. 1997. Washington D.C.
4. EPA 600/4-79/020 Method 150.1. pH, ElectrometricMethod.. 1982.
5. EPA 600/R-94/111 Method 200.8. Determination of Trace Metals by Inductively Coupled
Plasma - Mass Spectrometry. 1994.
6. EPA 600/4-79/020 Method 130.2. Hardness, Total (mg/L as CaCO3) Titrimetric, EDTA.
1982.
7. EPA 600/R-95/131. EPA Method 524.2. Purgeable Organic Compounds by Capillary
Column GC/Mass Spectrometry. Methods for Determination of Organic Compounds in
Drinking Water, Supplement III. 1995.
8. EPA 600/R-95/131. EPA Method 552.2. Haloacetic Acids andDalapon by Liquid-Liquid
Extraction, Derivatization and GC with Electron Capture Detector. Methods for the
Determination of Organic Compounds in Drinking Water, Supplement III. 1995.
9. Battelle SOP HMRC-IV-118-05: Standard Operating Procedure for the Determination of
CA in Wastewater.
10. EPA 600/R-95/131. EPA Method 515.1. Chlorinated Acids in Water by Gas
Chromatography with an Electron Capture Detector. Methods for Determination of
Organic Compounds in Drinking Water, Supplement III. 1995.
41
-------
11. Quality Management Plan (QMP)for the ETV Advanced Monitoring Systems Center,
Version 5.0, U.S. EPA Environmental Technology Verification Program, Battelle,
Columbus, Ohio, March 2004.
12. "National Primary Drinking Water Regulations" (40 CFR 141), U.S. Environmental
Protection Agency, Office of Ground Water and Drinking Water (OGWDW), July 1, 2002.
42
------- |