-------
Decontamination of Agent Yellow,
a Lewisite and Sulfur Mustard Mixture
Evaluation Report
National Homeland Security Research Center
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
-------
Disclaimer
The United States Environmental Protection Agency through its Office of Research and
Development's National Homeland Security Research Center funded and managed the research
described here under EPA Contract Number EP-C-10-001, Work Assignment Number 4-28 with
Battelle. This report has been peer and administratively reviewed and has been approved for
publication as an Environmental Protection Agency report. It does not necessarily reflect views
of the Environmental Protection Agency. No official endorsement should be inferred. The
Environmental Protection Agency does not endorse the purchase or sale of any commercial
products or services.
Questions concerning this document or its application should be addressed to:
Lukas Oudejans, Ph.D.
Decontamination and Consequence Management Division
National Homeland Security Research Center
Office of Research and Development
U.S. Environmental Protection Agency (MD-E343-06)
109 T.W. Alexander Drive
Research Triangle Park, NC 27711
Phone: 919-541-2973
Fax: 919-541-0496
E-mail: Oudeians.Lukas@epa.gov
m
-------
Acknowledgments
The following individuals are acknowledged for review of this document:
Dave Mickunas - EPA/Office of Solid Waste and Emergency Response/Office of
Superfund Remediation and Technology Innovation/ Environmental Response Team
Larry Kaelin - EPA/Office of Solid Waste and Emergency Response/Office of Emergency
Management/Consequence Management Advisory Division
Matthew Magnuson - EPA/Office of Research and Development/National Homeland
Security Research Center
Ramona Sherman - EPA/Office of Research and Development/National Homeland Security
Research Center (QA review).
Contributions by Battelle, Columbus, OH are greatly acknowledged
IV
-------
Table of Contents
Disclaimer iii
Acknowledgments iv
Acronyms and Abbreviations ix
Executive Summary x
1.0 Introduction 1
1.1 Purpose and Objectives 2
1.2 Test Facility Description 2
2.0 Experimental Methods 4
2.1 Chemical Agent and Spiking Coupons 4
2.2 Test Materials 5
2.3 Description and Application of Decontaminants 5
2.4 Extraction of Coupons 6
2.5 Derivatization of L 7
2.6 Analytical Methods 8
2.7 Method Development and Demonstration 9
2.7.1 Extraction Efficiency 9
2.7.2 Method Detection Limit 10
2.7.3 Neutralization of Decontaminant 10
2.7.4 Confirm Derivatization Does Not Interfere with HD Analysis 12
2.8 Persistence Testing 13
2.9 Decontamination Testing 13
2.10 Observation of Surface Damage 14
2.11 Extraction Efficiency 15
2.12 Decontamination Efficacy 15
3.0 Test Results 17
3.1 Method Development and Demonstration Results 17
3.1.1 Extraction Efficiency 17
3.1.2 Method Detection Limit 19
3.1.3 Neutralization of the Decontaminant 20
3.1.4 Confirm Derivatization Does Not Interfere with HD Analysis 21
3.2 Persistence Testing Results 22
3.3 Decontamination Testing Results 25
-------
3.3.1 Efficacy Results Using Bleach (Full Strength) 26
3.3.2 Efficacy Results Using Bleach (Dilute) 27
3.3.3 Efficacy Results Using Hydrogen Peroxide (3%) 29
3.4 Qualitative Evaluation of By-products 32
3.5 Observations of Damage to Coupons 37
4.0 Quality Assurance/Quality Control 38
4.1 Control of Monitoring and Measuring Devices 38
4.2 Equipment Calibrations 40
4.3 Performance Evaluation Audits 41
4.4 Data Quality Audit 42
4.5 QA/QC Reporting 42
5.0 Summary 43
6.0 References 48
VI
-------
List of Figures
Figure 1. L (L-l, L-2, and L-3) and degradation by-products 1
Figure 2. Derivatization of Lewisite with butanethiol, where R = C/iHg 7
Figure 3. DerL-1 mean recoveries by solvent and material (error bars show % relative standard
deviation) 18
Figure 4. HD mean recoveries by solvent and material (error bars show % relative standard
deviation) 19
Figure 5. Persistence of derL-1 on concrete and glass over time (error bars show % relative
standard deviation) 23
Figure 6. Persistence of HD on concrete and glass over time (error bars show % relative standard
deviation) 24
Figure 7. Average percent recovery of derL-1 from positive control coupons compared to spike
controls (100%) from all bleach (dilute) and hydrogen peroxide (3%) decontamination tests.... 24
Figure 8. Average percent recovery of FID from positive control coupons compared to spike
controls (100%) from all bleach (dilute) and hydrogen peroxide (3%) decontamination tests.... 25
Figure 9. Chromatogram showing peaks for derL-1, derL-2, and HD 33
Figure 10. Chromatogram showing peaks for the HD by-product bis (beta chloroethyl) sulfone.
34
Figure 11. Photographs of coupons before and after decontamination treatment 37
Figure 12. Comparison of the mass of HD recovered from metal positive control coupons and
test coupons after a 30 min reaction time (30), 60 min reaction time (60), and 30 min reaction
time with reapplication for additional 30 min (30+30) of bleach (dilute) or hydrogen peroxide
(3%) 46
VII
-------
List of Tables
Table ES-1. Summary of Average % Decontamination Efficacy with a 30 min Reaction Time., xi
Table 1. Test Materials, Descriptions, Sources, Size, and Preparation 6
Table 2. Gas Chromatographic/Mass Spectrometry Conditions 9
Table 3. DerL-1 and HD Extraction Efficiency Matrix (Repeated for Toluene, Hexane, and
Acetone) 10
Table 4. Quench Test Matrix 12
Table 5. Matrix to Test Impact of Derivatization on HD and IS Analysis 12
Table 6. Persistence Test Matrix 13
Table 7. Agent Yellow Decontamination Test Matrix 14
Table 8. DerL-1 and HD Extraction Efficiencies with Various Solvents 17
Table 9. MDLforderL-1 and HD in Hexane 20
Table 10. Hexane Extraction and Thiosulfate Quench Test Results 21
Table 11. Matrix to Test Impact of Derivatization on HD and IS Analysis 22
Table 12. Persistence Testing Results 23
Table 13. Bleach (Full Strength) Efficacy Results 27
Table 14. Bleach (Dilute) Efficacy Results 28
Table 15. Decontamination of Neat HD using Bleach (dilute) with a 30 min Reaction Time 29
Table 16. Hydrogen Peroxide (3%) Efficacy Results 30
Table 17. Decontamination of Neat HD using Hydrogen Peroxide (3%) with a 30 min Reaction
Time 31
Table 18. DF200 Efficacy Results 32
Table 19. Summary of Results for Qualitative Analysis of DerL-2 35
Table 20. Summary of Results for Qualitative Analysis of HD By-product (Sulfone) 36
Table 21. Quality Control Requirements 39
Table 22. Equipment Calibration Schedule 40
Table 23. Calibration Levels 41
Table 24. PE Results 42
Table 25. Summary of Average % Decontamination Efficacy with a 30 min Reaction Time 44
Table 26. Summary of Average % Decontamination Efficacy with a 60 min Reaction Time 45
Table 27. Summary of Average % Decontamination Efficacy with a 30 min Reaction Time with
Reapplication and Subsequent additional 30 min Reaction Time 47
vm
-------
Acronyms and Abbreviations
amu
BCVAA
°C
ccv
CVAA
cm
EPA
derL-1
derL-2
DF200
GC
HD
HMRC
hr
HSRP
L
L-l
L-2
L-3
MDL
mg
min
mL
mm
MS
NHSRC
NIST
PE
PTFE
QA
QAPP
QC
RSD
ISA
atomic mass units
bis(2-chlorovinyl) arsonous acid
degrees Celsius
continuing calibration verification
2-chlorovinyl arsonous acid
centimeter(s)
U.S. Environmental Protection Agency
common product from reaction of butanethiol with L-l and CVAA
common product from reaction of butanethiol with L-2 and
BCVAA
EasyDecon® DF200
gas chromatography
sulfur mustard
Hazardous Materials Research Center
hour(s)
Homeland Security Research Program
Lewisite; when L is used in this report it refers to the synthesized
product that is primarily L-l, low levels of L-2, and may contain
very low levels of L-3
2-chlorovinyl dichloroarsine
bis(2-chlorovinyl) chloroarsine
tris(2-chlorovinyl) arsine
microgram(s)
microliter
method detection limit
milligram(s)
minute(s)
milliliter(s)
millimeter(s)
mass spectrometry
National Homeland Security Research Center
National Institute of Standards and Technology
performance evaluation
polytetrafluoroethylene
quality assurance
quality assurance project plan
quality control
relative standard deviation
technical systems audit
IX
-------
Executive Summary
Limited data exist on decontamination approaches that neutralize vesicant properties of Lewisite
or chemical agent mixtures containing Lewisite. Recent work conducted under U.S.
Environmental Protection Agency's (EPA) Homeland Security Research Program (HSRP)
focused on decontamination solutions of neat Lewisite. Agent Yellow, a mixture of Lewisite (L)
and sulfur mustard (HD), is a chemical warfare agent mixture and decontamination approaches
to neutralize the vesicant properties of both components in this mixture are the focus of this
study.
The objective of this evaluation was to develop, demonstrate and apply methods to determine the
decontamination efficacies of various readily-available, liquids for the decontamination of Agent
Yellow. Agent Yellow, a mixture of the chemical warfare agents Lewisite (L) and sulfur mustard
(HD) was applied to 3.5 x 1.5 centimeter pieces (coupons) of four types of materials: sealed
concrete, wood flooring, galvanized metal, and glass. Residual Agent Yellow was extracted from
the coupons and analyzed for Lewisite (measured after derivatization [derL-1]) and HD using
gas chromatography/mass spectrometry (GC/MS). The extractions took place over time, to
evaluate persistence, or after decontamination, to evaluate efficacy of various decontaminants.
Persistence testing showed that natural attenuation aids in the removal of Agent Yellow from the
materials tested. Measurement of the persistence and decontamination of the Lewisite and HD
components of Agent Yellow were made at ambient laboratory conditions (17.8 degrees Celsius
(°C) to 20.3 °C and 7.5% to 53.5% relative humidity). Measurement of derL-1 includes detection
of the L-l vesicant (blister agent) hydrolysis by-product 2-chlorovinylarsonous acid (CVAA) as
both chemicals are derivatized to derL-1. Persistence measurements under these environmental
conditions showed that after application of Agent Yellow, less than 20% of the Lewisite
(measured as derL-1) was recovered from sealed concrete or glass after 4 hours; however 14%
and <3% of the Lewisite, respectively, were recovered after 18 hours. Similarly, 23% and 46%
of the HD was recovered from sealed concrete and glass, respectively after six hours at ambient
conditions. Little (4% from sealed concrete) or no (from glass) HD was recovered after 18 hours
at ambient conditions.
Decontamination efficacy was evaluated for four decontaminants: bleach (full strength), bleach
(10 fold dilute), hydrogen peroxide (3%), and EasyDecon® DF200 (DF200). A 30 min reaction
time was evaluated for all decontaminants and material combinations. Results are summarized in
Table ES-1. An additional reaction time (60 min) and a 30 min reaction time with a subsequent
reapplication of the decontaminant and additional 30 min reaction time were evaluated for some
combinations of decontaminants and materials.
With a 30 min reaction time, efficacy was observed for all four decontaminants against the
Lewisite (L-l and CVAA, measured as derL-1) component of Agent Yellow on all four material
-------
types. While Table ES-1 shows that all four decontaminants were generally efficacious against
the HD component of Agent Yellow, efficacies were generally lower for the HD component than
for the L-l component. Efficacies varied by material type as well as decontaminant. Increasing
the reaction time to 60 min or reapplying the decontaminant (for bleach [dilute] and hydrogen
peroxide [3%]) resulted in lesser amounts of L-l and HD being recovered. These improvements
in efficacy were modest at best.
Bleach (full strength; -6% sodium hypochlorite), bleach (dilute; about 0.6% sodium
hypochlorite), hydrogen peroxide (3%), and DF200 are highly efficacious against the Lewisite
component of Agent Yellow and exhibit varying levels of efficacy, depending on material and
decontaminant, against the HD component after 30 min reaction times. Bleach (full strength)
after a 30 min reaction time generally removed Lewisite to levels at or below the quantitation
limit (i.e., 2.0 jig/mL, the lowest value on the calibration curve) and exhibited efficacies for HD
of 37% to >95%. The efficacy ranges for bleach (dilute), hydrogen peroxide (3%), and DF200
were comparable to bleach (full strength) for Lewisite, but had lower efficacy ranges for HD.
Qualitative analysis for L-2 (bis[2-chlorovinyl] chloroarsine and its vesicant (blister agent) by-
product bis(2-chlorovinyl) arsonous acid (BCVAA; both derivatized to derL-2 prior to analysis)]
and the vesicant by-product of HD, bis(beta-chloroethyl)sulfone, showed that these chemicals
were generally not extracted from positive control coupons. Small chromatographic peaks
consistent with vesicant by-products were detected on some materials after application of each
decontaminant. DerL-2 chromatographic peaks ranging from 1% to 24% of the corresponding
derL-1 peak area were found on all coupon types after decontamination with bleach (dilute) for
both 30 and 60 min reaction times. HD by-product was found on all material types after
hydrogen peroxide (3%) decontamination with 30 and 60 min reaction times. Chromatographic
peak areas correlated with the HD by-product ranged from <1% to 38% of the corresponding HD
peak area across four materials. After the 60 min reaction time the ratio of HD by-product to HD
was generally higher than after the 30 min reaction time.
XI
-------
Table ES-1. Summary of Average % Decontamination Efficacy with a 30 min Reaction
Time
derL-l
Sealed Concrete
Wood Flooring
Galvanized Metal
Glass
HD
Sealed Concrete
Wood Flooring
Galvanized Metal
Glass
>92%
*t
>79%
>94%
>94%
37%
>37%
>95%
>87%
>66%
>94%
>86%
*
*
*
35%
>86%
>83%
>92%
>93%
*
80%
49%
42%
>92%
>86%
>87%
>95%
23%
*
*
57%
Efficacies (E) were calculated as E = [(C0 - Cf)/C0]'100%,where: C0 = mean mass per coupon of agent without decontamination
(determined from the positive control coupons of each material), and Cf = mass per coupon on a test coupon with
decontamination.
Efficacies shown as ">" had at least one and in most cases all test coupon extracts that had values less than the quantitation limit,
i.e., the calibration range (<20.0 jig/coupon).
*No significant difference (p>0.05) between the mean agent remaining on the positive control coupons and the mean agent
remaining on the test coupons, no significant efficacy is observed.
There was high variability in the positive controls so the p = 0.007; efficacy (although not significant) was 85%.
Impact of the Study:
Based on the results obtained in this study, the vesicant properties of Agent Yellow, a mixture of
HD and Lewisite can be neutralized by using full strength bleach. Other decontamination
products evaluated in this study neutralize the Lewisite component, but would leave significant
amounts of the HD component of Agent Yellow on all surfaces tested. Caution should be used in
extrapolating from bench testing to field application of these decontamination solutions.
Xll
-------
1.0 Introduction
Among its responsibilities related to homeland security, the U.S. Environmental Protection
Agency (EPA) has the goal of identifying methods and equipment that can be used for
decontamination following a terrorist attack using chemical, radiological, or biological agents.
The EPA's National Homeland Security Research Center (NHSRC) has been tasked to manage,
coordinate, support, and conduct a wide variety of homeland security research and technical
assistance efforts. In the interest of expanding our national readiness against highly-ranked threat
scenarios, NHSRC, as part of EPA's Homeland Security Research program (HSRP) is
conducting tests to evaluate the performance of products, methods, and equipment for
decontaminating contaminated materials. EPA has identified a lack of knowledge of simple
methods for decontaminating Agent Yellow as an important gap. Agent Yellow is a chemical
warfare agent that is a mixture of two blister agents or vesicants under the Chemical Weapons
Convention:1 Lewisite (L) and sulfur mustard (HD, bis[2-chloroethyl] sulfide; CAS 505-60-2).
Shown in Figure 1, L is a mixture of up to three compounds that are produced in the synthesis:
• L-l (2-chlorovinyldichloroarsine, CAS 541-25-3, Lewisite);
• L-2 (bis[2-chlorovinyl] chloroarsine, CAS 40334-69-8); and
• L-3 (tris[2-chlorovinyl] arsine, CAS 40334-70-1).2
L made for weapons in the United Kingdom comprised L-l, L-2, and L-3 in a 90:9:1 ratio.3 L-3
is not always present in L. (The L used in this study did not contain detectable quantities of L-3.)
"Lewisite" is sometimes used to refer only to L-l. In this report, L is used in the broad sense to
include L-l and L-2 (also L-3 if present).
L-l
C2H2AsCl3
L-2
C4H4AsCl3
L-3
C6H6AsCl3
^ I
Lewisite (L)
CVAA
ClCH=CHAs(OH)2
BCVAA
(ClCH=CH)2AsO
In water
solution
Slow dehydration
reaction; relatively
insoluble
Lewisite Oxide
ClCH=CHAs=0
Figure 1. L (L-l, L-2, and L-3) and degradation by-products.
Agent Yellow was commonly prepared by the Japanese in World War II as a 1:1 [by volume]
mixture of L and HD. Relatively undegraded Agent Yellow found in Yellow shells (the chemical
-------
weapon) comprised, weight/weight percent, HD (43.0%), LI (50.0%) and L2 (4.6%) of the
liquid content of the shells with other chemicals comprising less than 5% of the liquid.4
As shown in Figure 1, L-l rapidly hydrolyzes to 2-chlorovinyl arsinous acid (CVAA) in contact
with water and, with excess water and slowly, is transformed to 2-chlorovinyl arsine oxide
(Lewisite oxide).5'6 L-l, CVAA, and Lewisite oxide have similar vesicant properties.6 Lewisite
decontaminants would need to convert the various vesicant compounds to non-vesicant
compounds.
HD (C/jHgSCb) degradation by oxidation can yield a stable by-product with vesicant properties,
bis (beta-chloroethyl) sulfone (C4H8C12O2S) (CAS 471-03-4).7 Munro et al.6 provides an
extensive review of HD degradation products and impurities. Effective decontamination of HD
would need to degrade HD without producing this by-product.
/. / Purpose and Objectives
The overall purpose of this evaluation was to determine the decontamination efficacies of
various, readily-available, liquid-based methods for decontamination of Agent Yellow (including
L, HD, and by-products with vesicant properties) from various materials. Specific objectives
included:
• Evaluation of the persistence of Agent Yellow on sealed concrete and glass under
ambient laboratory conditions;
• Demonstration of efficiency of Agent Yellow extraction using three solvents: acetone,
toluene, and hexane;
• Determination of GC/MS method detection limits (MDL) for derL-1 and HD, i.e., bis[2-
chloroethyl] sulfide, in a mixture of L and HD using one solvent (hexane). DerL-1 is
ClCH=ChAs(SC4H9)2, the derivatization product of L-l, CVAA, and Lewisite oxide;
• Systematic evaluation of the efficacy of four decontaminants (bleach [full strength],
bleach [1:10 dilution], hydrogen peroxide solution [3%], and EasyDecon DF200) for
neutralization (conversion to a non-vesicant compound) of L and HD on 3.5 x 1.5
centimeter pieces (coupons) of four materials (sealed concrete, wood flooring, galvanized
metal, and glass); and
• Qualitative assessment of coupons for obvious visible damage resulting from application
of the decontaminants.
1.2 Test Facility Description
All testing was performed at the Battelle Hazardous Materials Research Center (HMRC) located
on the Battelle site in West Jefferson, Ohio. The HMRC is certified to work with chemical surety
material through its Bailment Agreement W911SR-10-H-0001 with the U.S. Department of the
Army.
-------
All testing was performed under ambient laboratory conditions. The temperature and relative
humidity in the laboratory were not controlled beyond normal heating and air conditioning. The
temperature and relative humidity were documented during each day of testing, both prior to and
following operations. The temperature in the laboratory during testing ranged from 17.8 degrees
Celsius (°C) to 20.3 °C and the relative humidity ranged from 7.5% to 53.5%.
-------
2.0 Experimental Methods
A posttest-only control group experimental design was used. In the following representation of
the experimental design, "X" is the experimental variable and "C represents a measurement or
observation. The experimental design is represented below in which time (represented by the
arrow) passes from left to right:
X - ^Ci
where coupons are spiked with Agent Yellow and randomly assigned as test coupons or positive
control coupons. Decontaminant applied to the test coupons is the experimental variable
(symbolized by "X"). Test coupons are decontaminated for a period of time and then extracted
and analyzed by GC/MS for derL-1 and HD (Observation 1; Ci). Positive control coupons are
not preceded by X, i.e., not decontaminated, but are extracted and analyzed for derL-1 and HD
(Observation 2; C2) at the same time as the test coupons. The effect of the treatment (efficacy) is
reported as the percentage of chemical agent remaining on treated coupons compared to the
control coupons:
Efficacy = [(C2-Ci)/C2]« 100% (1)
The higher the efficacy, the greater the effect of the decontamination. If there is no significant
difference (p>0.05) between the mean agent remaining on the positive control coupons and the
mean agent remaining on the test coupons, no significant efficacy is observed.
In addition to the test and control coupons, laboratory blank coupons (coupons that were neither
contaminated with Agent Yellow nor decontaminated) and procedural blank coupons (coupons
that were not spiked with Agent Yellow, but were decontaminated along with the test coupons)
were extracted and analyzed for derL-1 and HD. To verify the amount of Agent Yellow spiked
onto coupons, the same amount as applied to coupons was applied to polytetrafluoroethylene
(PTFE) disks and the disks were immediately extracted and analyzed as spike controls.
2. 1 Chemical Agent and Spiking Coupons
The L and HD (as received from the Army and owned by EPA) without any dilution is referred
to as neat L and neat HD. High (>90%) and stable agent purities are normally observed over long
periods of time (12 months or greater) following standard Battelle procedures for storage and
manipulation of L and HD.
A single batch of Agent Yellow ("neat") was prepared, sufficient for all testing performed as a
mixture of neat L and neat HD without any dilution. The purity of the neat L and neat HD used
to prepare Agent Yellow was determined by GC/flame ionization detector (FID). To determine
purity, the peak area value of the agent (L-l or HD) from GC/FID analysis was compared to the
-------
sum of all peaks present in the chromatogram (corrected by a solvent blank analysis result) and
calculated as a percentage. The neat L and neat HD sources were found to have purities of 92.2%
L-l (<0.2% L-2; L-3 not detected) and 96.2% HD, respectively, prior to use. Agent Yellow was
prepared as a mixture of neat L (63% by weight) and neat HD (37% by weight). Thus, the Agent
Yellow was theoretically 58% L-l and 36% HD by weight. Purity analysis of the Agent Yellow
following preparation yielded a value of 95.8%, indicating that L-l and HD peak areas
represented 95.8% of the compounds in the agent.
Neat Agent Yellow was dispensed using a 50 microliter (jiL) gas-tight syringe (P/N 80920,
Hamilton, Reno, NV) equipped with repeating dispenser (P/N PB600-1, Hamilton, Reno, NV)
and PTFE needle (P/N 9991326, Integrated Dispensing Solutions, Inc., Agoura Hills, CA). All
test and positive control coupons were spiked with 1 jiL of neat Agent Yellow. Each jiL of neat
Agent Yellow has a nominal mass of 1.6 mg consisting of 1.0 mg of Lewisite and 0.6 mg of HD.
The coupons were open to the atmosphere (uncovered in open Petri dishes) within a chemical
agent hood during a 30 min weathering period prior to application of decontaminants.
Subsequent to the weathering period the decontaminants were applied to the Agent Yellow on
the coupons. After an appropriate decontamination reaction time the coupons were transferred
into solvent for extraction.
2.2 Test Materials
Targeted materials included sealed concrete, wood flooring, galvanized metal, and glass (Table
1). Except for concrete, coupons were cut from larger pieces of material to 3.5 x 1.5 centimeters
(cm). Concrete coupons were poured into a mold and after setting were coated with sealer. Two
materials, glass and wood, were used for persistence testing.
2.3 Description and Application of Decontaminants
Four decontaminants were evaluated for efficacy against Agent Yellow on coupons:
• Bleach (sodium hypochlorite 5.65% to 6%, #SS290-1, Fisher Scientific)
• Bleach (sodium hypochlorite 5.65% to 6%, #SS290-1, Fisher Scientific) diluted 1:10
with deionized water, (#23-751-610, Fisher Scientific)
• Hydrogen peroxide (3%, # 88597-100ML-F, Fisher Scientific)
• EasyDECON® DF200 (DF200, EFT Holdings, Inc.).
-------
Table 1. Test Materials, Descriptions, Sources, Size, and Preparation
Material
Description
Manufacturer/
Supplier Name
Coupon
Surface Size
Length x
Width (cm)
Material
Preparation
Sealed
concrete
Wood
flooring
material
Galvanized
metal
ductwork
Glass
Epoxy (Sure Klean®
Weather Seal Siloxane PD;
PROSOCO, Inc., Lawrence,
KS) sealed concrete (5 parts
sand, 2 parts concrete);
custom preparation
Pine plywood (bare);
thickness 1.0 cm
Industry heating,
ventilation, and air
conditioning standard; 24
gauge galvanized steel;
thickness 0.7 millimeter
(mm) (Adept
Manufacturing)
Glass (clear window)
Wysong Concrete,
Cincinnati, OH
Lowe's,
Columbus, OH
Adept Products,
Inc., 3.5x1.5
West Jefferson, OH
Brooks Brothers,
West Jefferson, OH ' X '
Clean with
to remove
dust
Clean with
to remove
dust
Clean with
acetone
Clean with
to remove
dust
dry air
loose
dry air
loose
dry air
loose
The decontaminants were applied as a liquid to the test coupons 30 minutes (min) after the Agent
Yellow was spiked onto the coupons. The decontaminants were dispensed using a positive
displacement pipette (P/N M-250 [50-250 uL] and CP250 tip, Gilson Inc., Middleton, WI). The
amount of decontaminant applied to the coupons was 0.06 mL for non-porous materials (sealed
concrete, glass, and galvanized metal ductwork) and 0.09 mL for porous wood flooring. These
amounts reflect the approximate quantity of liquid decontaminant expected to remain on a
coupon of a specific type of material and size after a spray application.9
The initial reaction time for the decontaminants was 30 min. The decontamination testing was
repeated at a second reaction time (60 min) for combinations of coupons and decontaminant for
which the common product of the derivatization of derL-1 or HD was detected after the 30 min
reaction time. Application of decontaminant for a reaction time of 30 min followed by a
reapplication of the decontaminant for an additional 30 min reaction time was evaluated for
bleach (dilute) and hydrogen peroxide (3%) to determine if efficacy was improved.
2.4 Extraction of Coupons
All coupons in the test matrix and the blank coupons were extracted by placing each into a
separate 60 milliliter (mL) glass bottle (02-991-701, Fisher Scientific, Pittsburgh, PA) that
contained 10 mL of solvent/internal standard (IS), swirled by hand for about 5-10 seconds, and
-------
placing into a sonicator. Extraction bottles were sonicated at 40 to 60 kilohertz for 10 min at
ambient laboratory temperature. The IS is included in order to detect instrument drift and loss of
analyte in the injection process. The same extraction process was repeated for all coupons used
to determine extraction efficiencies, MDL, persistence, and decontamination efficacy. Samples
that were not analyzed the same day were stored at or below -20 °C.
2.5 Derivatization ofL
It is important to determine the presence of L-l, CVAA, and Lewisite oxide post
decontamination because these by-products each have harmful vesicant properties, rather than
analyzing only for L-l. This determination was accomplished in this work by derivatizing the
samples, as shown in Figure 2, and analyzing the samples using gas chromatography (GC)/mass
spectrometry (MS) to detect the derivative of L-l and its hydrolysis by-products (quantitative)
(derL-1) and the derivative of L-2 and its hydrolysis by-product (bis(2-chlorovinyl) arsonous
acid [BCVAA]) (qualitative) (derL-2). Lewisite oxide, not shown in Figure 2, yields the same
product by derivatization as L-l and CVAA. Derivatizing L by reaction with butanethiol prior to
GC/MS analysis also aids detection by creating products that are readily volatized and with
distinctive mass spectral fragmentation. Muir et al. (2005) showed that butanethiol reacted with
L-l and L-2, but did not react with L-3 or HD.3 L-3 can be detected without derivatization by
GC/MS. However, L-3 typically comprises only 1% of American L. No analysis for L-3 was
performed as the L used in this work was shown not to contain L-3.8
L-1
C2H2AsCI3
L-2
C4H4AsCI3
CICH=CHAs(SR)2 (CICH=CH)2AsSR
CVAA
CICH=CHAs(OH)2
BCVAA
(CICH=CH)2AsOH
Figure 2. Derivatization of Lewisite with butanethiol, where R = €419.
In the presence of water, L-l is converted to the vesicant CVAA. Derivatization is the preferred
approach for detection of both residual L and its vesicant by-product CVAA.6 Following the
method of Muir et al.3, 200 jiL of 50 milligram (mg) mL"1 butanethiol were added to 1 mL of
each Agent Yellow extract to be analyzed. Each sample/standard received 1% butanethiol per
volume. Triethylamine (50 ug) was added to each solution to catalyze the derivatization. Each
solution was mixed on a vortex mixer for 10 seconds. After mixing, samples were analyzed the
same day by GC/MS or were stored at or below -20 °C until they were thawed for analysis.
-------
2.6 Analytical Methods
Analysis for HD (quantitative) and its vesicant by-product bis(beta-chloroethyl)sulfone
(qualitative) was accomplished using GC/MS to identify the sulfone by its mass spectrum.
Quantitative analysis of derL-1 and HD was enabled by preparation of standard solutions of
these agents that were analyzed along with test samples. No standard was available for derL-2 or
the HD by-product and the results were therefore qualitative, indicating whether or not derL-2 or
the by-product is detected.
Blanks, positive control coupons, and decontaminated test coupons were extracted and
derivatized according to methods described in Sections 2.4 and 2.5. Aliquots of the sample
extracts were analyzed to quantify the amount of derL-1, and HD, and to detect derL-2 and bis
(beta-chloroethyl) sulfone, remaining on each coupon using GC/MS (6890 gas chromatograph
and 5973 mass selective detector, Agilent Technologies, Santa Clara, CA) operated in the full
scan mode for compounds ranging from 40 to 500 atomic mass units (amu). The combined L-l,
CVAA, and Lewisite oxide (if present) butanethiol derivatives (derL-1) were detected with
quantification ion 164 and qualifier ions 204, and 314. L-2 and BCVAA butanethiol derivatives
(derL-2) were detected with quantification ion 164 and qualifier ions 107 and 286. HD was
detected with quantification ion 109 and qualifier ions 158, 160 and 63. Bis (beta-chloroethyl)
sulfone was detected with quantification ion 63 and qualifier ions 65, 92 and 127. The GC/MS
parameters that were used in method demonstration and subsequent decontamination testing are
shown in Table 2.
The lowest standard used to establish the calibration curve (quantitation limit) was above, but
near, the instrument detection limit of the GC/MS. Samples with results below the lower
calibration level are reported as less than the quantitation limit.
-------
Table 2. Gas Chromatographic/Mass Spectrometry Conditions
Parameter
Analysis Method
Instrument
Data System
Liner Type
Column
Helium Carrier Gas Pressure and
Flow
Temperature Fields
Sample Size
Oven Program for Analysis
Target Ions, m/z
(% relative abundance versus first
ion)
Description
GC/MS
Agilent Model 6890 Gas Chromatograph equipped with a 5973
Mass Selective Detector and a Model 7683 Injector with
AutoSampler
MSD ChemStation
4mm Splitless
RTX-5MS, 30.0 meters x 0.25 millimeter, 0.25 micrometer film
15 pounds per square inch (40 ^C) at constant flow (37.7 cm s"1)
Inlet temperature: 250 °C; Detector temperature: 230 ?C
luL
50 °C (2.0 min), 325 °C (3.0 min) @ 30 °C/min ; total run time 14.5
min
derL-1: 164 (quantification ion), 204, 314
derL-2: 164 (quantification ion), 107, 286
HD: 109 (quantification ion), 158, 160, 63
Bis (beta-chloroethyl) sulfone: 63 (quantification ion), 65, 92, 127
2.7 Method Development and Demonstration
Method development and demonstration included determining the extraction efficiencies of three
solvents; determining the MDL for one solvent that would be used in subsequent testing; and
determining the adequacy of extraction alone for quenching decontamination.
2.7.1 Extraction Efficiency
Tests were performed in triplicate, as shown in Table 3 to determine Agent Yellow extraction
efficiency from four materials (sealed concrete, wood flooring, galvanized metal, and glass)
using three solvents (acetone, hexane, and toluene). Neat Agent Yellow (1 uL) was spiked onto
the test coupons and immediately extracted. In addition, three spike controls per solvent (a spike
of equal amount of Agent Yellow on a PTFE disk followed by immediate extraction) and
extraction of a single laboratory blank per material were included in the analysis. Coupons were
spiked, extracted, derivatized, and analyzed as described in Sections 2.1, 2.4, 2.5, and 2.6, except
that the extractions were immediate, rather than after a 30 min weathering period. Extraction
efficiencies were calculated as described in Section 2.11.
-------
Table 3. DerL-1 and HD Extraction Efficiency Matrix (Repeated for Toluene, Hexane, and
Acetone)
Material
Number of
Test Coupons
Number of Spike
controls
Number of
Laboratory Blank
Coupons
Sealed concrete
Wood flooring material
Galvanized metal ductwork
Glass
3
3
a
3
3
1
1
1
1
2.7.2 Method Detection Limit
After hexane was selected as the solvent that would be used for extraction in the
decontamination investigation, MDL studies for hexane extraction of a mixture of derL-1 and
HD on the four materials were performed. Seven coupons of each material type were spiked with
a dilute mixture of L and HD in hexane (relative proportions were selected to enable the MDL to
be determined for both agents using one solution). The single concentration design estimator
recommended by the EPA (40 CFR part 136, Appendix B (1984)) was completed as follows to
determine the MDL for derL-1 and HD:
• The coupons were extracted, the extracts derivatized and analyzed for derL-1 and HD
using GC/MS as described in Sections 2.4, 2.5, and 2.6.
• The standard deviations of the replicate measurements were calculated.
• The MDL = Student's t-value for n replicates appropriate for a 99% confidence level x
standard deviation estimate with n-1.
2.7.3 Neutralization of Decontaminant
The decontamination reaction must be stopped at the end of a specified contact period in order to
determine how much decontamination occurred during the contact period. The method
demonstration determined the conditions necessary to stop the decontamination reaction
(quench) without interfering with the extraction and analysis so that different decontamination
reaction times could be evaluated. Two approaches were used. The first approach, shown below,
evaluated the hypothesis that hexane extraction alone would be sufficient to quench the reaction.
The hypothesis that extraction alone, without additional neutralization, would be sufficient for
GC/MS analysis if the amount of derL-1 and HD recovered in Step 1 (decontamination solution
present) was each at least 70% of the amount of the respective CWA recovered in Step 2 (no
decontamination solution present). Initial method:
1. As a test solution, the amount of decontamination solution to be applied to coupons for
decontamination testing (60 microliter [uL]) was added (using a positive displacement
pipette (P/N M-100 [10-100 |iL] and D-200 [2-200 |iL] tip, Gilson Inc., Middleton, WI)
10
-------
to a vial containing 10 milliliters (mL) of hexane, IS (10 [j,g/mL naphthalene-dg [176044-
1G, Isotec [Sigma-Aldrich], St Louis, MO]), and 1 uL of neat Agent Yellow, shaken for
15 seconds, and allowed to stand for 10 min.
2. As a positive control, water (60 uL) was added to a vial containing 10 mL of solvent and
IS (naphthalene-dg) and 1 uL of neat Agent Yellow, the vial was shaken for 15 seconds
and allowed to stand for 10 min.
3. The extracts from Steps 1 and 2 were stored for at least 24 hours at or below -20 degrees
Celsius (°C) and then analyzed using GC/MS.
Because the initial approach did not confirm the hypothesis, a second approach was employed
and the addition of sodium thiosulfate as a quench was evaluated. Wood was selected as the most
challenging coupon type for the quench evaluation because more decontaminant was applied
during testing for wood (a porous material) than for other nonporous coupon materials (90 jiL
versus 60 jiL). Further, wood is the most absorptive of the coupon materials. In a porous
material, it may be more difficult for the quench to physically contact (and neutralize) the
decontaminant. A second set of coupons was used to test the effectiveness of the quench for the
worst case scenario (180 jiL total decontaminant in the reapplication test). The decontaminant
quench retest procedure follows:
• Positive control coupons were prepared by derivatizing L and HD (5.0 ug/mL final
concentration of each) in hexane with IS.
• A candidate decontamination solution (90 uL aliquot) was applied to a wood coupon and
two 90 uL aliquots of the decontaminant were applied to a second wood coupon. Both
wood coupons were then allowed to sit for 30 min. (This was repeated for each of the
four types of decontaminants.)
• Thirty min after the decontaminant was applied, the coupons were individually extracted
in hexane with 10 |ig/mL naphthalene-dg IS as described in Section 2.4.
• Eight 1-mL aliquots of the each wood coupon extracts (coupon spiked with 90 uL and
coupon spiked with 180 uL) were placed in vials.
• Quench (0.5 mL of 3M sodium thiosulfate) was added to four of the eight vials.
• 10 jiL of a 500 |ig/mL solution of HD and L were added to three of the four vials of each
condition (quench added and no quench added; final concentration of 5 |ig/mL HD and L
in each vial).
• All vials were vortexed for 10 seconds.
• Samples from each vial were aliquoted, derivatized and analyzed immediately, and then
analyzed again after being stored for 24 hours at room temperature.
11
-------
Table 4. Quench Test Matrix
Additives
No quench
addition
Wood (90 |iL decontaminant) Wood (180 |iL decontaminant)
Negative Control (no thiosulfate, L, or
HD)
Negative Control (no thiosulfate, L,
orHD)
Spike 10 ul of a 500 u.g/mL solution of
No quench HD and L in hexane and vortex
addition (5u.g/mLfinal concentration (triplicate
test)
Spike 10 u.L of a 500 u.g/mL solution
of HD and L in hexane and vortex
(5u.g/mLfinal concentration
(triplicate test)
0.5mL3M
sodium
thiosulfate
Negative Control (no L or HD)
Negative Control (no L or HD)
0.5mL3M
sodium
thiosulfate
Spike 10 u.L of a 500 u.g/mL solution of
HD and L in hexane and vortex
(5u.g/mLfinal concentration [triplicate
test])
Spike 10 u.L of a 500 u.g/mL solution
of HD and L in hexane and vortex
(5u.g/mL final concentration
[triplicate test])
2.7.4 Confirm Derivatization Does Not Interfere with HD Analysis
Tests were performed, summarized in Table 5, to confirm that the addition of triethylamine or
derivatization mixture (adding triethylamine and butanethiol), as described in Section 2.5, does
not interfere with the HD analysis. To perform the test, 1 uL of neat HD was spiked into 10 mL
of acetone. One-milliliter aliquots were placed into each of nine GC vials. IS consistent with
testing was added to three vials; triethylamine was added to three vials; and triethylamine and
butanethiol (consistent with the derivatization in Section 2.5) were added to three vials. Samples
were taken from each vial and analyzed for IS and HD as described in Section 2.6. Results were
considered acceptable if differences among the means of the treatments were within 30%.
Table 5. Matrix to Test Impact of Derivatization on HD and IS Analysis
Vial Containing HD and IS
Negative Control (triplicate test)
Additive
No addition
Triethylamine Enhancement Test (triplicate
test)
Triethylamine
Derivatization Test (triplicate test)
Triethylamine + butanethiol
12
-------
2.8 Persistence Testing
Persistence was evaluated at ambient conditions in the laboratory for Agent Yellow spiked onto
sealed concrete and glass as described in Section 2.1. These tests were included to assess the
natural attenuation of the agent on materials over time (up to 18 hours (hr) post contamination).
The persistence testing generally followed the same set-up and procedures (spike, extraction,
derivatization, and analysis) used for the decontamination efficacy testing with two exceptions:
no decontaminant was applied to any coupon, and coupons were covered (lids were placed on the
Petri dishes holding the coupons) during the Agent Yellow residence time. The temperature and
relative humidity during the persistence testing were in the range of 18.5 °C to 20.6 °C and 8.3%
to 13.7%, respectively.
The test matrix for Agent Yellow persistence is shown in Table 6. On the first day of testing,
coupons were extracted immediately after spiking ("0" hours) and at one, two, and four hours
after the coupons were spiked. On a second day of persistence testing, coupons were spiked with
Agent Yellow and extracted at 0, 6 and 18 hours after spiking. All coupons from the persistence
tests were extracted, the L derivatized, and the samples analyzed using GC/MS to quantify
residual derL-1 and HD. In addition to the test coupons, one laboratory blank coupon of each
material type was extracted and analyzed along with the test coupons on each day of persistence
testing.
Table 6. Persistence Test Matrix
Lab Day
1
Coupon
Type
Sealed
Concrete
Glass
Time Intervals, hours
3 coupons 3 coupons 3 coupons 3 coupons
3 coupons 3 coupons 3 coupons 3 coupons
3 coupons 3 coupons 3 coupons
3 coupons 3 coupons 3 coupons
2.9 Decontamination Testing
The Agent Yellow decontamination test matrix is shown in Table 7. For each combination of
time, material and decontamination method, five test coupons (spiked with neat Agent Yellow
and decontaminated), three positive control coupons (spiked with neat Agent Yellow, not
decontaminated) and two procedural control coupons (not spiked with Agent Yellow,
decontaminated) were included. Coupons were spiked, extracted, derivatized, and analyzed as
13
-------
described in Sections 2.1, 2.4, 2.5, and 2.6. One blank (negative control) coupon of each material
type was extracted and analyzed each day of testing. The two reaction times that were evaluated
were 30 min and 60 min. In addition, sequential 30 min reapplication of the decontaminant was
evaluated for bleach (dilute) and hydrogen peroxide (3%). Quantitative GC/MS analysis was
used for derL-1 and HD; qualitative GC/MS analysis was used to detect derL-2 and bis (beta-
chloroethyl) sulfone.
As shown in the last row in Table 7, decontamination of neat HD on two materials and with two
decontaminants was also evaluated. The materials selected were glass and wood and the
decontaminants applied were bleach (dilute) and hydrogen peroxide (3%).
Table 7. Agent Yellow Decontamination Test Matrix
Decontaminant
Repeated for each of the four
decontaminants and all four coupon
materials
Reaction Time
30 min
Coupons for each Combination of
Decontaminant and Material
5 test coupons; 3 positive control
coupons, two procedural blank
coupons; one blank coupon included
for each day of testing
Repeated for each of the four
decontaminants and all four coupon
materials, except full-strength bleach
on glass and sealed concrete were
excluded*
60 min
5 test coupons; 3 positive control
coupons, two procedural blank
coupons; one blank coupon included
for each day of testing
Repeated for bleach (dilute) and
hydrogen peroxide (3%) on all four
coupons types
30 min; then
reapplication after the
first 30 min of
reaction time; total
reaction time 60 min
5 test coupons; 3 positive control
coupons, two procedural blank
coupons; one blank coupon included
for each day of testing
5 test coupons; 3 positive control
coupons, two procedural blank
coupons; one blank coupon included
for each day of testing
Glass and wood were spiked with 1 ul
of neat HD and decontaminated with
bleach (dilute) and hydrogen peroxide
30 min
*No detectable derL-1 or HD remained on glass and sealed concrete after the 30 min decontamination with full-strength bleach,
so these material and decontaminant combinations were excluded from testing with a 60 min reaction time.
2.10 Observation of Surface Damage
The impact of decontamination on the building materials was assessed visually. Independent of
the agent work, one procedural blank of each material type was rinsed with deionized water and
allowed to dry. The procedural blank was visually inspected and compared to laboratory blank
coupons not exposed to the decontamination treatment to look for obvious changes in color,
reflectivity, or apparent roughness of the coupon surfaces. Observations were documented and
photographs of pre- and post-decontamination coupons were taken.
14
-------
2.11 Extraction Efficiency
Extraction efficiency was calculated using a series of equations as follows. Chemical agent
concentration in a coupon extract or solution sample was determined by the GC analysis
software (ChemStation) using the obtained sample area value.
GC concentration results (ng/mL) are converted to total mass by multiplying by extract volume:
Mm=CxEv (2)
where:
Mm = measured mass of chemical agent (microgram[s], ug)
C = GC concentration ug/mL), see Equation 1
Ev = volume of extract (mL).
Extraction efficiency was then defined as:
( M of Chemical Agent on Material \
Extraction Efficiency = —2-^ x 100% (3)
^ Mm of Chemical Agent on PTFE j
where:
Mm = measured mass of chemical agent (|j,g)
"Material" = coupons of materials used in the evaluation that are spiked with
chemical agent, extracted, and analyzed
"PTFE" = PTFE disks used in the evaluation as spike controls that are spiked with
chemical agent, extracted, and analyzed
Extraction efficiency = percent recovery of chemical agent from coupons.
2.12 Decontamination Efficacy
Decontamination efficacy was determined by measuring the amount of residual derL-1 and FID
on test coupons and comparing with positive control coupons (spiked with Agent Yellow, not
decontaminated) analyzed after the same elapsed time after spiking as the test coupons, i.e., 30
min weathering time plus reaction time. For derL-1 and HD, efficacy in percent was calculated
for each individual test coupon as shown in Equation 4. The primary efficacy results from the
coupon testing are provided in a matrix table in which each entry shows the mean and percent
relative standard deviation of efficacy results for derL-1 and HD on each of the materials.
15
-------
E = [(C0-Cf)/C0]-100% (4)
where:
E = efficacy
C0 = mean concentration of agent without decontamination (determined from the
positive control coupons of each material)
Cf = concentration on a test coupon with decontamination.
A Student's t-test was used to compare the amount of derL-1 and HD recovered from test
coupons to the amount of agent recovered from positive control coupons; p-values < 0.05 were
considered statistically significant. A separate t-test was performed for each material and for both
derL-1 and HD. If there is no significant difference (p>0.05) between the mean agent remaining
on the positive control coupons and the mean agent remaining on the test coupons, no significant
efficacy is observed.
Qualitative analysis of derL-2 was reported as the ratios of the areas of the chromatograph peaks
for derL-2 to derL-1. The presence of HD by-product was also noted.
16
-------
3.0 Test Results
3.1 Method Development and Demonstration Results
3.1.1 Extraction Efficiency
The results of extraction efficiency testing for Agent Yellow with acetone, hexane, and toluene
are shown in Table 8. Recoveries of derL-1 for the various solvent material combinations ranged
from 66% to 119%. In all cases the percent relative standard deviation was 19% or lower except
for derL-1 recovery from concrete using toluene; 1 of 3 concrete samples extracted with toluene
showed a 44% recovery while the other two samples showed recoveries >70%. Recoveries of
HD for the various solvent material combinations ranged from 78% to 122%. In all cases the
percent relative standard deviation (RSD) was at or below 15%. Note that extraction efficiencies
shown for PTFE spike compare the recovery from the PTFE disk to the theoretical mass applied.
The efficacy for materials used in testing compare the recoveries from materials to recoveries
from the PTFE disk (consistent with Equation 3).
Table 8. DerL-1 and HD Extraction Efficiencies with Various Solvents
Solvent
Acetone
Hexane
Toluene
Coupon
Material
PTFE
Spike
Concrete
Wood
Metal
Glass
Spike
Concrete
Wood
Metal
Glass
Spike
Concrete
Wood
Metal
Glass
derL-1
Mean,
Hg
1200
750
770
980
980
880
760
830
920
1000
1200
780
1100
1700
1100
derL-1
%RSD
13
15
17
15
6
4
9
15
9
7
12
30
11
19
5
derL-1
Extraction
Efficiency, %
H|
70
72
93
92
86*
87
95
100
120
110*
66
90
91
97
HD
Mean,
Hg
680
530
670
660
660
530
650
580
570
640
590
520
500
550
600
HD
%RSD
10
15
11
14
3
5
10
10
10
8
15
12
2
8
10
HD
Extraction Efficiency,
%
110*
78
97
97
96
89*
120
110
110
120
98*
88
85
92
100
*Compared to the theoretical mass of agent applied to the disk of 1.0 mg of Lewisite and 0.6 mg of HD.
17
-------
Figures 3 and 4 provide graphs showing a comparison of recovery efficiencies by material type.
Recoveries were generally lower for wood and concrete than for glass and metal for all three
solvents. Recoveries were at least 85% from all materials (including the PTFE disks for spike
controls) using hexane. Based on these data, hexane was selected as the extraction solvent for the
decontamination testing.
140.00%
120.00%
Spike Controls
I Concrete
I Wood
I Metal
I Glass
20.00%
0.00%
Acetone
Toluene
Solvent
Hexane
Figure 3. DerL-1 mean recoveries by solvent and material (error bars show % relative
standard deviation).
18
-------
140.00%
120.00%
I Spike Controls
I Concrete
Wood
I Metal
I Glass
20.00%
0.00%
Acetone
Toluene
Solvent
Hexane
Figure 4. HD mean recoveries by solvent and material (error bars show % relative
standard deviation).
3.1.2 Method Detection Limit
MDL studies for hexane extraction of a mixture of derL-1 and HD are summarized in Table 9.
For n = 7 coupons, t (n-1, l-a=0.99) = 2.998. Therefore:
MDL = 2.998 x standard deviation estimate with n-1.
(5)
For both derL-1 and for HD in hexane, MDL was lowest for wood (0.3 and 0.4 jig/coupon,
respectively). For derL-1, the highest MDL was for sealed concrete (0. 9 jig/coupon). For HD,
the highest MDL was for metal (1.2 jig/coupon).
19
-------
Table 9. MDL for derL-1 and HD in Hexane
Sample Source
Extracted from Sealed Concrete
Target
derL-1
HD
SD,
jig/coupon
0.3
MDL, lag/coupon
0.9
0.2
0.5
Extracted from Wood
derL-1
HD
0.1
0.1
0.3
0.4
Extracted from Metal
derL-1
HD
0.1
0.4
0.4
1.2
Extracted from Glass
derL-1
HD
0.2
0.2
0.5
0.5
3.1.3 Neutralization of the Decontaminant
The initial approach to evaluate the need for neutralization yielded anomalous results, e.g., HD
recovery in 5% bleach extract was greater than 200% higher than HD recovery from the positive
control condition. The results also suggested that extraction alone might not be a sufficient
quench for hydrogen peroxide (3%) or for DF200. A revised test procedure, described in Section
2.7.3, was used to both reevaluate the sufficiency of extraction alone to quench decontamination
and to evaluate the addition of sodium thiosulfate as a quench.
The results of the revised test (Table 10) showed recoveries of derL-1 and HD from hexane
containing decontaminants was 82% or more of the recoveries from hexane in the absence of
decontaminants. Similar results were observed when the extracts were stored for 24 hours at
room temperature and analyzed (data not shown). The addition of sodium thiosulfate was not
effective as a quench. Little or no derL-1 was found in extracts to which thiosulfate had been
added suggesting that thiosulfate degraded L. Thiosulfate did not impact the amount of HD
recovered from extracts. These results indicated that extraction alone was sufficient to terminate
decontaminati on.
20
-------
Table 10. Hexane Extraction and Thiosulfate Quench Test Results
Additives to
Hexane Containing
LandHD
Bleach (full
strength), %
Recovery vs.
Positive control
coupons
Bleach (dilute) %
Recovery vs.
Positive control
coupons
Hydrogen
Peroxide (3%)%
Recovery vs.
Positive control
coupons
DF200
% Recovery vs.
Positive control
coupons
derL-1 HD derL-1 HD derL-1 HD
90 nL
91 88 89 87 85 86
decontaminant
90 nL
decontaminant plus <37 86 <37 86 <37 87
thiosulfate
180 ul
90 87 89 87 82 85
decontaminant
180 ul
decontaminant plus <37 88 <40 87 <37 84
thiosulfate
derL-1 HD
100 92
<47 96
110 100
<40 100
Efficacies shown as "<" had at least one and in most cases all coupon extracts that were below the quantitation limit; quantitation
limit (2.0 ng/mL) substituted into calculation of the mean and reported as "<".
3.1.4 Confirm Derivatization Does Not Interfere with HD Analysis
Tests were performed, summarized in Table 11, to confirm that the addition of triethylamine or
derivatization mixture (adding triethylamine and butanethiol), as described in Section 2.5, does
not interfere with the HD analysis. To perform the test, 1 uL of neat HD was spiked into 10 mL
of acetone. One-milliliter aliquots were placed into each of nine GC vials. IS consistent with
testing was added to three vials; triethylamine and IS were added to three vials; and
triethylamine, IS and butanethiol (consistent with the derivatization in Section 2.5) were added to
three vials. Samples were taken from each vial and analyzed for IS and HD as described in
Section 2.6. The means of the results from the treatments differed by less than 8%. The results
with triethylamine or with triethylamine and butanethiol added were not significantly different
from the controls.
21
-------
Table 11. Matrix to Test Impact of Derivatization on HD and IS Analysis
Vial Containing HD (1 ul) in
10 ml acetone and IS
(naphthalene-dg [10 ng/mL])
plus:
Mean HD,
Hg/mL
(%RSD)
% HD Difference
Compared to
Control
IS Area
(%RSD)
% IS Difference
Compared to Control
(Student's t-test, n=3)
200 nL naphthalene-dg 10
Hg/mL (Control)
200 nL naphthalene-dg 10
Hg/mL and 250 ng/mL
triethylamine
200 nL naphthalene-dg 10
Hg/mL and 250 ng/mL
triethylamine and 1 mg/mL
butanethiol
100
(9.3)
94
(13)
94
(10.8)
7.4%
(p=0.73)
7.8%
(p=0.34)
419,857
(9)
389,573
(15)
386,282
(11)
7.2%
(p=0.63)
8.0%
(p=0.4)
3.2 Persistence Testing Results
Persistence was evaluated on sealed concrete and glass as described in Section 2.8. The test
results for Agent Yellow persistence are shown in Table 12. The percentages are in comparison
to recoveries from spike control disks (Agent Yellow spiked onto PTFE disks). Extraction
efficiencies are higher from glass than from the PTFE disks, resulting in time zero recoveries that
are greater than 100%. Persistence testing occurred on two separate days. Zero hour controls
were included on both days and results are similar. For derL-1, mean zero hour recoveries
(extracted immediately after spiking) were 62% on Day 1 and 69% on Day 2 for concrete.
Shown in Figure 5, derL-1 recoveries were somewhat higher from glass than from concrete
during the first hour, but were very similar at and after two hours. Recoveries of derL-1 from
both materials were about 20% after four hours. Shown in Figure 6, FID recoveries from concrete
parallels that of glass, but somewhat more HD is recovered from glass than from concrete during
the first six hours. Recoveries of HD from concrete and glass were minimal or non-detected at 18
hours.
22
-------
Table 12. Persistence Testing Results
Agent and
Test Day
Hours
Mean % of
Spike controls
%RSD
Mean % of
Spike
controls
%RSD
Concrete
0
1
Rc»rl 1 fR-31/ 1 \
uer L-I ^uay ij
4
0
DerL-1 (Day 2) 6
18
0
1
u n lr\-~t\i 1 \
HU ^uay ij
4
0
HD(Day2) 6
18
62
33
34
19
69
18
14
100
70
63
31
110
23
<6t
7
7
51
51
12
22
19
1
12
36
62
7
20
NAt
Glass
130
67
33
18
120
12
<3t
120
98
85
70
110
46
<2*
2
7
12
24
3
18
NAt
4
3
3
20
3
24
NA*
* Values below the quantitation limit, <20 (ig/coupon.
f Some values below quantitation limit.
1 ~)f) ^
>
sp mn
1-4"
i< Glass
5 4>
)C ^
\ ^
/ V
0 2 4 6 8 10 12 14 16 18 20
Time After Application, Hours
Figure 5. Persistence of derL-1 on concrete and glass over time (error bars show % relative
standard deviation).
23
-------
120 v
100 *
C?
x 80
g 60
o
01
c 40
ra
01
20
0
C
- £ '
If
/
f
)
f
\
+ Concrete
¥
2 4 6 8 10 12 14 16 18 20
Time After Application, Hours
Figure 6. Persistence of HD on concrete and glass over time (error bars show % relative
standard deviation).
During decontamination testing, the amount of VX and HD remaining on the coupons further
demonstrates persistence. Figures 7 and 8 show the average recoveries (compared to the spike
controls) from the extraction efficiency testing (with hexane) and the recoveries from positive
control coupons (compared to the spike controls) from all of the hydrogen peroxide (3%) and
bleach (dilute) testing. Persistence on wood and metal appears to follow a similar trend to
persistence on glass and concrete.
140
• Concrete
•Wood
• Metal
• Glass
60
90
Minutes after Applying VX to the Coupon
Figure 7. Average percent recovery of derL-1 from positive control coupons compared to
spike controls (100%) from all bleach (dilute) and hydrogen peroxide (3%)
decontamination tests.
24
-------
140
120
• Concrete
•Wood
• Metal
• Glass
0 60 90
Minutes after Applying HD to Coupon
Figure 8. Average percent recovery of HD from positive control coupons compared to spike
controls (100%) from all bleach (dilute) and hydrogen peroxide (3%) decontamination
tests.
3.3 Decontamination Testing Results
The results of decontamination testing after 30 min and 60 min reaction times with bleach (full
strength), bleach (dilute), hydrogen peroxide (3%), and DF200 are shown in Tables 13, 14, 15,
and 16, respectively. Sequential 30 min reapplication of the decontaminant was evaluated for
bleach (dilute) and hydrogen peroxide (3%). Quantitative GC/MS analysis was used to analyze
for derL-1 and HD. Qualitative GC/MS analysis was used to detect derL-2 and bis (beta-
chloroethyl) sulfone. Test results were defined as "efficacious" when the amount of agent
recovered from the test coupons was below the quantification limit. Efficacy results were
deemed "moderately efficacious" when the mean efficacy was at least 60% with quantifiable
amounts remaining on the test coupons. "Low efficacy" was defined for test results where the
efficacy values were less than 60% and amounts recovered from the test coupons were above the
quantification limit.
For all testing, no derL-1 or HD was measured on any of the laboratory blanks or procedural
control blanks. On some procedural blanks small peak areas consistent with derL-1 or HD were
sometimes observed but were below the quantitation limits.
25
-------
3.3.1 Efficacy Results Using Bleach (Full Strength)
Bleach (full strength) was efficacious against the L component of Agent Yellow decontaminated
from concrete, metal, and glass, as determined by measuring the decrease in derL-1. Efficacies,
shown in Table 13, ranged from >79 (metal) to >94 (glass) after a 30 min reaction time.
Recoveries from all coupon materials were at or below the limits of detection for derL-1 after a
30 min reaction time.
Bleach (full strength) bleach was generally efficacious against the HD component of Agent
Yellow decontaminated from all four materials. Efficacy was material dependent with the lowest
efficacies after 30 min reaction time (37%) observed on wood and metal. The highest efficacies
after a 30 min reaction time (>94%) were observed on concrete and glass. Increasing the bleach
(full strength) reaction time from 30 min to 60 min did not significantly improve removal of HD
or corresponding efficacies. (In these results and subsequent results a negative mean percent
efficacy is sometimes reported. This indicates that the amount of chemical agent recovered from
the test coupons after decontamination was greater than the amount of chemical agent recovered
from the positive control coupons. However, in no case was this figure significant. This should
be interpreted as no significant difference between agent recovered from test and positive control
coupons.)
The mass of HD recovered from the positive control metal coupons was unexpectedly low in the
30 min testing. The reason for this result is not known, but was observed again on metal during
testing with bleach (dilute) and hydrogen peroxide (3%).
26
-------
Table 13. Bleach (Full Strength) Efficacy Results
Analyte
Reaction
Time, min
Mean Positive Control Coupons
Total Mass of derL-1 or HD
Recovered, u,g (% RSD)
Concrete
Mean Test Coupon Total Mean%
Mass of derL-1 or HD Efficacy
Recovered, ug (% RSD) (p-value)
derL-1
HD
30
30
270 (50)
330 (38)
<20 (*)
<20 (*)
>92 (<0.05)
>94(<0.05)
Wood
r\c*r\ 1
der L-l
un
HU
30
60
30
60
140 (62)
99(8)
210(22)
120 (19)
21(9)
55 (53)
140 (25)
130 (29)
85 (0.07)
45 (<0.05)
37 (<0.05)
-1.5 (0.47)
Metal
-J — ,-1 1
aeri_--L
un
MU
30
60
30
60
94 (33)
180 (53)
32 (32)
260 (57)
<20(*)
22(20)
<20(*)
36 (87)
>79 (<0.05)
88(0.05)
>37 (<0.05)
86 (<0.05)
Glass
derL-1
HD
30
30
320(23)
410(21)
<20(*)
<20(*)
>94(<0.05)
>95 (<0.05)
*Not calculated because one or more values are below the quantitation limit, <20 jig/coupon).
3.3.2 Efficacy Results Using Bleach (Dilute)
Bleach (dilute), shown in Table 14, was moderately to fully efficacious (72% to 94%) against the
L component of Agent Yellow decontaminated from concrete, metal, and glass, as determined by
measuring the decrease in derL-1. Wood exhibited a lower efficacy, but little derL-1 was
recovered from wood after treatment or from positive control coupons. Low efficacy was
observed for dilute bleach (16% to 35%) against the HD component of Agent Yellow
decontaminated from concrete, wood, metal, and glass with a 30 min reaction time. Efficacy was
material dependent with the lowest efficacies observed on wood and metal.
27
-------
Table 14. Bleach (Dilute) Efficacy Results
Analyte
Reaction
Time, min
Mean Positive control
coupons Total Mass of derL-1
or HD Recovered, u.g (% RSD)
Concrete
Mean Test Coupon Total Mean%
Mass of derL-1 or HD Efficacy
Recovered, u.g (% RSD) (p-value)
derL-1
HD
30
60
30 + 30
30
60
30 + 30
160 (63)
140 (25)
330(20)
350 (55)
160 (31)
410(11)
<20 (*)
<20(*)
<20 (*)
230 (40)
110(51)
160 (42)
>87 (<0.05)
>86 (<0.05)
>94 (<0.05)
33 (0.14)
30 (0.14)
62 (<0.05)
Wood
derL-1
HD
30
60
30 + 30
30
60
30 + 30
170 (23)
150(21)
140 (24)
260 (14)
170 (26)
160 (19)
<59 (*)
<41 (*)
<24 (*)
220 (40)
140 (37)
120 (61)
>66 (<0.05)
>72 (<0.05)
>83 (<0.05)
16 (0.24)
18 (0.21)
27 (0.2)
Metal
derL-1
HD
30
60
30 + 30
30
60
30 + 30
350 (6.6)
160 (24)
120 (14)
420 (12)
260(27)
43 (23)
<21 (*)
<38(*)
<21 (*)
340 (28)
170 (4)
120(27)
>94 (<0.05)
>76 (<0.05)
>83 (<0.05)
19 (0.10)
35 (0.08)
-180 (<0.05)
Glass
derL-1
HD
30
60
30 + 30
30
60
30 + 30
300 (21)
290 (6.5)
230(5.7)
510(10)
460 (3.0)
380 (6.0)
41 (70)
80 (34)
<20(*)
340 (15)
180 (53)
170 (30)
86 (<0.05)
72 (<0.05)
>91 (<0.05)
35 (<0.05)
62(<0.05)
54 (<0.05)
*Not calculated because one or more values are below the quantitation limit, <20 jig/coupon).
Repeating the application of the dilute bleach generally appeared to increase efficacy more than
just increasing the reaction time. However, this does not always hold for HD, e.g., HD on glass
and metal.
Lower amounts of HD were recovered from all coupon types after reapplication of bleach
(dilute) compared to the single 30 min application. HD recovery results after reapplication of
bleach (dilute) were similar to the 60 min reaction time. The mass recovered from the positive
control metal coupons was unexpectedly low in the 30 min reapplication testing (30 + 30). In
28
-------
spite of low recoveries from metal after decontamination, efficacy was not observed because of
the unexpectedly low recoveries from the positive control coupons, i.e., there was little
difference between the recoveries from positive control or test metal coupons. Rust coloring on
the metal where the Agent Yellow was applied suggests corrosion may be related to the low
recoveries from the positive control coupons.
An additional decontamination test with bleach (dilute) was performed by spiking wood and
glass coupons with 1 jil of neat HD (no L) and repeating the decontamination test with a 30 min
reaction time. Results are shown in Table 15. The efficacies were 49% on wood and 30% on
glass. These efficacy values are similar to those observed for the HD component of Agent
yellow, namely, 16% on wood and 35% on glass.
Table 15. Decontamination of Neat HD using Bleach (dilute) with a 30 min Reaction Time
Analyte
Reaction Time,
Mean Positive control
coupons Total Mass of
derL-1 or HD Recovered, \ig
(% RSD)
Wood
Mean Test Coupon Total Mean%
Mass of derL-1 or HD Efficacy
Recovered, \ig (% RSD) (p-value)
HD
30
650 (41)
330 (44)
49 (<0.05)
Glass
HD
30
1200 (5.6)
830 (17)
30 (<0.05)
3.3.3 Efficacy Results Using Hydrogen Peroxide (3%)
Hydrogen peroxide (3%) results are shown in Table 16. Hydrogen peroxide (3%) was efficacious
(83% to 93%) against the L component of Agent Yellow decontaminated from all four materials,
as determined by measuring the decrease in derL-1 after a 30 min reaction time. In all cases,
after decontamination with hydrogen peroxide (3%) derL-1 extracted was below the limit of
detection.
Hydrogen peroxide (3%) was low to moderately efficacious (19% to 80%) against the HD
component of Agent Yellow decontaminated from all four materials after a 30 min reaction time.
Efficacy was material dependent with a moderate efficacy observed only on wood (80%); other
efficacies were below 60%; at 60 min recoveries were below the quantitation limits for one to all
five of the wood and metal coupons.
29
-------
Table 16. Hydrogen Peroxide (3%) Efficacy Results
Mean Positive control
coupons Total Mass of
derL-1 or HD Recovered,
ug(% RSD)
Analyte
Reaction Time,
Mean Test Coupon Total Mean%
Mass of derL-1 or HD Efficacy
Recovered, \ig (% RSD) (p-value)
Concrete
derL-1
HD
30
60
30 + 30
30
60
30 + 30
150(12)
180 (59)
390 (42)
310 (24)
230 (46)
370 (26)
<20(*)
<20(*)
<20(*)
260 (15)
83 (68)
140 (34)
>86 (<0.05)
>89 (<0.05)
>95 (<0.05)
19 (0.09)
63 (<0.05)
61 (<0.05)
Wood
derL-1
HD
30
60
30 + 30
30
60
30 + 30
120(13)
96 (16)
190 (12)
190 (22)
110 (14)
240 (31)
<20(*)
<20(*)
<20(*)
38(8)
<50(*)
<20(*)
>83 (<0.05)
>79 (<0.05)
>89 (<0.05)
80 (<0.05)
>55 (<0.05)
>91(<0.05)
Metal
30
derL-1 60
30 + 30
30
HD 60
30 + 30
250 (8)
250 (14)
110(19)
380 (7)
380(11)
<42 (*)
<20(*)
<20(*)
<20(*)
190 (36)
110 (26)
<43 (*)
>92 (<0.05)
>92 (<0.05)
>81 (<0.05)
49 (<0.05)
71 (<0.05)
-2.50(0.5)
Glass
30
derL-1 60
30 + 30
30
HD 60
30 + 30
270 (37)
240 (17)
200 (7)
420 (16)
400 (6.2)
360 (6.7)
<20(*)
<20(*)
<20(*)
240 (17)
120 (54)
190 (43)
>93 (<0.05)
>92 (<0.05)
>90 (<0.05)
42 (<0.05)
70 (<0.05)
47(<0.05)
*Not calculated because one or more values are below the quantitation limit, <20 (ig/coupon.
Lower amounts of HD were recovered from test coupons of all material types after reapplication
of hydrogen peroxide (3%) compared to the single 30 min application. HD recovery results after
30
-------
reapplication of hydrogen peroxide (3%) were similar to the 60 min reaction time. Shown in
Table 16, the mass of HD recovered from the positive control metal coupons was unexpectedly
low in the 30 min reapplication testing (30 + 30). In spite of low recoveries from metal after
decontamination, efficacy was not observed because of the unexpectedly low recoveries from the
positive control coupons, i.e., there was little difference between the recoveries from positive
control or test metal coupons. Rust coloring on the metal where the Agent Yellow was applied
suggests corrosion that may be related to the low recoveries from the positive control coupons.
No generalizations can be made regarding the improved efficacy of repeated application of the
hydrogen peroxide (3%) or the longer reaction times.
An additional decontamination test with hydrogen peroxide (3%) was performed by spiking
wood and glass coupons with 1 jil of neat HD (no L) and repeating the decontamination test with
a 30 min reaction time. Results are shown in Table 17. The efficacy was statistically significant,
but low (21% on wood and 10% on glass.)
Table 17. Decontamination of Neat HD using Hydrogen Peroxide (3%) with a 30 min
Reaction Time
Analyte
Reaction
Time, min
Mean Positive control coupons
Total Mass of derL-1 or HD
Recovered, u,g (% RSD)
Wood
Mean Test Coupon Total Mean%
Mass of derL-1 or HD Efficacy
Recovered, u,g (% RSD) (p-value)
HD
30
460 (26)
320(21)
21% (<0.05)
Glass
HD
30
1100 (5)
1000 (3.7)
10% (<0.05)
3.3.3.1 Efficacy Results Usins DF200
DF200, shown in Table 18, was efficacious (86% to 95%) against the L component of Agent
Yellow decontaminated from all four materials, as determined by measuring the decrease in
derL-1 after a 30 min reaction time. In all cases, after decontamination with DF200 derL-1
extracted was below the limit of detection. Because no derL-1 was detected after a 30 min
application of DF200, extending the reaction time to 60 min yielded no improvement in efficacy.
DF200 exhibited low efficacy (23% to 57%) against the HD component of Agent Yellow
decontaminated from all materials except metal after a 30 min reaction time. No significant
efficacy was observed for decontamination of HD on metal by application of DF200 after the 30
min and 60 min reaction times. Extending the reaction time for DF200 from 30 min to 60 min
resulted in little or no additional efficacy on the other three materials.
31
-------
Table 18. DF200 Efficacy Results
Analyte
Reaction Time,
Mean Positive control
coupons Total Mass of
derL-1 or HD Recovered,
ug(% RSD)
Concrete
Mean Test Coupon
Total Mass of derL-1 or
HD Recovered, u,g (%
RSD)
Mean%
Efficacy
(p-value)
-J — ,-1 1
aeri_--L
un
HU
30
60
30
60
250(15)
130(15)
380 (5)
270 (19)
<20(*)
<20(*)
290 (25)
240 (520)
>92 (<0.05)
>84 (<0.05)
23 (0.05)
12(0.3)
Wood
rlarl 1
HD
30
60
30
60
160 (28)
99 (15)
220 (33)
140 (12)
<22(*)
<20(*)
120
8
110 (38)
>86 (<0.05)
>80(*)
46
(0.07)
23 (0.13)
Metal
r\c*r\ 1
aer L-I
un
HU
30
60
30
60
160 (42)
110 (60)
90(120)
>150(*)
<20(*)
<20(*)
150 (42)
200 (29)
>87(*)
>81(*)
-68 (0.17)
-32 (0.22)
Glass
Hprl -1
un
MU
30
60
30
60
370 (35)
310(17)
520(1)
480 (7)
<20(*)
<20(*)
230 (29)
170 (51)
>95 (*)
>94(*)
57(<0.05)
64(<0.05)
*Not calculated because one or more values are below the quantitation limit, <20 (ig/coupon.
3.4 Qualitative Evaluation of By-products
Qualitative analysis for derL-2 and bis(beta-chloroethyl)sulfone (CAS 471-03-4), a vesicant by-
product of HD (qualitative) was accomplished using GC/MS. Results of the analyses are
summarized in Tables 19 and 20. The qualitative analysis of derL-2 and HD are reported as
detection of peaks and the ratios of the areas of the chromatograph peaks for derL-2 to derL-1
and the ratio of the areas of the chromatographic peaks of bis(beta-chloroethyl)sulfone (CAS
471-03-4) to HD.
On some metal and glass positive control coupons, small peaks were observed indicating the
presence of derL-2 (see Figure 9); none was detected on any concrete or wood positive control
coupons. After decontamination, significant ratios of derL-2 to derL-1 peaks were observed on
one or more types of materials. DerL-2 was always detected on metal after decontamination and
was found on all coupon types after decontamination with dilute bleach. In all cases, the peaks
32
-------
were small or very small compared to the peaks observed for derL-1 on the positive control
coupons.
Abundance
4800000
4600000
4400000
420000O
40ooooa
3800000
3600000
3400000
3200000
3OOOOOO
2800000
2600000
2400000
2200000
2000000
1800000
1SOOOOO
1400000
1200000
1000000
800000
600000
400000
200000
TIC: 030614017.D\data.ms
Time-;.
4.00 4.50 S.OO 5.50 6.00 6.50 7.00 7.50 8.00 8.50 9.00 9.30 10.0010.5011.0011.5012.0012,5013.0013.50
Figure 9. Chromatogram showing peaks for derL-1, derL-2, and HD.
The chromatographic peak representative of the HD by-product bis (beta chloroethyl) sulfone
(see Figure 10) was found on one of 24 wood positive control coupons, one of 24 metal positive
control coupons, and one of 24 glass positive control coupons, but not on any concrete positive
control coupons. Chromatographic peaks indicating the presence of the HD by-product were
observed on one or more coupon types after decontamination with all four products. After
hydrogen peroxide (3%) decontamination the by-product was detected on all coupon types and
the ratio of the by-product to residual HD was high (2% on concrete, 16% on wood, 20% on
33
-------
metal, and 38% on glass after 60 min reaction time). In all cases, the peaks were small or very
small compared to the peaks observed for HD on the positive control coupons.
Abundance
34000C
320000
300000
280000
260000
240000
220000
200000
180000
160000
140000
120000
100000
80000
60000
40000
20000
Naphthalene
Di-n-butyl
Sniffle
to
TIC; 030414056.D\data.ms
Bis(beta-
chloroethyl)
sulfone
"• I i i ' i I ' ' i i I i i i i I i I i i . . I 1.1.1.1.1, i i i i i , i , i i i i , i , i i , i
Time~> 3.50 4.00 4.50 5.00 5.50 6.00 6.50 7.00 7.50 8.00 8.50 9.00 9.50 10.00 10.50 11.00 11.50 12.00
Figure 10. Chromatogram showing peaks for the HD by-product bis (beta chloroethyl)
sulfone.
34
-------
Table 19. Summary of Results for Qualitative Analysis of DerL-2
Bleach (full strength) Bleach (diluted) Hydrogen Peroxide (3%) DF200
P°SitiV" DerL-2/derL-l (%)
control . '
or positive/ total
coupons,
test coupons, n=5
n=3
30: ND ND
Concrete
30:ND 2%
Wood
60:ND 1%
30:1% 5%
Metal
60:ND 3%
30:1/3 ND
Glass
P°SitiV" DerL-2/derL-l (%)
control . \
or positive/ total
coupons,
test coupons, n=5
n=3
30:ND 24%
60:ND 8%
30:ND 1%
60:ND 2%
30:ND 11%
60:ND 4%
30:ND 11%
60:ND 1%
Posffiv" DerL-2/derL-l (%)
control . \
or positive/ total
coupons,
test coupons, n=5
n=3
30:ND 1/5
60:ND ND
30:ND ND
60:ND ND
30: ND 9%
60: ND 2%
30:2/3 ND
60:ND 3/5
P°SitiV<: DerL-2/derL-l (%)
control : \
or positive / total
coupons,
test coupons, n=5
n=3
30:ND 5/5; for all
coupons but one
the derL-2 area
exceeded the
derL-1 area
60:ND ND
30:ND 2/5
60:ND 1/5
30: ND 1%
60:1/3 3/5
30:ND 1%
60:ND 24%
Notes: Two reaction times are show 30 min, indicated by "30", and 60 min, indicated by "60". "ND" indicates that no peaks indicative of derL-2 were noted on any coupon in that
group, e.g., in the first column, concrete, 30 min reaction time positive control coupons. A percentage shown indicates the ratio of the derL-2/derL-l converted to percentage. If the
ratio was <1%, the number of coupons positive for derL-2 out of total coupons in the group is shown, e.g., 2/5 indicates that derL-2 was detected on two of five coupons.
35
-------
Table 20. Summary of Results for Qualitative Analysis of HD By-product (Sulfone)
Material Bleach (full strength) Bleach (diluted) Hydrogen Peroxide (3%) DF200
Positive Sulfone/HD (%)
control or positive/ total
coupons, test coupons,
n=3 n=5
30:ND 1/5
Concrete
30:1/3 2%
Wood
60:ND ND
30:1/3 1/5
60:ND 5/5 and much
Metal higher area on
each coupon
than HD
30:ND 4/5
Glass
60:ND ND
Positive Sulfone/HD (%)
control or positive/ total
coupons, test coupons,
n=3 n=5
30:ND ND
60:ND ND
30:ND ND
60:ND ND
30:ND ND
60:ND 1%
30:ND 1/5
60:ND 4%
Positive Sulfone/HD (%)
control or positive/ total
coupons, test coupons,
n=3 n=5
30:ND 3%
60:ND 2%
30:ND 4/5
60:ND 16%
30:1/3 12%
60:ND 20%
30:ND 5%
60:ND 38%
Positive Sulfone/HD (%)
control or positive/ total
coupons, test coupons,
n=3 n=5
30:ND 1/5
60:ND 1%
30:ND ND
60:ND ND
30:ND 8%
60:ND 1/5
30:ND 1%
60:1/3 1%
Notes: Two reaction times are show 30 min, indicated by "30", and 60 min, indicated by "60". "ND" indicates that no peaks indicative of the HD by-product were noted on any
coupon in that group, e.g., in the first column, concrete, 30 min reaction time positive control coupons. A percentage shown indicates the ratio of the HD by-product converted to
percentage. If the ratio was <1%, the number of coupons positive for the HD by-product out of total coupons in the group is shown, e.g., 2/5 indicates that the HD by-product was
detected on two of five coupons.
36
-------
3.5 Observations of Damage to Coupons
Example photographs before and after the decontamination treatment are shown in Figure 11.
During the surface damage test, bleach (full strength) showed slight discoloration (lightening) of
the wood (Figure 1 Ib). Except for the bleach (full strength) on wood, the decontamination
treatment resulted in no obvious visible change to any of the coupons Figures 1 la, c, and d).
During the efficacy testing, Agent Yellow droplets would lead to rust-colored corrosion of metal
coupons where the drop was applied (caused by the Agent Yellow droplet rather than any of the
decontaminants).
a. Bleach (dilute) on concrete; left is before, right is after application.
b. Bleach (full strength) on wood; left is before, right is after application.
c. DF200 on metal; left is before, right is after application.
d. Bleach (dilute) on glass; left is before, right is after application.
Figure 11. Photographs of coupons before and after decontamination treatment.
37
-------
4.0 Quality Assurance/Quality Control
4.1 Control of Monitoring and Measuring Devices
Quality control (QC) requirements and results are shown in Table 21. All results were
acceptable.
One parameter in the test/Quality Assurance (QA) plan, "Agent (derL-1 and HD) on Glass
Positive Controls, ug/mL," to verify chemical agent amounts spiked onto coupons had limited
usefulness because the glass positive control coupons were not extracted immediately. Rather,
they controlled for agent losses, e.g., to evaporation, during the time required for
decontamination. Extraction of chemical agent from PTFE disks at time zero, the spike controls,
is the basis for determining extraction efficiencies from other materials. The spike control values,
rather than the positive control values, should indicate that the mass of agent extracted from the
disk (compared to the mass applied) was within an acceptable range. The spike controls using
PTFE disks were all in the range of 70% to 120% with %RSD <30% (data not shown).
While the spike controls are more appropriate for evaluating adequacy of recoveries, extraction
of L (derL-1) and FID from glass test coupons gave higher recoveries (derL-1 at 119% and HD at
120%) than have been observed from PTFE disks. The derL-1 recovery %RSD for the glass
positive control coupons after 30 min weathering were all <30% except during the hydrogen
peroxide (3%) test that had a %RSD of 37% and the DF200 test that had a %RSD of 35%.
38
-------
Table 21. Quality Control Requirements
Parameter
Temperature,
degrees Celsius (°C)
Measurement
Method
Thermometer
Data Quality Indicators
Compare to calibrated National Institute of
Standards and Technology (NIST)-traceable
thermometer once before testing, agree ±1 °C
Results and Corrective Action
Accuracy of thermometer was within
±1°C limits.
Relative humidity,
Hygrometer
Compare to calibrated NIST-traceable hygrometer
once before testing, agree ±10% (full scale)
Accuracy of hygrometer was acceptable.
Time, sec
Timer/data logger Compare once before testing; agree ±2 sec/hour Accuracy of laboratory clock was acceptable.
Volume, u.L
Calibrated pipette
and repeating
dispenser/syringe
Pipettes and repeating dispenser/syringe will be
checked for accuracy and repeatability before use
by determining the mass of water delivered. The
pipette will be acceptable if the range of observed
masses for five droplets is ±10% of expected.
1-10 u.L pipette - < 3% average error;
50-250 u.L pipette - < 1% average error;
100-1000 u.L pipette - < 1% average error;
50ul syringe - < 3% average error
Agent (derL-1 and
HD) on Glass
Positive Control
Coupons, u.g/mL
Extraction, GC/MS
The mean percent recovery for a known quantity of
each analyte added to a test coupon or an IS used
to gauge recovery must fall within the range of 70%
to 120% and have a coefficient of variation of
<30% between replicates
Recoveries of agent were acceptable and, with
two exceptions discussed in the text, coefficient
of variation was with acceptable limits. Variance
was noted. Because the extractions were
occurring 30 min after application rather than
immediately after application, and the recoveries
from the spike controls were within the target
range and variation, no changes were made. See
discussion in text.
Agent on
Laboratory Blank
Coupons, u.g/mL
Extraction, GC/MS
Laboratory blanks (coupons without applied agent
that are not decontaminated) should have less
than 1% of the amount of analyte compared to that
found on positive control coupons
No measurable agent detected on laboratory
blank coupons.
Agent on
Procedural Blank,
u.g/mL
Extraction, GC/MS
Procedural blanks (coupons without applied agent
that are decontaminated) should have less than
5% of the amount compared to that found on
positive control coupons
No measurable agent detected on procedural
blank coupons.
39
-------
4.2 Equipment Calibrations
The instrumentation used for the analyses are identified in Section 2.6. The required analytical
equipment was maintained and operated according to the quality requirements and
documentation of the HMRC. All equipment was calibrated at the time of use and at the
frequency specified in Table 22.
Table 22. Equipment Calibration Schedule
Equipment
Calibrated pipette and repeating
dispenser/syringe
Frequency
Prior to testing and every six months thereafter
Calibrated Hygrometer/Thermometer
Prior to testing and annually thereafter
GC/MS
Beginning of each batch of test samples
(calibration curve) and a calibration verification
standard every six samples and at the end of a
batch of samples
Neat L and neat HD were used to prepare stock solutions. The L and HD stocks
(concentrations corrected for percent purity) were used to create calibration standards
encompassing the appropriate analysis range. L calibration standards were derivatized prior to
use. The GC was maintained in calibration such that the coefficient of determination (r2) from
the regression analysis of the standards was more than 0.98. In addition, the percent bias for
the low standard was less than 25%, and the percent bias for the remaining standards was less
than 15%. The GC/MS was tuned initially and as needed following manufacturer's
guidelines. A tune check was performed before each analytical run using
decafluorotriphenylphosphine (DFTPP). A 12-hour tune time was not employed.
Five-point high and low, overlapping calibration curves for derLl and FID were used with an
overall lower calibration level of 2 |ig/mL and upper level of 150 |ig/mL. Any sample
exceeding the upper calibration limit was diluted to a concentration within the calibration
range and reanalyzed. Table 23 provides the high and low calibration curve standard levels
used during sample analysis.
Except as noted in the deviations, one continuing calibration verification (CCV) check
standard was analyzed at the beginning and end of each run and after every five samples.
Each analytical run included CCV standard at two different concentrations, with the low
standard set to the same level as the lowest calibration standard and the second set near the
midpoint of the curve (5.0 |ig/mL and 50.0 |ig/mL for the high curve and 2.0 |ig/mL and 5.0
|ig/mL for the low curve). The two concentrations were alternated throughout the run. The
40
-------
percent bias for the low calibration check standard was less than 35%, and the percent bias for
the remaining calibration check standards were less than 20%.
Table 23. Calibration Levels
Level 1
Level 2
Level 3
Level 4
Level 5
High Curve
5.0 |ig/mL
10.0 |ig/mL
25.0 |ig/mL
50.0 |ig/mL
150.0 |ig/mL
Low Curve
2.0 |ig/mL
2.5 |ig/mL
4.0 |ig/mL
5.0 |ig/mL
10.0 |ig/mL
Standards do not exist for derL-2 or BCAA so only a qualitative analysis of these species was
performed. The peak area values for derL-2 and BCAA detected in each sample were reported as
well as a peak area ratio to the corresponding derL-1 and HD values, respectively.
4.3 Performance Evaluation A udits
A performance evaluation (PE) audit was conducted, summarized in Table 24. Acceptable values
were: volume (±10%), time (±1 sec/min), chemical mass (>85%), IS (±10%), temperature (±1
°C), and relative humidity (±10%).
41
-------
Table 24. PE Results
Parameter
Volume
Audit Procedure
Pipette used for dispensing chemical agent will
be checked for accuracy and repeatability one
time before use by determining the mass of
water delivered
Results
Pipettes < 4%
Syringe < 6%
Time
Compare time to independent clock one time
before use
0.0 sec/min
Chemical Mass
Use GC/MS to determine mass of agent
delivered to PTFE spike control disks and
compare to target application level one time
DerL-1 and HD were above the
acceptance of >85%:
DerL-1: 103%
(range 88% to 115%)
HD: 113%
(range 101% to 122%)
Internal Standard
Use GC/MS to measure from a secondary
source(Supelco Product #48715-U, Lot
LC04085, Exp. 11/30/2016) and compare to
the primary source (Isotec [Aldrich] Product #
176044-1G, LotTV1320, Exp. 5/20/2015) one
time
2% relative percent difference
Temperature
Compare against calibrated National Institute
of Standards and Technology (NIST)-traceable
thermometer one time before use
Relative
Humidity
Compare against calibrated NIST-traceable
hygrometer one time before use
4.4 Data Quality Audit
The QA Manager audited at least 10% of the evaluation data and traced the data from initial
acquisition, through reduction and statistical comparisons, to final reporting. All data analysis
calculations were checked. Only minor calculation issues were noted with the data that did not
have a significant bearing on the reported data. These issues were corrected.
4.5 QA/QC Reporting
QA/QC procedures were performed in accordance with the QAPP for this study. Two deviations
that were not covered by two amendment to the initial QAPP were related to a longer time (by 7
min) between completion of the 10-min sonication and taking of an aliquot into GC vials during
the MDL study and one occasion were all required CCV standards were inadvertently omitted
from the repeat (low calibration) run for a decontamination test for process blanks and laboratory
blanks (no test coupons). Both deviations are expected to have a negligible impact on the results
of this study.
42
-------
5.0 Summary
The objective of this evaluation was to develop, demonstrate and apply methods to determine the
efficacies of various readily-available, liquid-based methods for the decontamination of Agent
Yellow, a mixture of L and HD, from materials. In addition, the persistence of Agent Yellow on
four building materials was determined.
Method development was used to determine extraction efficiencies for L and HD extracted from
four materials (sealed concrete, wood flooring, galvanized metal, and glass). Three solvents were
evaluated: acetone, hexane, and toluene. Efficiencies varied by material. Recoveries of L
(measured as derivatized L, derL-1) across the three solvents was lowest from concrete: 66%
with toluene, 70% with acetone, and 87% with hexane. Hexane overall had the highest
efficiencies (glass and metal at >100% (relative to PTFE recovery), wood at 95%. Recoveries of
HD from all material/solvent combinations were 78% (concrete extracted with acetone) or
higher. From hexane, recoveries were >100% (107% to 122%) relative to recoveries from PTFE.
Based on the results, hexane was selected for extractions in subsequent testing.
The GC/MS method MDL for derL-1 in hexane ranged from 0.03 |ig/mL from wood to 0.09
|ig/mL from concrete. The GC/MS MDL for HD in hexane ranged from 0.04 |ig/mL from wood
to 0.12 |ig/mL from metal.
Extraction with hexane alone was shown to provide adequate neutralization of the
decontaminants (with both 90 and 180 jiL of decontaminant included) with 82% to 108% of the
derL-1 measured in hexane to which decontaminant was added. Similarly, extraction with
hexane alone was shown to provide adequate neutralization of the decontaminants with 85% to
101% of the HD measured in hexane to which decontaminant was added compared to HD in
hexane to which no decontaminants were added. (Note that addition of thiosulfate as a quench
had little or no effect on HD recovery, but reduced measured amounts of derL-1 to <47% of the
positive control coupons with both 90 and 180 jiL of decontaminant included.)
The addition of triethylamine and butanethiol for derivatization of the Lewisite was shown not to
interfere with the measurement of HD by GC/MS. Measured concentrations in the presence of
triethylamine or triethylamine and butanethiol were 94% and 94%, respectively compared to
controls.
Measurement of the persistence of the L and HD components of Agent Yellow at ambient
laboratory conditions showed that after application of Agent Yellow, less than 20% of the L
(measured as derL-1) was recovered from concrete or glass after 4 hours; however 14% of the L
was recovered from concrete and <3% of the L was recovered from glass after 18 hours.
However, only one of three glass coupon had a measurable level of L after 18 hours.
43
-------
Decontamination efficacy was evaluated for four decontaminants: bleach (full strength; -6%
sodium hypochlorite), bleach (dilute, -0.6% sodium hypochlorite), hydrogen peroxide (3%), and
DF200. A 30 min reaction time was evaluated for all decontaminants and material combinations.
An additional reaction time (60 min) and a 30 min reaction time with a subsequent reapplication
of the decontaminant and additional 30 min reapplication) were evaluated for some combinations
of decontaminants and materials. Results are summarized in Tables 25, 26, and 27.
With a 30 min reaction time, high efficacy was observed for all four decontaminants again the L
(derL-1) component of Agent Yellow on all four material types. Efficacies were generally lower
for the HD component than for the derL-1 component with full strength bleach and hydrogen
peroxide (3%) more efficacious than diluted bleach or DF200. Efficacies varied by material type
as well as decontaminant.
An additional test applying bleach (dilute) or hydrogen peroxide (3%) with a 30 min reaction
time against HD only on wood and glass was performed. Dilute bleach demonstrated efficacy of
49% on wood and 30% on glass. Hydrogen peroxide (3%) demonstrated low levels (but
statistically significant) of efficacy (21% on wood and 10% on glass).
Table 25. Summary of Average % Decontamination Efficacy with a 30 min Reaction Time
'contamirv
Agent
Yellow
Material
Strength)
(Dilute)
Hydrog
Peroxide (3%)
derL-1
Sealed Concrete
Wood Flooring
Galvanized Metal
Glass
>92%
*t
>79%
>94%
>87%
>66%
>94%
86%
>86%
>83%
>92%
>93%
>92%
>86%
>87%
>95%
Decontaminant
Agent
Yellow
HD
Material
Sealed Concrete
Wood Flooring
Galvanized Metal
Glass
Bleach (Full Bleach
Strength) (Dilute)
>94%
37%
>37%
>95%
*
*
*
35%
Hydrogen
Peroxide (3%)
*
80%
49%
42%
23%
*
*
57%
Efficacies shown as ">" had at least one and in most cases all coupon extracts that were below the limit of detection.
* No significant difference (p>0.05) between the mean agent remaining on the positive control coupons and the mean agent
remaining on the test coupons, no significant efficacy is observed.
There was high variability in the positive controls so the p = 0.007; efficacy (although not significant) was 85%.
In the 60 min reaction time testing, derL-1 was below the quantitation limits for all extracts
decontaminated with hydrogen peroxide (3%) and DF 200 as well as concrete extract after
44
-------
decontamination with bleach (dilute). Natural attenuation of agent from positive control coupons
occurs during the 60 min period during which the decontamination process is occurring on the
test coupons. Thus, in spite of these non-detected recoveries, artificially lower efficiencies were
observed after 60 min compared to 30 min. The lower efficacies compared to 30 min reaction
times reflect bias arising from the declining derL-1 recoveries from the positive control coupons.
The lower recoveries from positive control coupons are likely due to the longer period of
evaporation. Because of this bias, the reduction in recovered Agent Yellow as a result of the
decontamination process (including both the decontaminant and natural attenuation) is likely
more informative than the efficacy value for longer reaction times. No operational significance
should be derived from these artificially declining efficacy values. (Because no derL-1 or HD
had been recovered after 30 min reaction time with bleach (full strength), testing with concrete
and glass at 60 min was not evaluated.)
Table 26. Summary of Average % Decontamination Efficacy with a 60 min Reaction Time
Efficacv for derL-1
Material Bleach (Full Strength)
Bleach (Dilute)
Hydrogen
Peroxide (3%)
DF200
Concrete
Wood
Metal
Glass
Not tested
45%
88%
Not tested
>86%
>72%
>76%
72%
>89%
| >79%
>92%
>92%
>84%
>80%
>81%
>94%
Material Bleach (Full Strength)
Efficacv for HD
Bleach (Dilute)
Hydrogen
Peroxide (3%)
DF200
Concrete
Wood
Metal
Glass
Not tested
*
86% |
Not tested
*
*
*
62%
63/o
>55%
71%
70%
*
*
*
64%
Efficacies shown as ">" had at least one and in most cases all coupon extracts that were below the limit of detection. Green color
indicates % efficacy was greater with a 60 min reaction time compared to the 30 min reaction time.
* No significant difference (p>0.05) between the mean agent remaining on the positive control coupons and the mean agent
remaining on the test coupons, no significant efficacy is observed.
The 30 min reaction time followed by reapplication for 30 min was only evaluated for two
decontaminants, bleach (dilute) and hydrogen peroxide (3%). Lower amounts of HD were
recovered from all coupon types after reapplication of bleach (dilute) or hydrogen peroxide (3%)
compared to the single 30 min application. HD recovery results after reapplication of bleach
(dilute) or hydrogen peroxide (3%) were similar to the corresponding 60 min reaction times for
these decontaminants. In spite of low recoveries from metal after decontamination (shown in
Figure 12), efficacy was not observed because of the unexpectedly low recoveries from the
45
-------
positive control coupons, i.e., there was little difference between the recoveries from positive
control or test metal coupons. Rust coloring on the metal where the Agent Yellow was applied
suggests corrosion that may be related to the low recoveries from the positive control coupons.
450
400
350
300
250
200
150
100
50
I Positive Control Coupons
I Test Coupons
30 60 30+30
Bleach (dilute)
30 60 30+30
Hydrogen peroxide (3%)
Figure 12. Comparison of the mass of HD recovered from metal positive control coupons
and test coupons after a 30 min reaction time (30), 60 min reaction time (60), and 30 min
reaction time with reapplication for additional 30 min (30+30) of bleach (dilute) or
hydrogen peroxide (3%).
Qualitative analysis for vesicant by-products for L [L-2 (bis[2-chlorovinyl] chloroarsine and
bis(2-chlorovinyl) arsinous acid (both derivatized to derL-2 prior to analysis)] and the vesicant
by-product of HD, bis(beta-chloroethyl)sulfone (BCVAA) showed that there were generally no
chromatographic peaks observed in positive control coupon extracts that correspond to these
chemicals. Small chromatographic peaks consistent with vesicant by-products were detected on
some materials after application of each decontaminant. Peak areas correlated with derL-2 were
found on all coupon types after decontamination with dilute bleach for both 30 and 60 min
reaction times and ranged from 1% to 24% of the corresponding derL-1 peak area. BCVAA was
found on all coupon types after hydrogen peroxide (3%) decontamination with 30 and 60 min
reaction times.
46
-------
Table 27. Summary of Average % Decontamination Efficacy with a 30 min Reaction Time
with Reapplication and Subsequent additional 30 min Reaction Time
Efficacy for derL-1
Material
Concrete
Wood
Metal
Glass
Bleach (Dilute)
>94%
>83%
>83%
>91%
Hydrogen Peroxide (3%)
>95%
>89%
>81%
>90%
EfficacvforHD
Material
Concrete
Wood
Metal
Glass
Bleach (Dilute)
62%
27%
-176%*
54%
Hydrogen Peroxide (3%)
61%
>91%
-2.5%* (not significant)
47%
Efficacies shown as ">" had at least one and in most cases all coupon extracts that were below the quantitation limit, <20
(ig/coupon.
* HD recovered from positive control coupons was unexpectedly very low.
In summary, all four methods of decontamination that were tested (bleach [full strength], bleach
[dilute], hydrogen peroxide [3%], and DF200 are highly efficacious against the L component of
Agent Yellow and exhibit varying levels of efficacy, depending on material and decontaminant,
against the HD component after 30 min reaction times. Bleach (full strength) generally removed
L below the levels of detection and exhibited efficacies for HD of 37% to >95%. The efficacy
ranges for bleach (dilute), hydrogen peroxide (3%), and DF200 were comparable to bleach (full
strength) for L, but had lower ranges of efficacy for HD.
Application of decontamination solutions on surfaces as described in this report is one approach
than would be part of a remediation strategy. Volumetric decontamination is another approach
for larger areas with lower contamination levels. Further research efforts would be required to
determine its functionality towards neutralization of these vesicant chemical agents.
47
-------
6.0 References
1. Organization for the Prohibition of Chemical Weapons. (2005) Convention on the
Prohibition of the Development, Production, Stockpiling, and Use of Chemical Weapons, p.
52. http://www.opcw.org/index.php?eID=dam_frontend_push&docID=6357. Accessed
7/23/13.
2. Goldman M. and Dacre, J.C. Lewisite: its chemistry, toxicology, and biological effects.
Reviews of Environmental Contamination and Toxicology 110 (1989): 75-115.
3. Muir B., Slater B.J., Cooper D.B., Timperley C.M. Analysis of chemical warfare agents. I.
Use of aliphatic thiols in the trace level determination of Lewisite compounds in complex
matrices. Journal of Chromatography A 1028 (2004): 313-320.
4. Hanaoka S., Nomura K., and Wada T. Determination of mustard and lewisite related
compounds in abandoned chemical weapons (Yellow shells) from sources in China and
Japan. Journal of Chromatography A 1101 (2006): 268-277.
5. Kroening K.K., Easter R.N., Richardson D.D., Willison S.A., and Caruso J.A. Analysis of
Chemical Warfare Degradation Products. West Sussex, UK: John Wiley & Sons (2011): 36-
42.
6. Munro N., Talmage S.S., Griffin G.D., Waters L.C., Watson A.P., King J.F., and Hauschild
V. The sources, fate, and toxicity of chemical warfare agent degradation products.
Environmental Health Perspectives 107 (1999): 933-974.
7. Ellison, Hank D. Handbook of Chemical and Biological Warfare Agents, 2n Edition. New
York: CRC Press (2008): p. 184.
8. EPA Report. Decontamination of Lewisite using Liquid Solutions: Neutralization and
Arsenic Removal (2014), EPA 600/R-14/119.
9. EPA Report. Evaluation of Household or Industrial Cleaning Products for Remediation of
Chemical Agents (2011), EPA 600/R-l 1/055.
48
-------
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
------- |