EPA/600/R-12/653 | January 2013 | www.epa.gov/ord
United States
Environmental Protection
Agency
Verification of Methods for
Selected Chemical Warfare
Agents (CWAs)
Office of Research and Development
National Homeland Security Research Center
-------
Disclaimer
This document was prepared as an account of work sponsored by an agency of the United
States government. U.S. Environmental Protection Agency (EPA) funded and collaborated in
this research under Interagency Agreement DW-89-92328201 with the U. S. Department of
Energy (DOE). It has been reviewed by EPA but does not necessarily reflect EPA's views.
Neither the United States government nor Lawrence Livermore National Security, LLC, nor
any of their employees makes any warranty, expressed or implied, or assumes any legal
liability or responsibility for the accuracy, completeness, or usefulness of any information,
apparatus, product, or process disclosed, or represents that its use would not infringe privately
owned rights. Reference herein to any specific commercial product, process, or service by trade
name, trademark, manufacturer, or otherwise does not constitute or imply its endorsement,
recommendation, or favoring by the United States government or Lawrence Livermore
National Security, LLC. The views and opinions of authors expressed herein do not necessarily
state or reflect those of the United States government or Lawrence Livermore National
Security, LLC, and shall not be used for advertising or product endorsement purposes.
Auspices Statement
This work performed under the auspices of the U.S. Department of Energy by Lawrence
Livermore National Laboratory under Contract DE-AC52-07NA27344; the Lawrence Livermore
National Laboratory report number is LLNL-TR-583012. The research team was comprised of
Heather Mulcahy, Roald Leif, Edmund Salazar, and Carolyn Koester.
Questions concerning this document or its application should be addressed to:
Romy Campisano (EPA Project Officer)
U.S. Environmental Protection Agency
National Homeland Security Research Center
26 W. Martin Luther King Drive, MS NG16
Cincinnati, OH 45268
513-569-7016
Campisano.Romy@epa.gov
11
-------
Abbreviations/Acronyms
Ace acetone
AMDIS Automated Mass Spectral Deconvolution and Identification System
ATL analytical target level
CAS Chemical Abstracts Service (CAS Registry Number®)
CCV continuing calibration verification
CWA chemical warfare agent
DCM dichloromethane
DFTPP decafluorotriphenylphosphine
DOE United States Department of Energy
EPA United States Environmental Protection Agency
ERLN Environmental Response Laboratory Network
EtOAc ethyl acetate
GA Tabun, dimethylamidoethoxyphosphoryl cyanide, formula C5HUN2O2P
GB Sarin, O-isopropyl methylfluorophosphonate, formula C/ftoFC^P
GC/MS gas chromatography/mass spectrometry
GC/TOF-MS gas chromatography/time-of-flight mass spectrometry
GD Soman, O-pinacolyl methylphosphonofluoridate, formula C7H16FO2P
GF Cyclosarin, cyclohexyl methylphosphonofluoridate, formula C7H14FO2P
HD distilled sulfur mustard, fc(2-chloroethyl)sulfide, formula C4H8C12S
FIN1 Nitrogen mustard 1, 6/5(2-chloroethyl)ethylamine, formula C6H13C12N
FINS Nitrogen mustard 3, fr/5(2-chloroethyl)amine, formula C6Hi2Cl3N
i.d. inner diameter
IDL instrument detection limit
IS internal standard
LLNL Lawrence Livermore National Laboratory
MDL method detection limit
NA not applicable
NHSRC National Homeland Security Research Center (Cincinnati, OH)
NIST National Institute of Standards and Technology
OTL optimum theoretical (detection) limit
PFTBA perfluorotributylamine
pKa dissociation constant
in
-------
ppm part(s) per million
PTFE polytetrafluoroethylene
RRF relative response factor
RRT relative retention time
RSD relative standard deviation
RT retention time
RVX Russian VX, O-isobutyl S-(2-diethylaminoethyl) methylphosphonothioate, formula
CnH26NO2PS
SAM Selected Analytical Methods for Environmental Remediation and Recovery (SAM)-2012
SIM selected ion monitoring (operating mode of a mass spectrometer)
S:N signal-to-noise ratio
TEA triethylamine
TIC total ion chromatogram (produced by GC/MS analysis)
TOC total organic carbon
Tris tris(hydroxymethyl)aminomethane
VOA volatile organic analyte
VX O-ethyl-S-(2-diisopropylaminoethyl) methylphosphonothioate, formula Cnf
IV
-------
Executive Summary
In its role as a reference laboratory, Lawrence Livermore National Laboratory (LLNL)
was tasked by EPA National Homeland Security Research Center (NHSRC) to evaluate, modify
when necessary, and perform a single-laboratory verification of sample preparation and analysis
protocols needed to support environmental restoration. The analytes studied in this work
included tabun (GA), nitrogen mustard 1 (HN1), nitrogen mustard 3 (HNS), and O-isobutyl S-(2-
diethylaminoethyl)methyl phosphothioate (Russian VX, or RVX). Main goals of the study were
to perform detection limit studies and calibration procedures to determine if protocols developed
for other G-agents, sulfur mustard and VX were applicable to these other agents in sample
matrices of water, soil, and wipes. This study investigated whether these protocols were able to
meet, in terms of both detection and quantification abilities, Analytical Target Level (ATL)
concentrations for the agents specified. ATLs reflect existing health benchmarks, such as risk-
based criteria and health-based environmental screening levels, based on realistic exposures
following a release and are intended to be interim targets levels for specific analytes that can be
used to guide analytical method development. ATLs are expected to be lower than actual
operational targets, i.e., method reporting limits.
Instrument detection limits (DDLs) for the above agents detected by gas chromatography/
(quadrupole) mass spectrometry (GC/MS) and gas chromatography/time-of-flight mass
spectrometry (GC/TOF-MS) were recorded. IDLs for the chemical warfare agents (CWAs) of
this study were 0.025 ng for GA, HN1, and HNS and 0.8 ng for RVX by GC/MS operated in the
full-scan mode and 0.01 ng for GA, HN1, and HNS and 0.1 ng for RVX by GC/MS with selected
ion monitoring (SIM). Using GC/TOF-MS, IDLs for the CWAs ranged from 0.005 ng to 0.1 ng.
While purely reflecting instrument response, these IDL values were utilized to provide a
preliminary estimate of an instrument's ability to detect the CWAs in the various sample
matrices. This comparison was performed through calculation of an Optimum Theoretical Limit
(OTL), which is an estimated, detectable level of an analyte based on merits of the sample
preparation and analysis procedure, including IDL, sample size, final extract volume, and 100%
extraction efficiency. The OTL is used to estimate the lowest concentration of analyte in a
specific matrix that would be expected to be detected successfully.
-------
Analytical Target Level (ATL) and Optimum Theoretical (Detection) Limit (OTL) Based on
Instrument Detection Limit
Agent
GA
HNla
HN3a
RVXb
Residential Soil (mg/kg)
ATL
0.01
0.01
0.042
OTL
GC/MS
0.0025
0.0025
0.0025
0.080
OTL
GC/
TOF-MS
0.0025
0.00050
0.0020
0.010
Water (ug/L)
ATL
1.4
0.25
0.25
0.021
OTL
GC/MS
1.4
1.4
1.4
46
Wipe (ng)
OTL
GC/MS
25
0.3
25
1.1
25
5.7
800
OTL
GC/
TOF-MS
25
20
100
Notes: Assumes that ATL values for HN1 and HNS are similar to those of HD.
Assumes that ATL values for RVX are similar to those of VX.
cAt this time, no risk-based criteria for surfaces have been designated as wipe ATLs.
Extraction methods of the protocol were then tested for each analyte and the method detection
limits (MDL) was determined by both GC/MS and GC/TOF-MS. The results of the study are
shown below. In many cases, MDLs were not lower than those of the ATLs, indicating that
further method optimization is required.
Selected Analyte Target Level (ATL) and Method Detection Limit (MDL) Calculated for This Study
Agent
GA
HNla
HN3a
RVX"
Residential Soil (mg/kg)
ATL
2.8
0.01
0.01
0.042
MDL
GC/MSC
0.026
0.035
0.057
0.142
MDL
GC/
TOF-MSC
0.00033
0.00057
0.00 16
0.015
Water (iig/L)
ATL
1.4
0.2
0.2
0.021
MDL
GC/MS
16
20
69
MDL
GC/
TOF-MS
0.13
0.084
0.72
22
Notes: Assumes that ATL values for HN1 and HN3 are similar to those of HD.
Assumes that ATL values for RVX are similar to those of VX.
CMDL for reagent sand.
At this time, no risk-based criteria for surfaces have been designated as wipe ATLs.
VI
-------
The results of the work provide a point of reference against which to judge future method
improvements. Method improvements are needed to allow detection of specific analytes at the
ATL level in most matrices. A comparison of ATLs and OTLs (see above table) indicate that for
some compounds neither GC/MS nor GC/TOF-MS provide detection limits to meet expected
analytical requirements. For example, the expected lower limit for RVX in water, 0.021 |ig/L,
cannot be detected using the tested sample preparation strategy and detection by either GC/MS
or GC/TOF-MS. In such cases, extraction methods with more specific detection techniques such
as gas chromatographic tandem mass spectrometry would be expected to improve instrument
detection limits (DDLs), perhaps even by as much as a factor of fifty.
Since this work was performed, the methods in the original protocol have been modified,
with dichloromethane (DCM) being the sole extraction solvent for GB, GD, GF, and HD in
wipes and soils. DCM should also be a reasonable extraction solvent for GA, HN1, and FINS on
solid matrices. Tris(hydroxymethyl)aminomethane buffer, at pH = 8.80, followed by back-
extraction of the of analyte into DCM, is now used for the determination of VX in soil. Work
funded by the U.S. Department of Homeland Security showed that this procedure could also be
used for the extraction of RVX from soils. In that same study, DCM was shown to be an
efficient extraction solvent for both VX and RVX from wipes. It is expected that changes in the
extraction solvent will have a lesser contribution to obtainable method detection limits, with
improvements on the order of five-fold or less. The use of the more efficient pressurized fluid
extraction may be helpful in this regard.
Ultimately, a method must be able to achieve a study's data quality objectives in order to
be useful. It has been assumed that measuring the target analytes at the ATL concentrations
cited here will be necessary (e.g. measured MDLs must be less than or equal to ATL
concentrations). It has been demonstrated that this is achievable for some analytes in certain
matrices (e.g. GA, HN1, and HNS in residential soils and GA in water). However, this is not
achievable for all analytes in all matrices. For this reason, the next step in the evolution of
CWA methods is to strive for lower detection limits. It is expected that the use of detection
technologies with greater specificity (e.g. tandem mass spectrometry), changes in sample
preparation (e.g. larger sample sizes, greater sample extract volume reduction, etc.), and changes
in solvent system will lead to improved method detection limits (MDLs).
vn
-------
Contents
Disclaimer ii
Abbreviations/Acronyms Hi
Executive Summary vi
List of Tables ix
1.0 Introduction and Background 1
2.0 Experimental Details 2
2.1 Chemical Standards, Solvents, and Materials 2
2.1.1 Standards 2
2.1.2 Solvents 2
2.1.3 Solid Chemicals 2
2.1.4 Matrices 3
2.1.5 Equipment 3
2.2 Extraction methods 4
Table 2.1. Summary of Extraction Methods for Wipe, Water, and Soil Samples 4
2.2.1 Soil Extraction Procedure 5
2.2.2 Water Extraction Procedure 5
2.2.3 Wipe Extraction Procedure 6
2.3 Analysis methods 7
2.3.1 Gas Chromatography/Mass Spectrometry 7
2.3.2 Gas Chromatography/ Time -of-Flight Mass Spectrometry 8
3.0 Results 9
3.1 Estimation of Instrument Detection Limits (IDLs) by GC/MS, Operated in the Electron
lonization Full-Scan Mode 9
3.2 Estimation of IDLs by GC/MS, Operated in the Electron lonization, Selected Ion Monitoring
(SIM) Mode 11
3.3 Estimation of IDLs by GC/TOF-MS Operated in the Electron lonization Mode 11
3.4 Calibration studies 12
3.5 Investigation of Method Detection Limits (MDLs) 17
3.5.1 Soils 17
3.5.2 Water 21
3.5.3 Wipes 22
4.0 Conclusions 23
References 26
Vlll
-------
List of Tables
Table 2.1. Summary of Extraction Methods for Wipe, Water, and Soil Samples 4
Table 3.1. Instrument Detection Limit (IDL) for Analytes Present in the Total Ion Chromatogram (TIC)
Produced by the GC/MS Operated in the Electron lonization Full-Scan Mode 10
Table 3.2. Instrument Detection Limit (IDL), for GC/MS Operated in the Full-Scan Mode, When the
Weakest of Two Qualifying Ions (i.e., 2° qual. ion) Must Be Detected at S:N> 3:1 10
Table 3.3. Instrument Detection Limit (IDL) for GC/MS Operated in the SIM Mode 11
Table 3.4. Instrument Detection Limit (IDL) for GC/TOF-MS 12
Table 3.5. Chemical Abstract Number, Retention Time, and Relative Retention Time for the Analytes and
Internal Standards Investigated 13
Table 3.6. Calibration Levels for GC/MS, Operated in Full-Scan Mode 14
Table 3.7. Relative Response Factor (RRF), Relative Standard Deviation (RSD) for RRF, and R2 Value
(Linear Regression) forthe Calibration Levels of Table 3.6, for GC/MS, Full-Scan 14
Table 3.8. Calibration Levels for GC/MS, Operated in SIM Mode 15
Table 3.9. Average RRF, RSD for RRFs, and R2 Value (Linear Regression) for the Calibration Levels,
Shown in Table 3.8, for GC/MS, SIM 15
Table 3.10. Calibration Levels for GC/TOF-MS 16
Table 3.11. Average RRF, RSD for RRF, and R2 Value (Linear Regression) for the Calibration Levels,
Shown in Table 3.8, for GC/TOF-MS 16
Table 3.12. Distribution of CWA Between the Different Extraction Solvents, with Analyses by Full-Scan
GC/MS 18
Table 3.13. Recovery and MDL for Selected of CWAs in Various Soils, With Analyses by Full-Scan,
GC/MS 19
Table 3.14. Distribution of CWA Between the Different Extraction Solvents, With Analyses by GC/TOF-
MS 20
Table 3.15. Recovery and MDL for Selected of CWAs in Various Soils, With Analyses by GC/TOF-MS
20
Table 3.16. Recovery, Standard Deviation, and MDL for GA, HN1, HNS and RVX in Laboratory Water
Analyzed by GC/MS, Full-Scan 21
Table 3.17. Recovery, Standard Deviation, and MDL for GA, HN1, HN3 and RVX in Laboratory Water
Analyzed by GC/TOF-MS 21
Table 3.18. Recovery, Standard Deviation, and MDL for GA, HN1, HNS and RVX From Wipes
Analyzed by GC/MS, Full-Scan (Recoveries FromExtract 1) 22
Table 3.19. Recovery, Standard Deviation, and MDL for GA, HN1, HNS and RVX From Wipes
Analyzed by GC/MS, Full-Scan (Recoveries From DCM Extraction) 23
Table 4.1. Analyte Target Level (ATL) and Optimum Theoretical Detection Limit (OTL) Based on
Instrument Detection Limits (IDLs) Presented in Tables 3.2 and 3.4 24
Table 4.2. Analyte Target Level (ATL) and Method Detection Limit (MDL) Calculated for This Study.24
IX
-------
1.0 Introduction and Background
In September 2002, the United States Environmental Protection Agency (EPA)
announced the formation of the National Homeland Security Research Center (NHSRC). One
research goal of the NHSRC is to ensure the availability of verified technologies and methods for
addressing risks posed by chemical warfare agent (CWA) contamination incidents. Verified
technologies and methods are needed to detect and accurately measure contaminants of concern
in environmental matrices following an intentional or unintentional release and to demonstrate
that an area is safe after remediation. Furthermore, such methods must be published in a form
that can easily be shared by multiple laboratories that could provide chemical analyses. The
Environmental Protection Agency established the Environmental Response Laboratory Network
(ERLN) to assist in addressing chemical, biological, and radiological threats during nationally
significant incidents. The ERLN is a national network of laboratories that can be ramped up as
needed to support large scale environmental responses by providing consistent analytical
capabilities, capacities, and quality data in a systematic, coordinated response.
As the first step in providing methods to ensure analytical consistency when multiple
laboratories must analyze a large number of samples resulting from a chemical scenario, EPA
now publishes Selected Analytical Methods for Environmental Remediation and Recovery
(SAM)1. In addition, ERLN laboratories have been testing draft methods for analyses of CWA.
The analytes of interest for which draft CWA methods are being tested by the ERLN laboratories
are sarin (GB), soman (GD), cyclosarin (GF), sulfur mustard (HD), and O-ethyl-S-(2-diisopropyl
aminoethyl) methylphosphonothioate (VX). However, due to the similarity of chemical structure,
the draft CWA methods should also be applicable to other CWAs. Thus, in its role as a reference
laboratory, Lawrence Livermore National Laboratory (LLNL) was tasked by NHSRC to
determine if other compounds, including tabun (GA), 6/X2-chloroethyl)ethylamine (HN1),
^m(2-chloroethyl)amine (HNS), and (9-isobutyl
-------
2.0 Experimental Details
Extraction experiments were performed in accordance with the draft CWA methods,
those previously developed and tested by LLNL for other CWAs, and incorporated quality
assurance protocols cited in EPA Method 8270.4
2.1 Chemical Standards, Solvents, and Materials
2.1.1 Standards
CWA standards of GA (purity 83%), HN1 (purity 65%), HNS (purity 79%), and RVX
(purity 46%) were synthesized in-house and diluted in dichloromethane (DCM) for DDL,
calibration, and all method performance studies.
Surrogate and internal standards used were those of EPA Method 8270D. The surrogate
standard mix included nitrobenzene-ds, 2-fluorobiphenyl, phencyclidine-ds, terphenyl-di4, and
triphenyl phosphate. Specific solutions purchased for this work included: Base/Neutrals
Surrogate Standard, 1000 ng/mL (Catalog number ERB-076, Cerilliant, Round Rock, TX);
Triphenylphosphate, 5000 |ig/mL (Catalog number ERT-1088, Cerilliant); and PCP-d5
(phencyclidine-ds), 1000 |ig/mL (Catalog number P-006, Cerilliant).
Internal standards included l,4-dichlorobenzene-d4, naphthalene-dg, acenaphthene-dio,
phenanthrene-dio, chrysene-di2, and perylene-di2. These standards were purchased as a
Semivolatile Internal Standard Mix, 2000 |ig/mL (Catalog number 861238, Supelco, Bellefonte,
PA).
Decafluorotriphenylphosphine (DFTPP), used to verify that the GC/MS was functioning
properly, was purchased as a solution at a concentration of 1000 |ig/mL in acetone (Catalog
number 47941, Supelco, Bellefonte, PA).
2.1.2 Solvents
Solvents used included:
• Acetone (Ace, PESTANAL®, solvent for residue analysis, >99.8%, Catalog
number 34480, Sigma-Aldrich, St. Louis, MO)
• Ethyl acetate (EtOAc, PESTANAL®, solvent for residue analysis, >99.8%,
Catalog number 34490, Sigma-Aldrich)
• Dichloromethane (DCM), stabilized with amylene at -25 mg/L (PESTANAL®,
solvent for residue analysis, >99.8%, Catalog number 34488, Sigma-Aldrich)
• Triethylamine (TEA, >99.5%, Catalog number 471283, Sigma-Aldrich)
2.1.3 Solid Chemicals
Solid chemicals used included:
-------
• Sodium chloride (puriss. p.a., ACS reagent, anhydrous, >99.5% (AT), Catalog
number 71379, Sigma-Aldrich, St. Louis, MO)
• Sodium sulfate (puriss. p.a., ACS reagent, anhydrous, >99.0% (T) powder (fine),
Catalog number 71960, Sigma-Aldrich, St. Louis, MO)
2.1.4 Matrices
Matrices evaluated included:
• Sand, purified, CAS No. 14808-60-7, Catalog No. 3382-05 (JT Baker, Inc.,
Phillipsburg, NJ)
• Nebraska Aglands Ap soil, with composition of 5.1% sand, 57.5% silt, 31.7%
clay, and 1.9% total organic carbon (TOC) and pH 5.5 in 1:1 soil:water (obtained
from National Exposure Research Laboratory, EPA, Las Vegas, NV)
• Georgia Bt2 soil, with composition of 46% sand, 22% silt, 32% clay, and 0.2%
TOC and pH 5.0 in 1:1 soil:water (obtained from National Exposure Research
Laboratory, EPA, Las Vegas, NV)
• Virginia soil, with composition of 64.5% sand, 28% silt, 7.5% clay, and 2.6%
TOC and pH 4.1 in 1:1 soil:water (obtained from National Exposure Research
Laboratory, EPA, Las Vegas, NV)
• Laboratory water (18 MQ from a Milli-Q® System, Millipore, Billerica, MA)
• Wipes, 3 in. x 3 in. (Kendall™-Curity™, 12-ply, P/N 1903, available from Tyco
Heathcare Group LP, Mansfield, MA)
2.1.5 Equipment
• Waterbath sonicator (Branson, Model 3510, Danbury, CT). While this model had
a temperature control feature, it was not used in this study.
• Rapid Vap® unit, customized to accommodate 40-mL vials (LabConco, Kansas
City, MO)
• Pierce Reacti-Therm™ III, #18824, heating module equipped with the Pierce
Reacti-Therm III, #188 evaporation module (ThermoScientific, Hudson, NH)
• Accuspin™ Model 400 Centrifuge, with a custom rotor to accommodate 50-mL
vials (Fisher Scientific, Pittsburg, PA)
• Glass beads (5 mm, P/N 18406, Sigma Aldrich, St. Louis, MO)
• 40-mL volatile organic analyte (VOA) vials, with a polytetrafluoroethylene
(PTFE)-lined screw cap (P/N 0040-0310-PC, Environmental Sampling Supply,
Oakland, CA)
50-mL, clear, glass, centrifuge vial with polytetrafluoroethylene (PTFE)-lined
screw cap (Kimble®-Contes , Kimble and Chase, LLC., Vineland, NJ, P/N
73785-50)
-------
2.2 Extraction methods
Samples were extracted by simple solvent methods. The extraction methods are
summarized in Table 2.1.
Table 2.1. Summary of Extraction Methods for Wipe, Water, and Soil Samples
Sample size
Add surrogates
Extraction
technique
Reduction of
solvent volume
Add internal
standard
Analysis by
GC/MS or
GC/TOF-MS
Wipe
1 wipe
Add 0.5 ug of each surrogate
Add 1 5 mL of 25/50/25
(v/v/v) Ace/DCM/EtOAc;
extract on shaker table for 15
min; retain Extract 1. Add 15
mL of 5/95 TEA/EtOAc to
previously -extracted wipe;
extract on shaker table for 15
min; retain Extract 2.
Keeping solvent extracts
separate and using nitrogen,
reduce the entire amount of
Extract 1 to 1.0 mL and entire
amount of Extract 2 to 1.0 mL.
Add internal standard mix so
that final concentration in
sample extracts is 1 ppm each
analyte.
Analyze the 1.0-mL aliquot of
Ace/DCM/EtOAc Extract 1;
Analyze the 1.0-mL aliquot of
TEA/EtOAc Extract 2.
Water
35 mL
Add 1.0 ug of each surrogate
Add -8.8 g NaCl and mix
until dissolved; add 2.00 mL
DCM and extract on shaker
table for 2 minutes; allow
layers to separate and
centrifuge if necessary; collect
DCM layer and dry with -50
mg anhydrous Na2SO4;
transfer 1.00 mL into
autosampler vial.
Not applicable.
Add internal standard mix so
that final concentration in
sample extract is 1 ppm each
analyte.
Analyze 1.0-mL aliquot of
DCM extract.
Soil
10 g
Add 0.5 ug of each surrogate
Mix 2.5 g anhydrous Na2SO4
and 5-10 glass beads with soil;
extract, 1 hr by waterbath
sonication, with 25 mL
25/50/25 (v/v/v)
Ace/DCM/EtOAc; add 1-2 g
NaSO4; retain Extract 1. Add
25 mL 5/95 TEA/EtOAc to
previously-extracted soil;
sonicate 1 hr in waterbath; add
1-2 g NaSO4; collect Extract 2.
Keeping solvent extracts
separate and using nitrogen,
reduce the entire amount of
Extract 1 to 1.0 mL and entire
amount of Extract 2 to 1.0 mL.
Add internal standard mix so
that final concentration in
sample extracts is 1 ppm each
analyte.
Analyze the 1.0-mL aliquot of
Ace/DCM/EtOAc Extract 1;
Analyze the 1.0-mL aliquot of
TEA/EtOAc Extract 2.
-------
2.2.1 Soil Extraction Procedure
Ten-gram samples of soil were weighed (to the nearest 0.01 g) into precleaned 40-mL
volatile organic analyte (VOA) vials. All CWAs were spiked as dilute multi-component
solutions in DCM directly into the soils. In addition, surrogate solutions were added, in
appropriate amounts, directly into the soil samples. Approximately 2.5 g of anhydrous sodium
sulfate and 5-10 glass beads were also added to each sample.
The soils were first extracted with 25 mL 25/50/25 (v/v/v) Ace/DCM/EtOAc. This
solvent was added to each sample, the vial capped tightly, and the sample was vortex-mixed for
approximately 30 seconds to ensure a free-flowing slurry. The samples were sonicated, in a
water bath at ambient temperature, for 1 hour. Samples were then removed from the water bath
and briefly mixed using the vortex mixer. One to two grams of anhydrous sodium sulfate were
added to each sample, which was then capped tightly and shaken well. Each sample was allowed
to settle by gravity or centrifuged for 3-5 minutes (or longer, if necessary to clarify the solution).
Care was taken not to exceed 870 x g, the relative centrifugal force over which the vials were
found to break. The solvent layer was decanted or pipetted into a new VOA vial (Extract 1).
The previously extracted soil was then extracted a second time using 25 mL of 5/95 (v/v)
TEA/EtOAc. Previous work had shown that this solvent system afforded efficient recovery for
VX and for the surrogate ds-phencyclidene (which were not recovered with the 25/50/25 (v/v/v)
Ace/DCM/EtOAc used in the first extraction step); therefore, this solvent system was thought to
be necessary for the recovery of RVX from soils. The solvent was added to each sample, and the
sample was vortex-mixed for approximately 30 seconds. The samples were sonicated, in a water
bath at ambient temperature for 1 hour. The samples were then briefly mixed with the vortex
mixer. One to two grams of anhydrous sodium sulfate were added to each sample, which was
then capped tightly and shaken well. Each sample was allowed to settle by gravity or
centrifuged. The solvent layer was decanted or pipetted into a new VOA vial (Extract 2).
Extracts from the two different solvent systems were collected and analyzed separately.
The sample extracts were concentrated prior to analysis. Each sample was evaporated to
just below 1 mL using a gentle stream of clean dry nitrogen provided by either a RapidVap unit,
and/or a Reacti-Therm unit. During evaporation, the internal wall of the vial was washed several
times with DCM. Care was taken to ensure that the extract was not allowed to evaporate to
dryness. The sample extract was adjusted to a final volume of 1.0 mL with either DCM or 5/95
(v/v) TEA/EtOAc, as appropriate. Just prior to analysis, internal standards were added to each
sample extract.
2.2.2 Water Extraction Procedure
To a 50-mL centrifuge vial were added 35 mL water. A 60-mL VOA vial may also be
used for sample extractions; however, the conical bottoms of the centrifuge vials allow the DCM
layer to be removed more easily than from the VOA vials. All CWAs were spiked as dilute,
multi-component solutions in DCM directly into the water. In addition, surrogate solutions were
added, in appropriate amounts, directly to the samples.
To each water sample were added 8.75 g of sodium chloride, and the sample was then
shaken vigorously or mixed with a vortex mixer for 2 minutes or until the sodium chloride
-------
dissolved completely. Then, 2.00 mL of dichloromethane was added to each sample, which was
capped tightly and shaken vigorously (or vortexed) for 2 minutes, venting periodically to reduce
pressure. The phases were allowed to settle by gravity or centrifuged for ~5 minutes. Note that if
centrifugation is used, do not exceed 35 x g, or the vials will break.
Using a glass pipette, approximately 1.5 mL (or as much as possible) of the DCM (lower)
layer was transferred to a 4-mL vial with a PTFE-lined screw cap, taking precautions to exclude
any water from the pipette. Approximately 50 mg (or more) of anhydrous sodium sulfate was
added to each sample. The sample was capped and shaken vigorously (or vortexed) for 2
minutes. Using a glass pipette, 1.0 mL of the extract processed with drying agent was transferred
to a 2.0 mL vial with a PTFE-lined screw cap. Prior to analysis, the appropriate amount of
internal standard was added, and the vial was capped and inverted several times to mix the
contents.
2.2.3 Wipe Extraction Procedure
A gauze wipe was placed in a precleaned 40-mL VOA vial. The wipes were used as
received and were not cleaned prior to use. While some organic compounds were extracted from
the wipes, these compounds did not interfere with the detection of the target analytes and none of
the target analytes were ever detected in blank samples. All CWAs were spiked as dilute multi-
component solutions in DCM directly onto the dry wipe. In addition, surrogate solutions were
added, in appropriate amounts, directly onto the wipe samples. Wipes were spiked first with
CWA, then spiked with surrogates, and extracted immediately after the spiking procedure was
complete.
All samples were extracted with 15 mL of 25/50/25 (v/v/v) Ace/DCM/EtOAc. This
solvent was added to each vial, the vial was sealed tightly, and the samples were extracted by
sonication, in a water bath, for 15 minutes. After the samples were removed from the sonicator,
they were briefly shaken by hand, and the solvent layer was transferred, by pipette, into a new
40-mL VOA vial (Extract 1).
The wipe was extracted a second time, as above, with 15 mL of 5/95 (v/v) TEA/EtOAc
and the resulting sample extract was transferred to a new VOA vial (Extract 2). Extracts from
the two different solvent systems were collected and analyzed separately.
Prior to analysis, the sample extracts were evaporated to just below 1 mL using a gentle
stream of clean dry nitrogen, through the use of a RapidVap unit and/or a Reacti-Therm unit. The
sample extract was not allowed to evaporate to dryness. During evaporation, the internal wall of
the vial was washed several times with DCM. The sample extract was adjusted to a final volume
of 1.0 mL with DCM and the appropriate amount of internal standard was added prior to
analysis.
-------
2.3 Analysis methods
2.3.1 Gas Chromatography/Mass Spectrometry
GC/MS analysis was performed with an Agilent 6890 GC coupled with an Agilent 5973 MS
(Agilent Technologies, Inc., Santa Clara, CA). The MS was tuned using the manufacturer's
software procedures, with perfluorotributylamine (PFTB A) as a mass calibrant. The MS was
operated in accordance with the laboratory's standard operating procedures. The performance of
the GC/MS was checked with a 50-ng injection of decafluorotriphenylphosphine (DFTPP) prior
to sample analysis. During analysis sequences, a continuing calibration verification (CCV)
standard near the midpoint of the calibration range was analyzed every 10 samples. The CWA
concentrations calculated for the CCV, using the most recent calibration curve, were required to
be within 20% of the expected value in order for the data collected between CCV checks to be
considered valid.
The standard GC parameters were:
Carrier gas: Helium, at a constant flow of 32 cm/s
Injection mode: Splitless for 0.75 min
Injector temperature: 250 °C
Sample injection volume: 1 jiL
GC Column: Agilent HP-5MS (5%-phenyl)-methylpolysiloxane
Column dimensions: 30 m x 0.25 mm inner diameter (i.d.) x 0.25 jim film
thickness
GC temperature program: 40 °C (3 min), 10 °C/min to 150 °C, 25 °C/min to 280 °C,
hold for 10.8 min
The standard MS conditions for full-scan analyses performed in electron ionization mode
were as follows:
MS transfer line temperature: 280 °C
MS source temperature: 230 °C
MS quadrupole temperature: 150 °C
Solvent delay time: 3 min
Scan range: 35-500 m/z
Electron energy: 70 eV
Scan time: 3.15 scans/s
Ionization polarity: Positive
Library searching: National Institute of Standards and Technology (NIST) 08
Mass Spectral Data Base
-------
The standard MS conditions for selected ion monitoring (SIM) analyses performed in
electron ionization mode were:
MS transfer line temperature: 280 °C
MS source temperature:
MS quadrupole temperature:
Electron energy:
Ion dwell time:
Ionization polarity:
230 °C
150°C
70 eV
100 ms per ion (each analyte was assigned its own SIM
group; depending on the number of ions monitored, cycle
times ranged from 1.44-2.86 cycles/s)
Positive
2.3.2 Gas Chromatography/ Time -of-Flight Mass Spectrometry
GC/TOF-MS experiments were performed with an Agilent 6890 gas chromatograph
(Agilent Technologies, Inc., Santa Clara, CA) coupled with a LECO Pegasus 4D mass
spectrometer (LECO Corp., St. Joseph, MI). Prior to use, the GC/TOF-MS was tuned with the
vendor's standard protocols and PFTBA as a mass calibrant. An injection of 15 ng DFTPP was
used to check the performance of the instrument prior to analyzing samples. Experimental data
were collected using the same instrument conditions, including electron multiplier voltages, as
those used to produce the DFTPP check samples. During analysis sequences, CCVs were
analyzed and evaluated as previously described for GC/MS.
Injection size:
Inlet type:
Injection mode:
Pulse pressure:
Purge time:
Carrier gas:
GC injection port:
GC columns:
GC oven (primary):
GC oven (secondary)
GC transfer line:
split/splitless
pulsed-splitless
40 psi for 0.5 min
35 sec at 30 mL/min
He with constant flow of 1.2 mL/min
250 °C
15 m x 0.18 mm i.d. x 0.18 |im film thickness, HP5-MS UI
(Agilent Technologies, Inc., Santa Clara, CA)
1 m x 0.1 mm i.d. x 0.1 jim film thickness, Rxi-17 (Restek,
Bellefonte, PA)
55 °C held for 0.5 min, 20 °C/min to 100 °C, 40 °C/min to 280 °C,
held for 2.75 min
70 °C held for 0.5 min, 20 °C/min to 115 °C, 40 °C/min to 295 °C,
held for 1.64 min
295 °C
-------
The following MS conditions were used for detection.
MS filament delay: 1.5 min
MS scan range: 35-500, at a data acquisition rate of 15 spectra/sec
MS source: 250 °C
Electron energy: 70 eV
3.0 Results
3.1 Estimation of Instrument Detection Limits (IDLs) by GC/MS, Operated in
the Electron lonization Full-Scan Mode
IDLs reflect only the response of the GC/MS to the analyte of interest and do not consider
factors attributed to sample preparation or matrix complications. Nonetheless, they are useful to
estimate the lower bound of instrument performance. DDL values were estimated by determining
the injected mass at which the analyte peak in the total ion chromatogram (TIC) produced a
signal-to-noise ratio (S:N) of 3:1. If S:N was greater than 3:1 for a given mass of analyte, the
standard solution was diluted and reinjected into the GC/MS. S:N values did not always scale
linearly with decreases in analyte mass; therefore, in some cases, S:N values greater than 3:1
were reported, as further reduction of the mass of analyte injected into the GC/MS did not
produce a S:N value that was at least 3:1. In these cases, if the S:N was between 3:1 and 10:1,
the mass injected into the GC/MS and reliably detected was estimated to be the DDL; see Table
3.1. The S:N value was reported as the "Peak-Peak S:N," which is a corrected signal divided by
the peak-to-peak noise, as calculated by Agilent's ChemStation® software. For GA, HN1, HN3,
and RVX, IDLs were 0.05 ng, 0.025 ng, 0.025 ng, and 0.4 ng, respectively.
In addition, at the IDL concentration, the goodness of fit between the mass spectra
produced and that of the agent as reported in the NIST 08 library was determined. A match, or
fit, factor of 999 indicates a perfect correspondence between a mass spectrum and that of an
authentic standard. The forward fit is a measure of how well all of the mass peaks in an
unknown's mass spectrum (even those that might correspond to interfering compounds) match
those found in the library spectrum. The reverse fit is a measure of how well the number and
intensities of the peaks in the library spectrum are represented in an unknown's spectrum. The
reverse fit is usually a larger number (i.e., a better match) than the forward fit because the reverse
fit value is not influenced by mass spectral interferences. The Automated Mass Spectral
Deconvolution and Identification System (AMDIS) fit is comparable to the forward fit value;
however, AMDIS applies a spectral deconvolution program to remove extraneous mass peaks
prior to searching the databases, which results in greater fit factors. Note that for AMDIS, a fit
of 100 indicates a perfect match between a mass spectrum of a sample and that of a compound in
the mass spectral database. In general, fits >800 (or >80 for AMDIS) represent reasonable
matches with library spectra. For all analytes, AMDIS fits were >80%, indicating that good
-------
library matches were obtained at DDL concentrations of analytes. However, the fits in Table 3.1
represent optimal library fits because of the high purities of the injected standards.
Table 3.1. Instrument Detection Limit (IDL) for Analytes Present in the Total Ion
Chromatogram (TIC) Produced by the GC/MS Operated in the Electron lonization Full-
Scan Mode
Analyte
GA
HN1
HNS
RVX
TIC IDL (ng)
0.05
0.025
0.025
0.4
Average S:N for
TIC peak
4.8
5.4
6.1
5.3
Average
Forward Fit
826
746
795
837
Average
Reverse Fit
893
781
873
844
Average
AMDIS Fit
92
82
93
79
Note: Signal-Noise (S:N) and match factors represent the average of triplicate measurements of the same standard
solution.
IDLs were also estimated based on the mass that provided S:N of at least 3:1 for the
weakest qualifying ion produced by full scan electron ionization GC/MS (when detection calls
for the presence of the quantification ion and only two qualifying ions). This IDL reflects the
practice of analyte quantitation based on the response of a single quantitation ion when
acceptable signals of at least two qualifying ions are also present. The results of these IDL
studies are found in Table 3.2. IDLs were 0.025 ng for GA, HN1, and HN3 and 0.8 ng for RVX.
Table 3.2. Instrument Detection Limit (IDL), for GC/MS Operated in the Full-Scan Mode,
When the Weakest of Two Qualifying Ions (i.e., 2° qual. ion) Must Be Detected at S:N > 3:1
Analyte
GA
HN1
HN3
RVX
Quant.
Ion
(m/z)
70
120
154
86
Qual. Ions
(m/z)
1°
133
92
92
99
2°
162
154
168
71
IDL based
on
Qualifying
Ions (ng)
0.025
0.025
0.025
0.8
Average S:N for
Peak in Mass
Chromatogram
of 2° Qual. Ion
6.3
4.9
5.8
5.3
Average
Forward
Fit
851
919
869
860
Average
Reverse
Fit
939
959
896
865
Average
AMDIS
Fit
91
87
87
77
Note: Signal-Noise (S:N) and match factors represent the average of triplicate measurements of the same standard
solution.
10
-------
3.2 Estimation of IDLs by GC/MS, Operated in the Electron lonization,
Selected Ion Monitoring (SIM) Mode
DDLs for SIM were determined by making successive injections of individual standards
of decreasing analyte concentration until S:N of 3:1 to 10:1 was obtained, as described
previously, for the signal from the chromatographic peak produced by the weakest confirming
ion monitored. The analyte mass at which S:N of 3:1 to 10:1 was obtained for three successive
injections was reported as the DDL. DDLs were 0.01 ng for GA, HN1, and HN3 and 0.1 ng for
RV: see Table 3.3.
Table 3.3. Instrument Detection Limit (IDL) for GC/MS Operated in the SIM Mode
Analyte
GA
HN1
HN3
RVX
Quant.
Ion
(m/z)
70
120
154
86
Qual. Ions
(m/z)
1°
133
92
92
99
2°
162
154
168
71
IDL (ng)
0.01
0.01
0.01
0.1
S:N TIC
8.6
11
4.2
5.6
Average S:N
for Peak in
Mass
Chromatogram
of 2° Qual. Ion
6.4
3.2
7.3
3.4
Note: Signal-Noise (S:N) values represent the average of triplicate measurements of the same standard solution.
Qual., qualifying; quan, quantifying; SIM, selected ion monitoring
3.3 Estimation of IDLs by GC/TOF-MS Operated in the Electron lonization
Mode
DDLs for GC/TOF-MS were determined by making successive injections of individual
standards of decreasing analyte concentration until S:N of 3:1 to 10:1 was obtained, as described
previously, for the signal from the chromatographic peak produced by the weakest confirming
ion monitored. The analyte mass at which S:N of 3:1 to 10:1 obtained for three successive
injections was reported as the IDL. IDLs ranged from 0.005 ng to 0.1 ng injected; see Table 3.4.
However, the spectral matches at the DDL, particularly for GA, were not as good as for the
GC/MS. The lower-quality matches might, in part, be attributed to the fact that most of the NIST
library data were collected using a quadrupole GC/MS instead of the slightly different GC/TOF-
MS analyzer. Most of the DDLs determined by GC/TOF-MS are lower than those measured by
GC/MS (full scan) and within a factor of two (usually lower) of IDLs measured by GC/MS
(SIM). And, in contrast to GC/MS SIM, because of the nature of the GC/TOF-MS detector, full
mass spectral data are always available.
11
-------
Table 3.4. Instrument Detection Limit (IDL) for GC/TOF-MS
Analyte
GA
HN1
HN3
RVX
Quant.
Ion
(m/z)
70
120
154
86
Qual. Ions
(m/z)
1°
133
92
92
99
2°
162
154
168
71
IDL (ng)
0.025
0.005
0.02
0.1
Average S:N for
Peak in Mass
Chromatogram
of 2° Qual. Ion
3
o
J
4
4
Average
Forward Fit
445
590
839
722
Average
Reverse Fit
445
760
865
722
Note: Signal-Noise (S:N) values represent the average of triplicate measurements of the same standard solution.
Qual., qualifying; quan, quantifying
3.4 Calibration studies
Using practices that were consistent with EPA Method 8000C,5 a procedure was
implemented for quantifying GA, HN1, HNS, and RVX using the internal standards of EPA
Method 8270D,6 which include d4-l,4-dichlorobenzene, dg-naphthalene, and dio-phenanthrene.
Method 8270D surrogates, used in this study, were ds-nitrobenzene, ds-phencyclidine, di4-
terphenyl, and triphenyl phosphate. These internal standards and surrogates are generally
available in EPA laboratories; therefore, they are used in the draft CWA method. Single
laboratory verification experiments were performed using these standards; however, because
these surrogates and CWAs are different in their chemical and physical behaviors, the surrogates
are useful only in providing assurances that no gross problems in the extraction process occurred.
The surrogates cannot be used to gauge how efficiently a CWA might be extracted from a
sample matrix.
Table 3.5 lists CAS Registry Number® (Chemical Abstracts Service, Columbus, OH),
retention time, and relative (to the internal standard) retention time for the analytes determined in
this method. Data were collected using GC/MS conditions previously described in Section 2.3.
The concentrations chosen for full-scan and SIM calibrations are shown in Tables 3.6 and 3.8,
respectively. Table 3.10 shows GC/TOF-MS calibration ranges used for the CWA and for the
surrogates. The calibration range is narrow because of the ERLN restriction of working with the
ultradilute CWA standards (i.e., the ERLN laboratories must work with CWA concentrations of
10 ppm or less). Multi-component CWA standards would most probably be used by the ERLN in
method verification studies and all mixed standards must be made from 10-ppm single-agent
stock solutions. Thus, 3-5 ppm was chosen as the highest possible calibration level. Working
with a narrow calibration range was acceptable because it is expected that the method will be
used to determine levels of CWAs at, or below, some health-based guideline level used to
demonstrate that CWAs have been effectively removed from a previously contaminated, and
subsequently remediated, area. After an area has been decontaminated, high concentrations of
residual CWAs are not expected to be present.
12
-------
Once calibration levels were established, calibration curves and relative response factors
(RRFs) were determined. Table 3.7 shows the data obtained for full-scan GC/MS calibration
levels 0.025/0.8-1.0/3.0 ng/uL for the CWAs and 0.2- 2 ng/uL for the surrogates (internal
standards were at 0.5 ng/uL). Because the percent relative standard deviation (RSD) for the
response factor values was not always less than or equal to 15% (based on the guidance in 8000-
series methods), all quantitation for the CWAs was based on linear regression. Table 3.9 shows
comparable data for GC/MS, SIM analyses for a calibration range of 0.01/0.1-0.25/1 ng/uL for
the CWAs and 0.01-0.5 ng/uL for the surrogates (internal standards were at 0.5 ng/uL). For the
reasons stated above, linear regressions were also used for quantitation with GC/MS, SIM.
Table 3.11 shows GC/TOF-MS calibration ranges used for the CWA (0.01/0.1-1/5 ng/uL) and
for the surrogates 0.01-1 ng/uL; internal standards were at 0.5 ng/uL. Linear regression was
also used for quantification with the GC/TOF-MS; although, the RSD for the response factor
values were in an acceptable range for all analytes with the exception of RVX, which showed an
RSD of 40%.
Table 3.5. Chemical Abstract Number, Retention Time, and Relative Retention
Time for the Analytes and Internal Standards Investigated
Analyte
d4-l,4-dichlorobenzene (IS)
GA
d8-naphthalene (IS)
HN1
HNS
dio-phenanthrene (IS)
RVX
CAS Number
3855-82-1
77-81-6
1146-65-2
538-07-8
555-77-1
1517-22-2
159939-87-4
RT
(min.)
9.64
11.56
12.51
11.99
15.31
17.99
17.56
RRT
NA
1.20
NA
0.96
1.23
NA
0.98
Notes: CAS=Chemical Abstracts Service Registry Number; RT=retention time;
RRT=relative (to the internal standard) retention time; IS=internal standard;
NA=not applicable.
13
-------
Table 3.6. Calibration Levels for GC/MS, Operated in Full-Scan Mode
Contaminant
GA
HN1
HNS
RVX
CAS
77-81-6
538-07-8
555-77-1
159939-87-4
Calibration Level (ng/uL)
1
0.025
0.025
0.025
0.8
2
0.05
0.05
0.05
1.0
3
0.1
0.1
0.1
1.6
4
0.5
0.5
0.5
2.0
5
0.8
0.8
0.8
2.5
6
1.0
1.0
1.0
3.0
Surrogates
ds-nitrobenzene
ds-phencyclidine
d14-terphenyl
triphenyl phosphate
4165-60-0
60124-79-0
1718-51-0
115-86-6
0.2
0.2
0.2
0.2
0.4
0.4
0.4
0.4
0.8
0.8
0.8
0.8
1.0
1.0
1.0
1.0
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
Internal Standards
d4-l,4-dichlorobenzene
d8-naphthalene
dio-phenanthrene
3855-82-1
1146-65-2
1517-22-2
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
Notes: All concentrations in parts-per-million (ng/(iL). All injections into the GC/MS were 1 (iL.
Table 3.7. Relative Response Factor (RRF), Relative Standard Deviation (RSD) for RRF,
and R2 Value (Linear Regression) for the Calibration Levels of Table 3.6, for GC/MS, Full-
Scan
Analyte
GA
HN1
HN3
RVX
Internal Standard
d4-l,4-dichlorobenzene
d8-naphthalene
d8-naphthalene
d10-phenanthrene
Mean RRF
0.221
0.503
0.287
0.124
RSDofRRF(%)
11.9
29.0
30.3
39.8
R2
0.9973
0.9964
0.9976
0.9973
Note: Quantitation ions are as identified in Table 3.2.
14
-------
Table 3.8. Calibration Levels for GC/MS, Operated in SIM Mode
Contaminant
GA
HN1
HNS
RVX
CAS
Number
77-81-6
538-07-8
555-77-1
159939-87-4
Calibration Levels (ng/uL)
1
0.01
0.01
0.01
0.1
2
0.02
0.02
0.02
0.2
3
0.04
0.04
0.04
0.4
4
0.08
0.08
0.08
0.6
5
0.1
0.1
0.1
0.8
6
0.25
0.25
0.25
1.0
Surrogates
nitrobenzene, ds
d5-phencyclidine
di4-terphenyl
triphenyl phosphate
4165-60-0
60124-79-0
1718-51-0
115-86-6
0.01
0.01
0.01
0.01
0.05
0.05
0.05
0.05
0.08
0.08
0.08
0.08
0.1
0.1
0.1
0.1
0.25
0.25
0.25
0.25
0.5
0.5
0.5
0.5
Internal Standards
d4-l,4-dichlorobenzene
d8-naphthalene
d10-phenanthrene
3855-82-1
1146-65-2
1517-22-2
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
Notes: All concentrations in parts-per-million (ng/(iL). All injections into the GC/MS were 1
Table 3.9. Average RRF, RSD for RRFs, and R2 Value (Linear Regression) for the
Calibration Levels, Shown in Table 3.8, for GC/MS, SIM
Analyte
GA
HN-1
HN-3
RVX
Internal Standard
1,4-dichlorobenzene, d4
naphthalene, d8
naphthalene, d8
phenanthrene, d10
Mean RRF
0.217
0.171
0.148
0.049
RSDofRRF(%)
13.8
44.0
35.0
35.7
R2
0.9934
0.9956
0.9972
0.9946
Note: Quantitation ions are as identified in Table 3.3.
15
-------
Table 3.10. Calibration Levels for GC/TOF-MS
Contaminant
GA
HN1
HNS
RVX
CAS
Number
77-81-6
538-07-8
555-77-1
159939-87-4
Calibration Levels (ng/uL)
1
0.01
0.01
0.01
0.10
2
0.02
0.02
0.02
0.2
3
0.04
0.04
0.04
0.4
4
0.08
0.08
0.08
0.6
5
0.1
0.1
0.1
1.0
6
0.2
0.2
0.2
2
7
0.4
0.4
0.4
3
8
0.8
0.8
0.8
4
9
1
1
1
5
Surrogates
nitrobenzene, ds
d5-phencyclidine
di4-terphenyl
triphenyl
phosphate
4165-60-0
60124-79-0
1718-51-0
115-86-6
0.01
0.01
0.01
0.01
0.02
0.02
0.02
0.02
0.04
0.04
0.04
0.04
0.08
0.08
0.08
0.08
0.1
0.1
0.1
0.1
0.2
0.2
0.2
0.2
0.4
0.4
0.4
0.4
0.8
0.8
0.8
0.8
1
1
1
1
Internal Standards
d4-l,4-dichloro-
benzene
d8-naphthalene
d10-phenanthrene
3855-82-1
1146-65-2
1517-22-2
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
Notes: All concentrations in parts-per-million (ng/(iL). All injections into the GC/TOF-MS were 1
Table 3.11. Average RRF, RSD for RRF, and R2 Value (Linear Regression) for the
Calibration Levels, Shown in Table 3.8, for GC/TOF-MS
Analyte
GA
HN-1
HN-3
RVX
Internal Standard
1,4-dichlorobenzene, d4
naphthalene, d8
naphthalene, d8
phenanthrene, d10
Mean RRF
0.398
0.263
0.211
0.050
RSDofRRF(%)
5.82
6.41
9.61
40.2
R2
0.9995
0.9998
0.9994
0.9999
Note: Quantitation ions are as identified in Table 3.3.
16
-------
3.5 Investigation of Method Detection Limits (MDLs)
Using the extraction procedures described in Table 2.1, environmental matrices
representative of soils, waters, and wipes were spiked with CWA, extracted, and analyzed to
study method performance and to estimate MDLs. The sample extraction procedure for solids
consisted of two sequential extractions - the first was an extraction with 25/50/25 (v/v/v)
Ace/DCM/EtOAc, and the second was extraction with 5/95 TEA/EtOAc. Only the RVX was
expected to be recovered to any significant extent in the 5/95 TEA/EtOAc extraction. Water was
extracted with DCM. Analyses were performed by electron ionization, full-scan, GC/MS and by
GC/TOF-MS. The work performed and the results obtained are summarized, organized by matrix
type, in the following sections.
3.5.1 Soils
Analyte recovery and MDL varied by soil type; see Tables 3.10-3.13. The percentages
of CWA removed by each extraction solvent varied; see Table 3.12 and Table 3.14. The intent
of Tables 3.12 and 3.14 is to show the solvent fraction in which the analyte of interest is
predominantly found and to show the matrices for which the target analytes are not recovered.
For all soils, RVX was extracted primarily in the TEA/EtOAc solvent. Neither GA nor RVX
were extracted efficiently (i.e., >20%) from Georgia Soil.
Total analyte recovery and MDL are shown in Tables 3.13 and 3.15. CWA recoveries on
the different soils were variable, indicating the complexity of the processes by which CWAs are
sorbed to and removed from the soils. The sorption of CWAs to soils results from combined
effects of interactions with many different inorganic and organic soil components and is not
easily predictable7. All of the CWAs tested in this study contain a nitrogen atom. The presence
of a basic nitrogen provides a point for chemisorption of the CWA onto soil particles and the
formation of hydrogen bonds. For soils with a pH lower than the pKa of the analyte, one might
not expect efficient removal of that analyte with organic solvents alone. Recoveries greater than
100% were observed, indicating matrix enhancement effects. For example, the RVX was
susceptible to the same matrix enhancement effects previously observed for VX. In general,
MDLs calculated using the GC/TOF-MS were lower than those observed by GC/MS.
Data suggest that the method used for extraction of GA, HN1, HN3, and RVX is not
optimal. Since the time that these data were collected, the CWA protocol has been modified so
that the two extraction solvents of Ace/DCM/EtOAc and TEA/EtOAc are no longer used; DCM
alone is now used to extract soils and wipes. GA, FDST1, and HN3 are expected to be efficiently
extracted from soils with DCM. Also, in the latest edition of the CWA protocol, a buffer of
tris(hydroxymethyl)aminom ethane at pH = 8.80 (Tris), followed by back-extraction of the VX
into DCM, is used to extract VX from soils. Both VX and RVX have recently been
o
demonstrated to be extracted from soils using the Tris buffer procedure. In that study, in which
GC/TOF-MS was used for the final analysis, the calculated MDL for RVX from Virginia A soil
(64.5% sand, 28% silt, 7.5% clay, and 2.6% TOC, with pH 4.1) was 18 jig/kg (calculated from a
spike level of 40 jig/kg).
17
-------
Table 3.12. Distribution of CWA Between the Different Extraction Solvents, with Analyses
by Full-Scan GC/MS
Matrix
Sand
Analyte
GA
HN1
HNS
RVX
Spike
Level
(ug/kg)
100
150
300
500
Percent Recovered in
Extract 1
25/50/25
Ace/DCM/EtOAc
75
62
40
30
Percent Recovered in
Extract 2
5/95 TEA/EtOAc
11
8
9
70
Nebraska
Soil
GA
HN1
HNS
RVX
100
150
300
500
82
36
95
22
18
42
0
39
Georgia
Bt2 Soil
GA
HN1
HNS
RVX
200
300
600
1000
7
0
32
0
12
100
68
11
Virginia
Soil
GA
HN1
HNS
RVX
100
150
300
500
81
2
77
27
19
55
19
73
Percent not
Recovered
14
30
51
0
0
22
5
39
81
0
0
89
0
43
4
0
18
-------
Table 3.13. Recovery and MDL for Selected of CWAs in Various Soils, With Analyses by
Full-Scan, GC/MS
Analyte
GA
HN1
HNS
RVX
GA
HN1
HNS
RVX
GA
HN1
HNS
RVX
GA
HN1
HNS
RVX
Spike
Level
(jig/kg)
100
150
300
500
100
150
300
500
200
300
600
1000
100
150
300
500
Total Recovery
(Extract 1 + Extract 2)
Sand
Measured
Cone, (ug/kg)
86 ±8
106 ± 11
147 ± 18
578 ±45
% Rec
86 ±8
70 ±7
49 ±6
116±9
MDL
(^g/kg)
26
35
57
142
Nebraska Ag Soil
170 ±33
118±15
286 ± 77
310±26
170 ±33
78 ±10
95 ±26
61±5
106
48
243
81
Georgia Bt2 Soil
37 ±2
445 ±81
831 ±63
66 ±29
19± 1
148 ±8
138±11
11±5
8
81
198
91
Virginia Soil
169 ±22
85 ± 19
287 ±39
631 ±68
169 ±22
57 ±12
96 ±13
126 ±14
70
58
121
213
Notes: Internal standards were added to the final sample extract to yield a concentration of 1 ug/mL.
Average percent recoveries (Rec.) and percent relative standard deviation (RSD) represent the
average of seven samples that were spiked, extracted, and analyzed independently.
aMDL = s x ^a=o.oi, for 6 degrees of freedom (3.143) and were based on the total amount of analyte
extracted in Extract 1 and Extract 2.
19
-------
Table 3.14. Distribution of CWA Between the Different Extraction Solvents, With Analyses
by GC/TOF-MS
Matrix
Sand
Analyte
GA
HN1
HNS
RVX
Spike
Level
(ug/kg)
1.25
2.5
12.5
50
Percent Recovered in
Extract 1
25/50/25
Ace/DCM/EtOAc
93
73
82
0
Percent Recovered
in
Extract 2
5/95 TEA/EtOAc
0
27
18
87
Virginia
Soil
GA
HN1
HNS
RVX
1.25
2.5
12.5
50
94
0
74
22
0
100
26
78
Percent not
Recovered
7
0
0
13
6
0
0
0
Table 3.15. Recovery and MDL for Selected of CWAs in Various Soils, With Analyses by
GC/TOF-MS
Analyte
GA
HN1
HN3
RVX
GA
HN1
HN3
RVX
Spike
Level
(US/kg)
1.25
2.5
12.5
50
1.25
2.5
12.5
50
Total Recovery
(Extract 1 + Extract 2)
Sand
Measured
Cone, (ug/kg)
1.17±0.10
4.00±0.18
15.4±0.50
43.3 ±4.6
% Rec
93 ±8
159 ±7
123 ±4
87 ±9
MDL
(ug/kg)
0.33
0.57
1.6
15
Virginia Soil
1.18±0.13
3.00 ±0.73
22.2 ±4.0
94.6 ± 15.7
94 ±10
120 ±29
177 ±29
189±31
0.39
2.3
12
49
Notes: Internal standards were added to the final sample extract to yield a concentration of 1 ug/mL.
Average percent recoveries (Rec.) and percent relative standard deviation (RSD) represent the average of
seven samples that were spiked, extracted, and analyzed independently by GC/MS operated in full-scan
mode. MDL = 5 x ^a=00i, for6 degrees of freedom (3.143) and were based on the total amount of analyte
extracted in Extract 1 and Extract 2.
20
-------
3.5.2 Water
Laboratory water was spiked with CWA and extracted with DCM, as previously
described. Analyte recovery and MDL for GA, HN1, HNS, and RVX in laboratory water are
shown in Tables 3.16 and 3.17. Extraction recoveries were highly variable. Nitrogen mustards
are susceptible to hydrolysis;9'10 for this reason, recoveries of HN1 and HN3 may be low.
Likewise, GA may also hydrolyze in aqueous solution.n MDLs for GA, HN1, HN3, and RVX
in water were 16, 2, 20, and 69 |ig/L, respectively, when determined using GC/MS and 0.1, 0.08,
0.7, and 22 |ig/L, respectively, when determined using GC/TOF-MS.
Table 3.16. Recovery, Standard Deviation, and MDL for GA, HN1, HN3 and RVX in
Laboratory Water Analyzed by GC/MS, Full-Scan
Analyte
GA
HN1
HNS
RVX
Spiked
concentration
(Hg/L)
28.6
28.6
28.6
114
Measured
concentration
(Hg/L)
30.0 ±5.2
9.0 ±0.6
18. 8 ±6.4
100 ±22
Recovery
(%)
105 ± 18
31±2
64 ±22
88 ± 18
MDL
(Mg/L)
16
1.8
20
69
Notes: Each recovery represents the average of seven samples, which were independently spiked.
extracted, and analyzed. Analyses were performed by full-scan GC/MS;
MDL = s x ta=0 01, for 6 degrees of freedom (3.143).
Table 3.17. Recovery, Standard Deviation, and MDL for GA, HN1, HN3 and RVX in
Laboratory Water Analyzed by GC/TOF-MS
Analyte
GA
HN1
HN3
RVX
Spiked
concentration
0.86
1.66
8.33
33.3
Measured
concentration
0.64 ± 0.04
0.16 ±0.03
0.97± 0.23
35±7.1
Recovery
75 ±4.7
9.5 ±1.6
11±2.8
108 ±21
MDL
0.13
0.084
0.72
22
Notes: Each recovery represents the average of seven samples, which were independently spiked,
extracted, and analyzed. Analyses were performed by GC/TOF-MS;
MDL = s x ta=0 .01, for 6 degrees of freedom (3. 143).
21
-------
3.5.3 Wipes
For determination of MDL by GC/MS, wipes were spiked with CWA, first extracted with
Ace/DCM/EtOAc, and then extracted with TEA/EtOAc, as described in Table 2.1. All of the
CWAs were effectively removed from the wipe with Ace/DCM/EtOAc, the first extraction
solvent (data not shown). Table 3.18 shows recoveries and calculated MDLs for the target
analytes, which are based on their recoveries from the first extraction solvent. As previously
observed with VX, RVX shows susceptibility to matrix enhancement effects. MDLs for GA,
HN1, HNS, and RVX on wipes were 1.7, 1.4, 3.7, and 46 ng/cm2, respectively (assuming
conversion to a surface area of 100 cm2 being sampled).
Table 3.18. Recovery, Standard Deviation, and MDL for GA, HN1, HN3 and RVX From
Wipes Analyzed by GC/MS, Full-Scan (Recoveries From Extract 1)
Analyte
GA
HN1
HN3
RVX
Spike
Level
(Mg)
0.50
0.75
1.5
2.5
Measured
concentration
(Mg)
0.80 ±0.05
0.88 ±0.04
1.85 ±0.12
4.05 ±1.46
Recovery
(%)
159 ±10
117±6
123 ±8
162 ± 58
MDLa
(Mg)
0.17
0.14
0.37
4.56
MDL
Wipe Area
(ng/cm2) b
1.7
1.4
3.7
45.6
Notes: Estimated MDL is based on the standard deviation of seven independently spiked, extracted, and
analyzed samples.
a Method detection limit (MDL) = s x ta=0,0i, for 6 degrees of freedom (3.143).
Assumes wipe area of 100 cm2.
For determination of MDL by GC/TOF-MS, wipes were spiked with CWA, and extracted
with a newer procedure that used a 15-minute extraction with one 15-mL aliquot of DCM.
MDLs for GA, HN1, HNS, and RVX on wipes were 0.11, 0.023, 0.35, and 4.4 ng/cm2,
respectively (assuming conversion to a surface area of 100 cm2 being sampled); see Table 3.19.
22
-------
Table 3.19. Recovery, Standard Deviation, and MDL for GA, HN1, HN3 and RVX From
Wipes Analyzed by GC/MS, Full-Scan (Recoveries From DCM Extraction)
Analyte
GA
HN1
HN3
RVX
Spike
Level
(ng)
25
50
250
2500
Measured
concentration
(ng)
31.8±3.6
8.3 ±0.7
105 ±11
2226 ± 140
Recovery
(%)
127 ±15
16.6 ±1.4
41.8 ±4.5
89.0 ±5.6
MDLa
(ng)
11
2.3
35
441
MDL
Wipe Area
(ng/cm2) b
0.11
0.023
0.35
4.41
Notes: Estimated MDLs are based on the standard deviation of seven independently spiked, extracted, and
analyzed samples.
a Method detection limit (MDL) = s x ta=OM, for 6 degrees of freedom (3.143).
Assumes wipe area of 100 cm2.
4.0 Conclusions
Sample extraction and analysis procedures were tested for their applicability to the
analysis of GA, HN1, HNS, and RVX. The utility of the method tested, in part, is dependent on
whether the analytical instruments can provide an optimum theoretical (detection) limit (OTL)
that is equal to or lower than the analytical target level (ATL) for a given CWA. The OTL of an
analyte was defined as the lowest detectable concentration of an agent that is calculated based on
a measured DDL and considering an extraction efficiency of 100% for the analytical procedure
being used (e.g., the ideal case). ATL is an analyte concentration that has been determined,
based on experiments and modeling, to be protective of human health and is used as one
benchmark by which an analytical method can be assessed. ATLs and OTLs based on analysis
by GC/MS and GC/TOF-MS are shown in Table 4.1. GC/TOF-MS was able to provide OTLs
capable of meeting analytical requirements for the methods tested in residential soils, waters, and
wipes for the four CWAs tested. GC/MS OTLs met analytical requirements for less than half of
the analytes tested in the varying matrices. Of the agents tested, RVX requires the lowest
detection limits and its detection at expected ATL limits by both GC/MS and GC/TOF-MS is
problematic in some sample types.
In addition to assessing OTLs, a more realistic gauge of method performance is the MDL.
Table 4.2 shows a comparison of MDLs determined in this study and ATLs. MDLs are expected
to be worse (i.e., higher concentrations) than OTLs, as they are partially influenced by analyte
recoveries. However, MDLs calculated using the standard EPA method12 may sometimes be
unrealistically low (i.e., the calculated value is a concentration at which the signal produced by
the analyte cannot be reliably distinguished from instrument noise). This may be attributed to
single-instrument, single-calibration, and(or) single-operator tests that result in estimates of the
standard deviation that are too small. Spiking the samples used to calculate MDL at too high of
a concentration (>5-fold higher than the expected MDL) may also result in the calculated MDL
value to be too low. GC/TOF-MS analysis provided satisfactory MDLs for all CWAs in sand;
GC/MS provided satisfactory MDLs for GA, HN1, and HNS; the MDL for RVX was
23
-------
approximately 3-fold higher than required to reach the ATL. For water, MDLs were able to meet
ATLs only for GA and HN1 by GC/TOF-MS. However, this observation may be misleading as
the recovery for HN1 in water was less than 10%. For wipes, as expected, GC/TOF-MS yielded
the best MDLs - MDLs for GA and FEN1 were capable of meeting expected ATLs; however,
MDLs for FINS and RVX were not able to meet the expected ATLs.
Table 4.1. Analyte Target Level (ATL) and Optimum Theoretical Detection Limit (OTL) Based on
Instrument Detection Limits (IDLs) Presented in Tables 3.2 and 3.4
Agent
GA
HNla
HN3a
RVX"
Residential Soil (mg/kg)
ATL1
0.01
0.01
0.042
OTL
GC/MS
0.0025
0.0025
0.0025
0.080
OTL
GC/
TOF-MS
0.0025
0.00050
0.0020
0.010
Water (iig/L)
ATL1
1.4
0.2
0.2
0.021
OTL
GC/MS
1.4
1.4
1.4
46
Wipe (ng)
OTL
GC/MS
25
25
1.1
25
5.7
800
OTL
GC-
TOF-MS
25
5
20
100
Notes: Assumes that ATL values for HN1 and HNS are similar to those of HD.
Assumes that ATL values for RVX are similar to those of VX.
c At this time, no risk-based criteria for surfaces have been designated as wipe ATLs.
Table 4.2. Analyte Target Level (ATL) and Method Detection Limits (MDL) Calculated for
This Study
Agent
GA
HNla
HN3a
RVX"
Residential Soil (mg/kg)
ATL*
0.01
0.01
0.042
MDL
GC/MSC
0.026
0.035
0.057
0.142
MDL
GC-
TOF-MSC
0.00033
0.00057
0.0016
0.015
Water (ug/L)
ATL
10
1.4
0.2
0.2
0.021
MDL
GC/MS
16
20
69
Wipe (ng)
MDL
GC/MS
170
140
0.72
370
22
4600
MDL
GC-
TOF-MS
11
2.3
35
441
Notes: Assumes that ATL values for HN1 and HN3 are similar to those of HD.
Assumes that ATL values for RVX are similar to those of VX.
CMDL for reagent sand.
At this time, no risk-based criteria for surfaces have been designated as wipe ATLs.
24
-------
The results of this work provide a point of reference against which to judge future
method improvements. Method improvements are needed to allow detection of CWAs at
expected target levels in most matrices. A comparison of ATLs and OTLs (see Table 4.1)
indicate that for some compounds neither GC/MS nor GC/TOF-MS provide detection limits to
meet expected analytical requirements. For example, the expected lower limit for RVX in water,
0.021 ng/L, cannot be detected using the tested sample preparation strategy and detection by
either GC/MS or GC/TOF-MS. In such cases, extraction methods with more specific detection
techniques such as gas chromatographic tandem mass spectrometry would be expected to
improve instrument detection limits (DDLs), perhaps even by as much as a factor of fifty.
Since this work was performed, the methods in the original protocol have been modified,
with dichloromethane (DCM) being the sole extraction solvent for GB, GD, GF, and HD in
wipes and soils. The major reason for this change was that, while the analysis method may have
been acceptably sensitive, the extraction method was not sufficiently robust - especially for
soils. For example, the use of ethyl acetate necessitated frequent maintenance of the gas
chromatography column. The use of DCM is expected to minimize maintenance times, possibly
reduce the matrix enhancement observed during sample analysis, and provide a reasonable
extraction solvent for GA, FDST1, and FENS on solid matrices. Tris(hydroxymethyl)-
aminomethane buffer, at pH = 8.80, followed by back-extraction of the of analyte into DCM, is
now used for the determination of VX in soil; this procedure has been shown, as discussed in
Section 3.5.1, to be adequate for the extraction of RVX from soils. In that same study, DCM
was shown to be an efficient extraction solvent for both VX and RVX from wipes. It is expected
that changes in the extraction solvent will have a lesser contribution to obtainable method
detection limits, with improvements on the order of five-fold or less.
Another method improvement shown to be useful in a previous study was the use of an
isotopically-labeled internal standard (d!4-VX). The affects of matrix enhancement were
mitigated because d!4-VX, nearly identical to that of the analytes of interest, was used as an
extracted internal standard against which to base quantification. The use of the more efficient
pressurized fluid extraction should also be considered. Such strategies are not expected to
significantly improve method detection limits, but will assist in improving method robustness
and reproducibility.
Ultimately, a method must be able to achieve a study's data quality objectives in order to
be useful. It has been assumed that measuring the target analytes at the ATL concentrations
cited here will be necessary (e.g. measured MDLs must be less than or equal to ATL
concentrations). It has been demonstrated that this is achievable for some analytes in certain
matrices (e.g. GA, HN1, and HNS in residential soils and GA in water). However, this is not
achievable for all analytes in all matrices. For this reason, the next step in the evolution of
CWA methods is to strive for lower detection limits. It is expected that the use of detection
technologies with greater specificity (e.g. tandem mass spectrometry), changes in sample
preparation (e.g. larger sample sizes, greater sample extract volume reduction, etc.), and changes
in solvent system will lead to improved method detection limits (MDLs).
25
-------
References
1 U.S. Environmental Protection Agency. Standardized Analytical Methods for Environmental
Restoration Following Homeland Security Events -SAM (Revision 6.0). U.S. Environmental Protection
Agency, Washington, DC, EPA/600/R-10/122, 2010.
2 U.S. Department of Defense. Chemical Agent Health-Based Standards and Guidelines Summary Table
2: Criteria for Water, Soil, Waste, as of July 2011. U.S. Department of Defense, U.S. Army Public Health
Command, Aberdeen Proving Ground, MD. Public Health Notice No. 0711-03, 2011.
3 U.S. Environmental Protection Agency. Risk-Based Criteria to Support Validation of Detection Methods
for Drinking Water and Air. U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-
08/021,2008.
4 U.S. Environmental Protection Agency. Semivolatile Organic Compounds by Gas
Chromatography/Mass Spectrometry (GC/MS). SW-846Method 8270D. Revision 4. February 2007.
U.S. Environmental Protection Agency: Washington, DC.
5 U.S. Environmental Protection Agency. Determinative Chromatographic Separations. Method 8000C.
March 2003.
6 U.S. Environmental Protection Agency. Semivolatile Organic Compounds by Gas
Chromatography/Mass Spectrometry (GC/MS). SW-846 Method 8270D. Revision 4. February 2007.
7 McGinley, P. M.; Katz, L. E.; Weber, W. J. Jr. A distributed reactivity model for sorption by soils and
sediments. 2. Multicomponent systems and competitive effects. Environ. Sci. Technol. 1993, 27, 1524-
1531.
8 Koester, C., Mulcahy, H., and Leif, R., Fast Methods for Analysis of Environmental Samples
Contaminated by VX: Sample Throughput Study, LLNL-TR-497391, August 2011.
9 Lee, J. Y.; Lee, Y. H.; Byun, Y.G. Characterization and study of piperazinium salts, degradation
products of nitrogen mustards by nuclear magnetic resonance spectroscopy and liquid chromatography-
mass Spectrometry. J. Chromatogr. A. 2012,1227(2), 163-173.
10 Chua, H.-C.; Lee, H.-S.; Sng, M.-T. Screening of nitrogen mustards and their degradation products in
water and decontamination solution by liquid chromatography-mass Spectrometry. J. Chromatogr. A.
2006,1102(1-2), 214-223.
11 Munro, N. B.; Talmage, S. S.; Griffin, G. D.; Waters, L. C.; Watson, A. P.; King, J. F.; Hauschild, V.
The sources, fate, and toxicity of chemical warfare agent degradation products. Environ. Health
Perspect. 1999, 707(72,), 933-974.
12 Glaser, J. A.; Foerst, D. L.; McKee, G. D.; Quave, S. A.; Budde, W. L. Trace analysis for waste waters
- method detection limit. Environ. Sci. Technol. 1981, 75, 1426-1435.
13 U.S. Department of Defense. Chemical Agent Health-Based Standards and Guidelines Summary Table
2: Criteria for Water, Soil, Waste, as of July 2011. U.S. Department of Defense, Washington, DC, 0711-
03,2011.
14 U.S. EPA. Risk-Based Criteria to Support Validation of Detection Methods for Drinking Water and
Air. U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-08/021, 2008.
26
-------
United States
Environmental Protection
Agency
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT NO. G-35
Office of Research and Development (8101R)
Washington, DC 20460
Official Business
Penalty for Private Use
$300
------- |