June 2lri4- | www.epa.gov/or
United States
Environmental Protection
Agency
Assessment and Evaluation Report
Decontamination of
Lewisite using Liquid Solutions:
Neutralization and Arsenic
Removal
Office of Research and Development
National Homeland Security Research Center
-------
Decontamination of Lewisite using Liquid
Solutions: Neutralization and Arsenic
Removal
Assessment and 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-11-038, Task Order 0007 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)
109T.W. Alexander Dr
Research Triangle Park, NC 27711
Phone: 919-541-2973
Fax: 919-541-0496
E-mail: Oudeians.Lukas@epa.gov
m
-------
Acknowledgments
The following individuals and organizations are acknowledged for review of this document:
United States Environmental Protection Agency:
Office of Solid Waste and Emergency Response, Office of Emergency Management
Larry Kaelin
Office of Research and Development, National Risk Management Research Laboratory
Dennis Tabor
Office of Research and Development, National Homeland Security Research Center
Stuart Willison
Emily Snyder
Ramona Sherman (QA review)
Contributions of the following organization are acknowledged:
Battelle
IV
-------
Table of Contents
Disclaimer iii
Acknowledgments iv
Table of Contents v
List of Figures vii
List of Tables vii
Acronyms and Abbreviations ix
Executive Summary xi
l.OIntroduction 1
1.1.Purpose 2
1.2.Project Objectives 2
1.3 .Test Facility Description 3
2.0Experimental Methods 4
2.1 .Chemical Agent and Spiking Coupons 4
2.2.Test Materials 5
2.3 .Description and Application of Lewisite Decontaminants 6
2.4. Description and Application of Arsenic Removal Technologies 6
2.5 .Extraction of Coupons 7
2.6.Derivatization of Subset of Extracts in the Test Matrix 7
2.6.1 Derivatized Lewisite 7
2.7 .Analyzing for Lewisite and Degradation Products 7
2.7.1 GC/MS with Cool On-Column Injection System 8
2.7.2 GC/MS with Derivatization of Lewisite 9
2.7.3 LC/MS for CVAOA analysis 10
2.8.Method Development and Demonstration 12
2.8.1 Extraction Efficiency for L-l and derL-1 12
2.8.2 Method Detection Limit for Derivatized Lewisite 13
2.8.3 Extraction Efficiencies for Arsenic 13
2.8.4 Method Detection Limit for Arsenic 14
2.9.Test Matrices 15
2.10 Observation of Surface Damage 16
2.11 Extraction Efficiency 18
S.OTest Results 20
-------
3.1.Method Development and Demonstration Results 20
3.1.1 Extraction Efficiency for L-1 Measured with GC/MS using a Cool On-Column Inlet.. 20
3.1.2 Extraction Efficiency and MDL for derL-1 Analyzed with GC/MS 21
3.1.3 Extraction Efficiency and MDL for Arsenic 21
3.2.Decontamination Results 23
3.2.1 Solution Test 23
3.2.2 Efficiency Results Using Water to Decontaminate Lewisite on Glass or Wood 24
3.2.3 Efficiency Results Using Hydrogen Peroxide to Decontaminate Lewisite on Glass or
Wood 26
3.2.4 Efficiency Results Using Bleach to Decontaminate Lewisite on Glass or Wood 27
3.2.5 Efficiency Results Using DF200 to Decontaminate Lewisite on Glass or Wood 28
3.3 .Arsenic Removal Results 29
3.4. Observations of Damage to Coupons 30
4.0Quality Assurance/Quality Control 33
4.1 .Control of Monitoring and Measuring Devices 33
4.2.Equipment Calibrations 34
4.3 .Technical Systems Audit 36
4.4.Performance Evaluation Audits 36
4.5 .Data Quality Audit 37
4.6 .Amendments 37
4.7. Deviations 37
S.OSummary 39
6.0References 46
VI
-------
List of Figures
Figure 1. Lewisite and common degradation products for L-l and L-2 2
Figure 2. Details of the on-column inlet system 9
Figure 3. Te sting sequence for Lewisite and arsenic removal efficacy evaluation 15
Figure 4. Results for extraction efficiencies and %RSD (n = 3) 20
Figure 5. Photographs of coupons before and after decontamination treatment 32
List of Tables
Table ES-1.Summary of L-l and DerL-1 Efficacy Results xii
Table ES-2.Summary of L-2 and DerL-2 Efficacy Results xiii
Table 1. Test Materials, Descriptions, Sources, Size, and Preparation 5
Table 2. GC/MS Conditions for Lewisite Analysis 8
Table 3. Ions Monitored for Target Chemicals Using GC/MS for Lewisite 9
Table 4. GC/MS Conditions for Derivatized Lewisite Analysis 10
Table 5. Target Ions Monitored in GC/MS Analysis of Derivatized Lewisite 10
Table 6. LC/MS Conditions 11
Table 7. Pertinent Parameters for CVAOA Using LC/MS 11
Table 8. L-l Extraction Efficiency Matrix 13
Table 9. Evaluation of Total Arsenic Extraction Efficiency 14
Table 10. Test Matrix for Lewisite Decontamination 17
Table 11. Hexane Extraction Efficiencies and Method Detection Limits for derL-1 21
Table 12. Extraction Efficiencies (% Arsenic Applied that was Recovered) from Test Coupons and
Solution Controls 22
Table 13. Background Levels of Arsenic Detected in Coupon Materials 22
Table 14. Determination of Arsenic Method Detection Limits Using the Certifier 6 Method 23
Table 15. Solution Decontamination Test (15 min Reaction Time) 24
Table 16. Water Decontamination Efficacy (measuring L-l) 24
Table 17. Water Decontamination Efficacy (measuring derL-1) 25
Table 18. Hydrogen Peroxide Decontamination Efficacy Measured as L-l 26
Table 19. Hydrogen Peroxide Decontamination Efficacy (measuring derL-1) 27
Table 20. Bleach Decontamination Efficacy 28
Table 21. DF200 Decontamination Efficacy 29
Table 22. Arsenic Removal Efficiencies Using Water and LeadOff 30
vii
-------
Table 23. Data Quality Objectives and Results for the Test Measurements 33
Table 24. Equipment Calibration Schedule 34
Table 25. TSA Results 36
Table 26. PE Results 36
Table 27. Efficacy (% reduction) of Decontaminants Analyzed as L-l or derL-1 40
Table 28. Mean Efficacy of Decontamination Methods Analyzed as DerL-1 41
Table 29. Summary of Efficacy Results for L-l and DerL-1 42
Table 30. Summary of Efficacy Results for L-2 and DerL-2 43
Table 31. Results of Qualitative Analysis for CVAOA 44
Table 32. Summary of Arsenic Removal Efficiency Using Gauze Wetted with Water or LeadOff 45
Vlll
-------
Acronyms and Abbreviations
amu
BBRC
BCVAA
°C
ccv
COR
CVAA
CVAO
CVAOA
cm
EPA
derL-1
derL-2
DF200
DI
g
GC
GFAA
HPLC
kHz
L-l
L-2
L-3
LC
m
MDL
mg
min
mL
mm
MS
ng
NHSRC
atomic mass units
Battelle Biomedical Research Center
bis(2-chlorovinyl) arsonous acid
degrees Celsius
continuing calibration verification
Contracting Officer Representative
2-chlorovinyl arsonous acid
2-chlorovinyl arsine oxide (Lewisite oxide)
2-chlorovinyl arsonic 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
deionized
gram(s)
gas chromatography
graphite furnace atomic absorption
high-performance liquid chromatography
kilohertz
2-chlorovinyl dichloroarsine, Lewisite
bis(2-chlorovinyl) chloroarsine
tris(2-chlorovinyl) arsine
liquid chromatography
meter(s)
microgram(s)
micrometer(s)
microliter(s)
method detection limit
milligram(s)
minute(s)
milliliter(s)
millimeter(s)
mass spectrometry
nanogram(s)
National Homeland Security Research Center
ix
-------
NIST National Institute of Standards and Technology
NRT National Response Team
PE performance evaluation
ppm part(s) per million
QA quality assurance
QC quality control
RSD relative standard deviation
SAM Selected Analytical Methods
SIM selected ion monitoring
TSA technical systems audit
-------
Executive Summary
Lewisite is an arsenical, vesicant, chemical warfare agent. During construction of the National
Response Team Quick Reference Guides, U.S. Environmental Protection Agency (EPA)
scientists discovered a lack of data on the decontamination of Lewisite-contaminated surfaces.
The objective of this evaluation was to determine the neutralization efficacies of various
Lewisite decontamination methods. Lewisite comprises three organo-arsenic vesicants: L-l (2-
chlorovinyldichloroarsine, Lewisite), L-2 (bis-[2-chlorovinyl]-chloroarsine), and L-3 (tris-[2-
chlorovinylj-arsine). L-l, L-2, and L-3 typically constitute approximately 90 %, 9 %, and 1 %,
respectively, of Lewisite. It is known that in the presence of water, L-l rapidly hydrolyzes to 2-
chlorovinyl arsonous acid (CVAA) which is a vesicant. Because decontamination of Lewisite
generates arsenical compounds, residual risk may be associated with the decontamination
products. Therefore, an additional objective was to evaluate the amount of residual arsenic
remaining on building materials after decontamination and wiping with gauze wetted with water
or the commercial lead removal product Hygenall® LeadOff Surface Decontamination Spray
Cleaner (LeadOff).
Decontamination efficacy was evaluated for four decontaminants: water, hydrogen peroxide (3
%), bleach (8.7 % hypochlorite), and EasyDecon DF200 (DF200). Results are summarized in
Table ES-1. In the presence of water, a significant decrease in the amount of L-l occurred that
may be attributed to conversion to CVAA. Derivatization during analysis that converts L-l and
CVAA to a common product, derL-1, was used. Significant amounts of derL-1 (i.e. L-l or
CVAA) were recovered relative to the L-l, indicating most of the L-l had been converted to
CVAA. DerL-1 (i.e., L-l or CVAA) remained on glass and wood- even after a 60 minute (min)
contact period with water. Water exhibited the lowest efficacy of the four methods tested at 30
min and no additional efficacy was observed with a longer (60 min) reaction time.
With hydrogen peroxide applied to either glass or wood, neither L-l nor CVAA (i.e., no derL-1)
was detected after a 30- or 60- min reaction time. A small amount of derL-1 was detected in the
derivatized extract from wood (but not on glass) after a 30- min reaction time with both bleach
and DF200. After a 60- min reaction time with bleach or DF200, no derL-1 (i.e., no L-l or
CVAA) was detected on wood. Hydrogen peroxide, bleach, and DF200 all showed significant
efficacy against Lewisite (measured as derL-1).
Table ES-2 summarizes the qualitative results from decontamination of L-2. While analysis
showed that the amounts of L-l recovered from test coupons decontaminated with water were
lower compared to the amounts recovered from positive controls, the relative amount of L-2
(qualitative) recovered from the test coupons was not reduced by the 30- min reaction time with
water. The L-2 remained on the test coupons in the presence of water, while most of the L-l was
no longer detected. DerL-2 (the common product when derivatizing both L-2 and bis[2-
chlorovinyl] arsonous acid) was detected in the derivatized extracts from all glass and some
wood coupons after a 60- min contact with water, but was not detected after decontamination
with hydrogen peroxide, bleach, or DF200 for 30- or 60- min.
XI
-------
Table ES-l.Summary of L-l and DerL-1 Efficacy Results.
Form of Agent Analyzed and Decontaminant
Efficacy on Building Materials
Reaction Time
30 min
60 min
L-l conversion by water
Glass
Wooda
Not tested
L-l conversion by hydrogen peroxide (3 %)
DerL-1 conversion by water
Not tested
Glass
Glass
Wood
Wood
DerL-1 conversion by hydrogen peroxide (3 %)
Wood
DerL-1 conversion by bleach (8.7 % hypochlorite)
Not tested
DerL-1 conversion by DF200
Wood
a: Insufficient amount of L-l recovered from positive controls after 30 min to assess efficacy.
Key:
Efficacy less than 87 % for agent in specified form, e.g., L-l or derL-1.
Agent detected on some of the test coupons in specified form with efficacy greater than 87 %.
No agent detected in specified form and efficacy greater than 94 %.
XII
-------
Table ES-2.Summary of L-2 and DerL-2 Efficacy Results.
Form of Agent Analyzed and Decontaminant
Reaction Time
L-2 conversion by water
L-2 conversion by hydrogen peroxide (3 %)
DerL-2 conversion by water
DerL-2 conversion by hydrogen peroxide (3 %)
DerL-2 conversion by bleach (8.7 % hypochlorite)
DerL-2 conversion by DF200
Detection on Building Materials
30 min
Glass
Glass
60 min
Not Tested
Not Tested
Not Tested
Not Tested
Glass
Wood (Detected on Two
of Five Coupons)
Not Tested
Not Tested
Key:
Detected in specified form, e.g., L-2 or derL-2.
No agent detected in specified form.
Qualitative analysis was performed to detect a potential oxidation product of CVAA, 2-
chlorovinyl arsonic acid (CVAOA). After decontamination by hydrogen peroxide for 30 or 60
minutes, detectable amounts of CVAOA were observed, except that after 60- min contact with
water, CVAOA was no longer detected on glass or wood. The dynamics of CVAOA formation
and degradation were not obvious from these results.
Removal of residual arsenic from coupons after decontamination with water and hydrogen
peroxide was evaluated by wiping the coupons with a wetted gauze pad after spraying with either
water or LeadOff followed by analysis of residual arsenic on the coupons. Wiping with gauze
after spraying with either water or LeadOff was efficacious in removing arsenic from glass.
LeadOff removal efficiencies from glass were 93 %-98 %, which were slightly better than using
water (84 % -92 %). Removal of arsenic from wood by wiping with wetted gauze after spraying
with water or LeadOff was ineffective. The arsenic was assumed to have soaked into the wood
where it was not readily removed by wiping.
Based on results obtained in this study, the vesicant properties can be neutralized by using
bleach, hydrogen peroxide, or hydrogen peroxide containing products such as DF200. Water
only converts the main L-l component of Lewisite into a different chemical with significant
Xlll
-------
vesicant properties and is therefore not recommended. Caution should be used in extrapolating
from bench testing to field application of these decontamination solutions.
XIV
-------
1.0 Introduction
Protecting human health and the environment from the release of hazardous materials is the
mission of the U.S. Environmental Protection Agency (EPA). During construction of the
National Response Team (NRT) Quick Reference Guides, EPA's National Homeland Security
Research Center (NHSRC) scientists discovered there was a lack of data on the persistence of
Lewisite and on decontamination of Lewisite-contaminated surfaces. (For NRT Guides, see:
NRT Quick Reference Guides for Chemicals). Figure 1 highlights Lewisite constituents and
potential decontamination products that are discussed in this report. As manufactured, Lewisite
(L) comprises three compounds: L-l (2-chlorovinyldichloroarsine, Lewisite); L-2 (bis[2-
chlorovinyl] chloroarsine); and L-3 (tris[2-chlorovinyl] arsine).1 Lewisite, including L-l, L-2,
and L-3, is a Schedule 1 organo-arsenic vesicant under the Chemical Weapons Convention.2 In
the presence of water, Lewisite rapidly hydrolyzes to 2-chlorovinyl arsonous acid (CVAA) and,
with excess water, is rapidly converted to 2-chlorovinyl arsine oxide (Lewisite oxide or
CVAO).3'4 L-l, CVAA, and CVAO are vesicants.4 Lewisite has only low solubility in water and
is more volatile than CVAA, which is water soluble,4 thus complicating extraction and analysis
of Lewisite and degradation products. Further, oxidation of CVAA during decontamination may
generate 2-chlorovinyl arsonic acid (CVAOA).4 No single method is known for analysis of the
three constituents of Lewisite and the hydrolysis and oxidation products.
While Lewisite released into the environment may not persist, various inorganic arsenic
decontamination products may persist.4 Arsenic, as an element, is expected to remain in
degradation products that arise naturally or by applying decontaminants. The residual arsenic
may be toxic and needs to be removed to release a contaminated site for reoccupation and use.
Efficacious Lewisite decontaminants would be expected to neutralize the three different
compounds that comprise Lewisite (L-l, L-2, and L-3), converting them to non-vesicant
compounds. Observing efficacy against Lewisite (L-l, L-2, and L-3) is not sufficient to ensure
that products with vesicant properties have not been generated. Additional or alternative testing
was, therefore, used to detect such products. In addition, the decontamination strategy must also
consider how to remove the remaining toxic arsenical compounds after Lewisite degradation and
remove arsenic in totality if materials are to be reused.
-------
Lewisite (L1) j
C2H2AsCI3 |
Hydrolyze,
+H2O
CVAA
CICH=CHAs(OH)2
Oxidize,
+H2O2
CVAOA
CICH=CHAs=O(OH)2
Lewisite 3 (L3)
C6H6AsCI3
Dehydrate,
Lewisite 2 (L2)
C4H2AsCI3
Hydrolyze,
+H20
BCVAA
(CICH=CH)2AsOH
Dehydrate,
i [(CICH=CH)2As]0
Figure 1. Lewisite and common degradation products for L-l and L-2.
1.1 Purpose
The overall purpose of this evaluation was to determine the neutralization efficacies of various
readily-available, liquid-based methods for Lewisite decontamination (including arsenic
removal) from surfaces.
A review of the literature, as part of this effort, showed that methods for evaluating residual
Lewisite and decontamination products were limited. Therefore, prior to performing
decontamination and removal experiments, method development for the analysis of Lewisite and
degradation products was required. The method development and subsequent evaluation of
decontamination and arsenic removal technologies are presented here.
1.2 Project Objectives
Specific project objectives to achieve the overall purpose included:
• systematic evaluation of the efficacy of four surface decontaminants for neutralization of
Lewisite (conversion to a non-vesicant compound) on two surfaces (glass and wood), and
• evaluation of the efficacy of subsequent arsenic removal by scrubbing the surface with
water and with Hygenall® LeadOff Surface Decontamination Spray Cleaner (LeadOff).
-------
1.3 Test Facility Description
All testing was performed at the Battelle Biomedical Research Center (BBRC) located on the
Battelle site in West Jefferson, Ohio. Battelle is certified to work with chemical surety materials
through its contract with the Defense Threat Reduction Agency (contract number: W81XWH-
ll-D-0002).
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 at least once during each day of testing. The
temperature in the laboratory during testing ranged from 19.0 °C - 20.7 °C and the relative
humidity ranged from 45 % to 62 %.
-------
2.0 Experimental Methods
Decontamination testing was conducted on a bench-scale to evaluate the efficacy of four
decontamination methods against surfaces contaminated with Lewisite. Four types of building
materials were included in method development that included determination of extraction
efficiencies of alternative methods. Four decontaminants were evaluated for efficacy against
Lewisite (L-l and L-2) applied to two materials. Not all combinations of building materials and
decontamination methods were tested as a result of initial testing results.
A post-test-only control group experimental design was used for the determination of the
decontamination efficacy against L-l and L-2. Test coupons were decontaminated (experimental
variable) and then extracted and analyzed for L-l and L-2 (Observation 1; O-i). Positive control
coupons were not decontaminated but were extracted and analyzed for L-l and L-2 (Observation
2; 02) along with the test coupons. The effect of the treatment (efficacy) was reported as the
percentage of L-l or L-2 remaining on treated coupons compared to the control coupons:
Efficacy = [(O2 - d)/ O2]«100% (1)
The higher the efficacy, the greater the effect of the decontamination.
In addition to the test and control coupons, laboratory blank coupons (coupons that were neither
contaminated with Lewisite nor decontaminated) and procedural blank coupons (coupons that
were not spiked with Lewisite, but decontaminated along with the test coupons) were extracted
and analyzed for L-l and L-2. To verify the amount of Lewisite spiked onto coupons, the same
amount as applied to coupons was directly pipetted into the extraction solvent and analyzed as a
spike control.
Likewise, a post-test-only control group experimental design was used for the determination of
the removal of arsenic subsequent to the application of two Lewisite decontaminants. Coupons
were spiked with Lewisite, decontaminated, and randomly assigned as test coupons or positive
control coupons for the arsenic removal test. Arsenic removal test coupons were wiped with
water or with Leadoff (experimental variable) and then extracted and analyzed for arsenic
(Observation 1; O-i). Arsenic removal positive control coupons were not wiped with water or
Leadoff but were extracted and analyzed for arsenic (Observation 2; ©2) The effect of the
treatment (efficacy) was reported as the percentage of arsenic remaining on treated coupons
compared to the control coupons and was calculated using Equation 1. The higher the efficacy,
the greater the arsenic removal.
2.1 Chemical Agent and Spiking Coupons
The neat Lewisite that was used was supplied by the U.S. Army and owned by the EPA. Because
no standards exist for Lewisite, relative composition of L-l (CAS 541-25-3), L-2 (CAS 40334-
69-8), and L-3 (CAS 40334-70-1) in the agent was determined using gas chromatography/mass
spectrometry (GC/MS). The Lewisite used in this evaluation was determined to be 92 % L-l and
3 % L-2; no L-3 was detected. The impurities constituting the remainder of the area under the
chromatographic peaks were not identified.
-------
All test and positive control coupons were spiked with 1 microliter (jiL; nominally 1.88
milligrams [mg]) of neat Lewi site onto a coupon surface of about 5.25 square centimeters (cm2).
This area represented a contamination level of approximately 4 grams (g)/meter2 (m2). Each
microliter contained approximately 0.68 mg of arsenic. A positive displacement pipette (P/N M-
10 [1-10 jiL] and CP10 tip, Gilson Inc, Middleton, WI) was used to apply the Lewisite to the test
and positive control coupons.
2.2 Test Materials
Four types of building materials were included in method development that included
determination of extraction efficiencies of alternative methods. These materials included sealed
concrete, wood flooring, galvanized metal, and glass (Table 1). Except for sealed concrete,
coupons were cut from larger pieces of material to 3.5 centimeters (cm) x 1.5 cm. Concrete
coupons were poured into a mold and coated with sealer (Sure Klean® Weather Seal Siloxane
PD). Two materials, glass and wood, were selected by the EPA for subsequent use in the
evaluation of the decontaminants based on the ability to extract sufficient Lewisite from these
surfaces.
Table 1. Test Materials, Descriptions, Sources, Size, and Preparation
Material
Sealed
concrete
Wood
flooring
material
Galvanized
metal
ductwork
Glass
Description
Epoxy-sealed concrete
(5 parts sand; 2 parts
concrete); custom
preparation
Fir plywood (bare);
thickness 0.9 cm
Industry heating, ventilation,
and air conditioning
standard; 24 gauge
galvanized steel;
thickness 0.7 millimeters
(mm) (Adept
Manufacturing)
Glass (clear window)
Manufacturer/
Supplier Name
Wysong
Concrete
Cincinnati, OH
Lowe's,
Columbus, OH
Adept Products,
Tnr
West Jefferson,
OH
Brooks Brothers,
West Jefferson,
OH
Coupon
Surface Size
Length (cm) x
Width (cm)
35x15
3.5 x 1.5
3.5 x 1.5
3.5 x 1.5
Material Preparation
Clean with dry air to
remove loose dust
Clean with dry air to
remove loose dust
Clean with acetone
Clean with dry air to
remove loose dust
-------
2.3 Description and Application of Lewisite Decontaminants
Four decontaminants were evaluated for efficacy against Lewisite (L-l and L-2) as applied to the
materials. The first three of four decontamination solutions are, in general, readily available from
local retail stores:
• Deionized (DI) water (#23-751-610, Fisher Scientific)
• Bleach (sodium hypochlorite 5-10 %, Clorox® Regular Concentrated bleach (#003-07-
0755, Target)
• Hydrogen peroxide (3 %, #3819132, Fisher Scientific)
• EasyDECON DF 200 (DF200, EFT Holdings, Inc.) applied as a liquid.
The decontaminants were applied as a liquid to the test coupons 30 min after the Lewisite was
spiked onto the coupons. The initial reaction time for the decontaminants was 30 min. The
decontamination testing was repeated at a second reaction time (60 min) for selected
combinations of coupons and decontaminant to determine whether extended interaction times of
the decontaminant with Lewisite would enhance the decontamination efficacy. Each
decontaminant was applied as a single droplet (using a positive displacement pipette (P/N M-100
[100 uL] and D-200 [2-200 uL] tip, Gilson Inc., Middleton, WI). Decontamination volumes
ranged from 60 to 90 jiL for decontamination of glass and wood, respectively.
2.4 Description and Application of Arsenic Removal Technologies
Two technologies were evaluated for efficacy in removing residual arsenic on glass and wood
after decontamination with two of the four decontaminants, namely water or hydrogen peroxide:
• Deionized water (#23-751-610, Fisher Scientific), and
• LeadOff (#HN21131QCS, The LeadOff Store, NewingtonNH 03801).
The LeadOff manufacturer's instructions are: (1) apply generously; (2) allow to sit 5-10 seconds;
(3) wipe with a damp cloth. Based on these instructions, the following approach was used for
both deionized water and the LeadOff cleaner:
1. Apply generously with mist setting of spray bottle until entire coupon surface appears
visually "wet".
2. Allow treated coupon to sit for -5-10 seconds.
3. Wet a 5 cm x 5 cm gauze pad (#22-362-178, Fisher) with 2 mL DI water and fold in half.
4. Grasp wetted gauze with disposable forceps and wipe the coupon once from one end to
the other using the folded end of the gauze pad.
5. Repeat Step 4 twice with a freshly wetted gauze pad for a total of three wipes.
6. Extract the coupon for arsenic using the Certifier 65 method.
-------
2.5 Extraction of Coupons
At the time of this study, the recommended methods for analysis of Lewisite using the EPA's
Selected Analytical Methods for Environmental Remediation and Recovery (SAM)6 database
only relate to measurement of total arsenic. These methods do not address the change in vesicant
properties of this agent during decontamination and were not considered here.
After the appropriate period of contact with the decontaminant, the test coupons were transferred
individually into separate 40 mL glass bottles (05-719-120, Fisher Scientific, Pittsburgh, PA)
that contained 10 mL of solvent that was added to the vial using a 0.10-110 mL bottletop
dispenser (unknown model number, Barnstead), then sonicating at 50-60 kilohertz (kHz) for 10
min. Solvent selection was based on the outcome of the extraction efficiency studies as part of
the method development. The same extraction process was repeated for all samples until each
test coupon, positive control coupon, and procedural blank coupon had been extracted and an
aliquot taken for analysis. Extraction removes L-l, L-2, and L-3 from the aqueous
decontaminating conditions thus halting further decontamination processes in the extract. After
the extraction was completed, a 1-mL sample was transferred to a GC vial (P/N HP-5181-8801,
VWR [Agilent Technologies], West Chester, PA) using a 1 mL positive displacement pipette
(P/N M1000 [100-1000 |iL] and CP1000 tip, Gilson Inc., Middleton, WI) and sealed. Samples
that were not analyzed the same day were stored at <-70 °C.
In a similar manner, 10 mL acetone was used to extract a replicate set of coupons for liquid
chromatography/MS (LC/MS) analysis for the degradation product CVAOA.
2.6 Derivatization of Subset of Extracts in the Test Matrix
Extracts were analyzed for L-l and L-2 by GC/MS or derivatized prior to detection by GC/MS
for detection of decontamination byproducts. The derivatization process of the extract is
described in this section.
2.6.1 Derivatized Lewisite
Based on the results from method development and demonstration, analysis of derivatized
Lewisite was used as the primary metric of decontamination efficacy. The derivative of L-l and
its hydrolysis byproduct, CVAA, yield the same product (derL-1). Similarly, the derivative of L-
2 and its hydrolysis byproduct, bis(2-chlorovinyl) arsonous acid (BCVAA), yield a common
product, derL-2, that can be distinguished from derL-1 using GC/MS.
The derivatization followed the method of Muir et al.7 Lewisite derivatives were formed by
adding 200 uL of 1 mg mL"1 butanethiol to a 1 mL aliquot of each Lewisite extract (no coupon
material present) to be analyzed. Triethylamine (50 micrograms [ug]) was added to the solution
to catalyze the derivatization. The solution was mixed on a vortex mixer for 10 seconds.
2.7 Analyzing for Lewisite and Degradation Products
Four analysis techniques were investigated but only three were evaluated to analyze the extracts
for Lewisite and degradation products: GC/MS with cool on-column injection system;
-------
derivatization of the extract followed by GC/MS; and LC/MS. These methods are described in
the following sections.
2.7.1 GC/MS with Cool On-Column Injection System
Method development was required for quantitative analysis of L-l and qualitative analysis of L-2
and L-3 by GC/MS. Replicate sets of samples were analyzed using an Agilent® 6890N Series GC
interfaced to a 5973 network quadrupole mass-selective detector (Palo Alto, CA).
Chromatographic separation of the analytes was conducted using a Restek Rtx-5 fused silica
capillary column (Bellefonte, PA), 30 m x 0.25 mm x 0.25 micrometers (|im). The GC/MS
conditions used for separation of L-l, L-2, and L-3 from co-extractives are outlined in Table 4.
A conventional application of GC/MS, using the same parameters as in Table 2 but using
splitless injection (rather than cool on-column injection) was evaluated for analysis of neat
Lewisite. L-l and L-2 were detected (no L-3 was observed), but with poor sensitivity (no data
shown; development funded by other Federal Agency). A revised method was therefore
evaluated using a cool on-column injection system, shown in Figure 2, and using the parameters
described in Table 2.
Table 2. GC/MS Conditions for Lewisite Analysis
Parameter
Instrument
Column
Injection Temperature
Injection Mode
Injection Volume
Oven Program
MS Transfer Line
Temperature
MS Source Temperature
Electron Multiplier Voltage
Mode
Description
Agilent Model 6890 Gas Chromatograph equipped
Selective Detector and a Model 7683 Injector with
with a 5973 Mass
Auto Sampler.
Restek Rtx-MS5 fused silica capillary column (30 m, 0.25 mm inside
diameter, 0.25 um film thickness)
Track oven temperature ± 3 degrees Celsius (°C)
Cool on-column
2(iL
40 °C Initial temperature (hold 2 min)
250 °C @ 25 °C/min (hold 0 min)
300 °C Post temperature (hold 0 min)
250 °C
230 °C
-2200 V
Selective ion Monitoring (SIM)
Table 3 outlines the ion masses that were used to quantitate L-l, and identify L-2 and L-3 using
the SIM mode.
-------
Table 3. Ions Monitored for Target Chemicals Using GC/MS for Lewisite
Analyte
L-l
L-2
L-3
SIM Ions, m/z
206, 208, 145
87, 145,51,210
136, 77, 145
No incompatible metal components
Avoids heat-induced degradation of
Lewisite
Column:
30m
0.25mm ID
0.25 Mm film
thickness
Figure 2. Details of the on-column inlet system.
2.7.2 GC/MS with Derivatization of Lewisite
Method development was required to demonstrate a GC/MS method for analyzing for derL-1
(the common product of derivatizing both L-l and CVAA) and derL-2 (the common product of
derivatizing both L-2 and bis[2-chlorovinyl] arsonous acid).
GC/MS detection was also demonstrated for analyzing derL-1 (the common product of
derivatizing L-l and CVAA) and derL-2 (the common product of derivatizing both L-2 and
bis[2-chlorovinyl] arsonous acid). The GC/MS conditions (oven temperature, oven temperature
program rate, injection port/detector temperatures, and column flow rates) used for separation of
the derivatized Lewisite products are outlined in Table 4. The method of analysis for derL-1 was
a procedure similar to that of Hanaoka et al.8
-------
The mass selective detector was operated in the full-scan mode for compounds ranging from 40
to 400 amu. Table 5 outlines the target ion masses that were used to identify derL-1 and derL-2.
Table 4. GC/MS Conditions for Derivatized Lewisite Analysis
Parameter
Instrument
Column
Injection Temperature
Injection Mode
Injection Volume
Oven Program
MS Transfer Line Temperature
MS Source Temperature
Electron Multiplier Voltage
Mode
Description
Agilent Model 6890 Gas Chromatograph equipped with a 5973 Mass
Selective Detector and a Model 7683 Injector with AutoSampler.
Restek Rtx-MS5 fused silica capillary column (30 m, 0.25 mm
diameter, 0.25 um film thickness)
inside
Track oven temperature ± 3 °C
Splitless
IjiL
40 °C Initial Temp (hold 2 min)
250 °C @ 15 °C/min (hold 0 min)
300 °C Post Temp (hold 0 min)
250 °C
230 °C
-1400V
Scan (40-400 atomic mass units [amu])
Table 5. Target Ions Monitored in GC/MS Analysis of Derivatized Lewisite
Analyte
L-l (derivatized)
L-2 (derivatized)
Target Ions, m/z
164,204,314
107, 164, 286
2.7.3 LC/MS for CVAOA analysis
Method development and demonstration included demonstrating an LC/MS method to analyze
for the degradation product CVAOA.
LC/MS was used to analyze CVAOA a Lewisite oxidation product that is not detectable by
GC/MS. Replicate sets of samples extracted in acetone were analyzed using high performance
liquid chromatography (HPLC) coupled to a tandem mass spectrometer (HPLC-MS/MS). A
Shimadzu 20 XR series HPLC (Columbia, MD) coupled to an AB SCIEX Triple Quad™ 5500
mass spectrometer with the TurboIonSpray® probe installed (Framingham, MA) was used for
CVAOA analysis. Analyst® software was used for data acquisition, instrument control, and data
analysis. The method demonstrated followed Battelle's existing Standard Operating Procedure,
10
-------
summarized in Table 6, to analyze for CVAOA. Table 7 shows the ion transitions that were used
to identify CVAOA.
A single CVAOA standard was included in the LC/MS analyses. This standard was not used to
create a set of calibration standards as this was beyond the scope of the intended semi-
quantitative analysis of CVAOA by LC/MS. The use of the single standard enabled the results to
be reported as greater than or less than the value of the standard.
Table 6. LC/MS Conditions
LC/MS
Mass Spectrometer
HPLC
Data Acquisition Software
Analytical Column
Guard Column
Column Temperature
Mobile Phase
Mobile Phase Gradient
Flow Rate
Injection Volume
Run Time
AB SCIEX Triple Quad™5500 with the Turbolon Spray® probe,
Positive Ion Mode
Shimadzu 20 XR Series
Analyst 1.5.1
Phenomenex Prodigy ODS-3 150 x 2 mm, 5 (im (Torrance, CA)
Phenomenex Security Guard CIS, 2.1 x 4 mm (Torrance, CA)
30 °C
A: 98:2 H2O/Acetonitrile
B: 0.2 % formic acid in 80:20 acetonitrile/isopropanol
Time (min)
0
1.0
13.0
15.0
19.0
21.0
25.0
%B
0
0
25
100
100
0
0
0.2 mL/min
50 (iL
25 min
Table 7. Pertinent Parameters for CVAOA Using LC/MS
Analyte
CVAOA
MS/MS transitions monitored
187.0> 109.0 (primary)
187.0 > 61.0, 187.0 > 91.0, 187.0 > 123.0,
187.0 > 169.0 (all confirmation)
11
-------
2.8 Method Development and Demonstration
Additional method development and demonstration included:
• Determining extraction efficiency of L-l and derL-1 from coupons of each of the four
material types using three alternative solvents;
• Determining method detection limits for derL-1 by GC/MS;
• Determining extraction efficiencies of arsenic using two alternative methods: modified
EPA Method 200.99 and Certifier 65;
• Determining method detection limits for arsenic by graphite furnace atomic absorption
(GFAA) spectroscopy.
2.8.1 Extraction Efficiency for L-l and derL-1
Three solvents (toluene, hexane, and acetone) were selected for extraction efficiency testing for
extracting L-l from the coupon materials used in this evaluation. Extraction of L-l with each
solvent was evaluated for all four building materials (sealed concrete, wood flooring material,
galvanized metal ductwork, and glass). The extraction efficiency matrix is shown in Table 8.
Neat Lewisite (1 uL) was spiked onto the test coupons, as described in Section 2.1, and
immediately extracted. The test coupons and laboratory blank coupons were extracted as
described in Section 2.5. The positive solution controls were prepared by directly injecting 1 uL
of neat Lewisite into the same type of vials containing solvent.
Replicate sets of samples were analyzed for L-l, L-2, and L-3 as described in Section 2.7.1. The
extraction efficiency was also measured for derL-1 as described in Section 2.7.2 for the selected
extraction solvent for L-l. Samples that were not analyzed the same day were stored at <-70 °C.
12
-------
Table 8. L-l Extraction Efficiency Matrix
Solvent
Toluene
Hexane
Acetone
Material
Sealed concrete
Wood flooring material
Galvanized metal ductwork
Glass
Sealed concrete
Wood flooring material
Galvanized metal ductwork
Glass
Sealed concrete
Wood flooring material
Galvanized metal ductwork
Glass
Total Coupons
Number of Test
Coupons
3
3
3
3
3
3
3
3
3
3
3
3
36
Number of
Solution Spike
Controls
1
1
1
Number of
Laboratory
Blank Coupons
1
1
1
1
1
1
1
1
1
1
1
1
12
2.8.2 Method Detection Limit for Derivatized Lewisite
A method detection limit (MDL) study was performed for derL-1 for all four materials. For
MDL determination, seven samples of each material were spiked with 20 |iL of a 4.2 mg/mL L-l
solution in hexane, allowed to sit undisturbed for approximately five minutes, and extracted as
described in Section 2.3.2. The hexane extract was derivatized as described in Section 2.5.1. The
mass of derL-1 in the extract was determined by GC/MS. The MDL was calculated following the
single concentration design estimator (40 CFR Part 136, Appendix B [1984]) as follows:
MDL = t(n-l,l-ct = 0.99) X SD
(2)
where:
t(n-l,l-a = 0.99) = the Student's t-value for a 99 % confidence level and standard
deviation estimate with n-1 degrees of freedom.
SD = standard deviation of the replicate measurements.
2.8.3 Extraction Efficiencies for Arsenic
Two extraction methods for total arsenic from the same coupon materials were demonstrated
during method development, as shown in Table 9. The method for analysis for total arsenic was
GFAA spectrometry.6 Coupons for determining extraction efficiency were spiked with 1 uL of
13
-------
arsenic reference standard solution (1000 parts per million [ppm] arsenic in 7 % nitric acid;
#SA449-500, Fisher Scientific). Extraction efficiency tests were also performed in the presence
of selected Lewisite decontamination methods (water and hydrogen peroxide) from glass and
wood to ensure that the decontamination method did not interfere with the arsenic extraction
analysis. The two extraction methods that were demonstrated were modified from EPA Method
200.99 (aqueous solutions of nitric acid and hydrochloric acid are added to the sample and
refluxed at approximately 95 °C) and modified from the method of Certifier 6 described in De La
Calle et al. (2010)5 (aqueous solutions of nitric acid are added to the sample and microwaved).
Analysis was performed following the Certifier 65 method.
Table 9. Evaluation of Total Arsenic Extraction Efficiency
Solvent
Nitric Acid +
Hydrochloric Acid6
Nitric Acid7
Material
Sealed concrete
Wood flooring
material
Galvanized metal
ductwork
Glass
Sealed concrete
Wood flooring
material
Galvanized metal
ductwork
Glass
Total Coupons
Test
Coupons
3
3
3
3
3
3
3
3
24
Number of
Solution Controls
1
1
Laboratory Blank
Coupons
1
1
1
1
1
1
1
1
8
2.8.4 Method Detection Limit for Arsenic
A MDL study was performed for total arsenic measurement for all four materials. To determine
the MDL, seven samples of each material were spiked with 1 jiL of the 1000 ppm arsenic
reference solution, allowed to sit undisturbed for approximately five minutes, and extracted and
analyzed as described in Section 2.8.3. The MDL was calculated according to Equation 2.
14
-------
2.9 Test Matrices
The overall testing sequence for Lewisite decontamination is diagrammed in Figure 3. The test
matrix is shown in Table 10. For each combination of time, material and decontamination
method, five test coupons (spiked with neat Lewisite, decontaminated for Lewisite), three
positive control coupons (spiked with neat Lewisite, not decontaminated for Lewisite) and two
procedural control coupons (not spiked with neat Lewisite, decontaminated for Lewisite) were
included. 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, for water and hydrogen peroxide 30 min reaction times, additional coupons were
included for subsequent analysis for residual arsenic. These included, for each combination of
decontaminant and materials, five test coupons (spiked with neat Lewisite, decontaminated for
Lewisite, and sprayed/wiped for arsenic removal), three positive control coupons (spiked with
neat Lewisite, decontaminated for Lewisite, but not sprayed/wiped for arsenic removal), and two
laboratory blank coupons (not spiked with neat Lewisite, decontaminated for Lewisite, and
sprayed/wiped for arsenic removal).
No controls that were not wiped were included in the quality assurance project plan. To remedy
this oversight, an additional set of coupons was prepared and decontaminated for Lewisite on a
different day than the day the arsenic test coupons were prepared. The additional set of coupons
included the positive controls for the Lewisite test. These coupons were not wiped prior to
extraction and analysis of the arsenic.
Extract with acetone:
glass and wood
Extract with hexane:
glass and wood
Glass and wood
coupons for As removal
LC/MS ana lysis
analysis forCVAOA
Derivatize; GC/MS
analysis for derL-1 and
derL-2
Cool on-column; GC/MS
analysis for L-l, L-2, and
L-3
Water spray and wipe
per As removal method;
analyze for As
LeadOff spray and wipe
per As removal method;
analyze for As
Figure 3. Testing sequence for Lewisite and arsenic removal efficacy evaluation.
15
-------
2.10 Observation of Surface Damage
The possible 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 and
photographs of pre- and post-decontamination coupons are included in Section 3.4. No visual
changes to the building materials were observed during testing.
16
-------
Table 10. Test Matrix for Lewisite Decontamination
Test
Day
1
2
Decontaminant
Water
Hydrogen
Peroxide
Bleach
DF200
Material
Wood
flooring
material
Glass
Wood
flooring
material
Glass
Wood
flooring
material
Glass
Wood
flooring
material
Glass
30 Min Reaction Time
5 test, 3 positive controls, 2 procedural
blanks
[Ten additional coupons included for use as
test, positive control, and laboratory blank
coupons in arsenic removal testing.]
5 test, 3 positive controls, 2 procedural
blanks
[Ten additional coupons included for use as
test, positive control, and laboratory blank
coupons in arsenic removal testing.]
5 test, 3 positive controls, 2 procedural
blanks
[Ten additional coupons included for use as
test, positive control, and laboratory blank
coupons in arsenic removal testing.]
5 test, 3 positive controls, 2 procedural
blanks
[Ten additional coupons included for use as
test, positive control, and laboratory blank
coupons in arsenic removal testing.]
5 test, 3 positive controls, 2 procedural
blanks
5 test, 3 positive controls, 2 procedural
blanks
5 test, 3 positive controls, 2 procedural
blanks
5 test, 3 positive controls, 2 procedural
blanks
60 Min Reaction
Time
5 test, 3 positive
controls, 2
procedural blanks
5 test, 3 positive
controls, 2
procedural blanks
5 test, 3 positive
controls, 2
procedural blanks
5 test, 3 positive
controls, 2
procedural blanks
5 test, 3 positive
controls, 2
procedural blanks
5 test, 3 positive
controls, 2
procedural blanks
5 test, 3 positive
controls, 2
procedural blanks
5 test, 3 positive
controls, 2
procedural blanks
17
-------
2.11 Extraction Efficiency
Extraction efficiency was calculated using a series of equations as follows. Chemical agent
concentration in a coupon extract or positive control solution sample was determined by
Equation 3:
As = a'Cs 2 + b'Cs + c (3)
where:
As = Area of the target analyte peak in the sample
Cs = Concentration (|j,g/mL) of the target analyte in the sample
a, b, c = Constants.
GC concentration results (ng/mL) are converted to total mass by multiplying by extract volume:
Mm=CxEv (4)
where:
Mm= measured mass of chemical agent (ng)
C = GC concentration (|j,g/mL), see Equation 3
Ev = volume of extract (mL).
Extraction efficiency is then defined as:
r- • T-/V- • I Mm of Chemical Agent on Coupon } ,nnn/ ,,-\
Extraction Efficiency = ^—^ x 100% (5)
[^Mm of Chemical Agent Spiked in Solvent j
where:
Mm= measured mass (ng)
Extraction efficiency = percent recovery of chemical agent from coupons.
The primary assessment of efficacy relies upon comparing the concentration of the target agent,
i.e., L-l or derL-1, on the test coupons, before and after the application of the decontaminant.
The purity of derL-1 and the ratio of derL-2 to derL-1 was determined prior to determining
extraction efficiency. (No L-3 was detected by qualitative analysis.) Efficacy in percent was
calculated as follows:
E = (Co - CfyCo-100% (6)
where:
E = efficacy
Co = mean concentration of agent without decontamination (determined from the
positive control coupons of each surface material)
Cf = concentration on a test coupon with decontamination.
A separate efficacy calculation was performed for each of the surface materials for L-l and/or
derL-1. For each material, a mean and percent relative standard deviation (%RSD) of efficacy
results were reported. Thus, the primary efficacy results from the coupon testing were placed in a
matrix table in which each entry shows the mean and %RSD of efficacy results for L-l or derL-1
18
-------
on each of the surface materials. The ratios of L-2 and derL-1 to L-l and derL-1, respectively,
were also reported.
A Student's t-test was used to compare the amount of L-l or derL-1 recovered from test coupons
to the amount of agent recovered from positive control coupons; p-values < 0.05 were considered
statistically significant.
Similar efficacy determination calculations were performed to compare the mass of arsenic
remaining on the coupons after decontamination and subsequent arsenic removal by wiping with
water or LeadOff.
19
-------
3.0 Test Results
3.1 Method Development and Demonstration Results
3.1.1 Extraction Efficiency for L-l Measured with GC/MS using a Cool On-Column Inlet
As shown in Figure 4, the efficiency with which L-l was extracted from test coupons depends on
the type of surface material onto which the Lewisite was applied and the solvent that was used to
extract the Lewisite. For all three solvents (acetone, hexane, and toluene), the highest extraction
efficiency was observed from glass, ranging from 65 % (hexane) to 100 % (toluene). The lowest
extraction efficiencies were observed from sealed concrete, ranging from 0 % (acetone) to 4 %
(toluene).For sealed concrete, only one of the three coupons returned a quantifiable amount of
LI.
Low extraction efficiencies may have alternative causes. First, extraction efficiencies may be low
due to adsorption or absorption of the L-l by the material or sealant. Second, extraction
efficiencies may be low due to a chemical reaction that converts L-l to other compounds, e.g.,
CVAA or CVAOA, that were not detected using the cool on-column GC/MS method. Hexane
was selected as the extraction solvent of choice for all GC/MS analysis of LI, L2, derL-1, and
derL-2 based on better extraction of L-l from wood.
100%
90%
80%
70%
60%
50%
40%
30%
20%
10% -
0%
I Glass
I Concrete
Metal
I Wood
Mean %RSD
Acetone
Mean %RSD
Hexane
Mean %RSD
Toluene
Figure 4. Results for L-l extraction efficiencies and %RSD (n = 3) by cool on-column
GC/MS.
20
-------
3.1.2 Extraction Efficiency and MDL for derL-1 Analyzed with GC/MS
As shown in Table 11, the extraction efficiency with which L-l and CVAA (measured as derL-1)
were extracted with hexane from glass and wood test coupons was demonstrated to meet data
quality objectives (recoveries within the range of 40 % to 120 % of the mass applied to the
coupons). Recoveries from wood were much lower and more variable than from glass. The MDL
from solution, glass, and wood were all 1.7 |ig/mL or less. These MDLs are at least a factor of
100 lower than the nominal mass of Lewisite (-1800 jig) extracted in 10 mL solvent. Therefore,
decontamination efficacies as high as 99 % can be determined without additional extraction
sample concentration steps.
Table 11. Hexane Extraction Efficiencies and Method Detection Limits for derL-1
Sample Source
Solution (no extraction)
Extracted from Glass
Extracted from Wood
Average %
Recovered, n=10
78
77
43
%RSD
7
6
17
MDL, jig/niL
1.3
1.1
1.7
3.1.3 Extraction Efficiency and MDL for Arsenic
As shown in Table 12, the efficiency with which arsenic, applied as arsenic trioxide, was
extracted from test coupons ranged from 85 % (metal) to 128 % (galvanized metal) using the
EPA Method 200.9.9 The efficiency with which arsenic, applied as arsenic trioxide was extracted
from the test coupons ranged from 90 % (wood) to 100 % (galvanized metal) using Certifier 65
method. Background levels of arsenic were detected in some materials, as indicated in Table 12
and shown in Table 13. Additional extraction efficiency tests in the presence of two
decontaminants (water and hydrogen peroxide) showed no impact on the observed extraction
efficiency values (data not shown).
21
-------
Table 12. Extraction Efficiencies (% Arsenic Applied that was Recovered) from Test
Coupons and Solution Controls
Solvent
Nitric Acid + Hydrochloric
Acid (EPA 200.9)9
Nitric Acid (Certifier 6)5
Material
Sealed concrete*
Wood flooring material*
Galvanized metal ductwork
Glass
Sealed concrete*
Wood flooring material
Galvanized metal ductwork*
Glass
Test Coupons,
n=3
110%
85%
128 %
105 %
97%
90%
100%
92%
Solution
Controls, n=3
93%
96%
*Background levels of arsenic were extracted from negative controls; see Table 13.
Table 13. Background Levels of Arsenic Detected in Coupon Materials
Sample Type
Negative Sealed
Concrete Controls
Negative Glass Controls
Negative Metal Controls
Negative Wood Controls
Certifier
Statistics,
Mean=
StDev=
%RSD=
Mean=
StDev=
%RSD=
Mean=
StDev=
%RSD=
Mean=
StDev=
%RSD=
6
MS
6.6
0.270
4%
<0.25
n/a
n/a
14.2
0.89
6%
<0.25
n/a
n/a
EPA 200.9
Statistics, ug
Mean= 13.8
StDev= 0.43
%RSD= 3 %
Mean= <0.25
StDev= n/a
%RSD= n/a
Mean= <0.25
StDev= n/a
%RSD= n/a
Mean= 1.90
StDev= 0.15
%RSD= 7.7 %
The nominal amount of arsenic spiked onto the coupons (as 1 |iL Lewisite) is 680 jig. Arsenic
background levels across all materials are, therefore, always less than 2 % from expected.
22
-------
As shown in Table 14, determining the extraction efficiencies using the Certifier 65 method to
extract arsenic from sealed concrete, wood, and glass samples spiked with arsenic trioxide met
data quality objectives (recoveries within the range of 40 % to 120 % of the mass of arsenic
applied to the coupons). MDLs from solution, glass, wood, and sealed concrete were all 3 ng/mL
or less. No MDL results were determined for metal because of high background levels of arsenic
extracted from the metal.
Table 14. Determination of Arsenic Method Detection Limits Using the Certifier 6 Method
Sample Source
Solution (no extraction)
Extracted from Glass
Extracted from Wood
Extracted from Sealed Concrete
Extracted from Metal
Average % Recovered,
n=7 coupons
101
112
90
91
*
%RSD
2.6
13.0
10
6.8
*
MDL, ng/mL
0.45
1.5
3.0
1.2
*
*No results obtained due to high background levels of arsenic.
3.2 Decontamination Results
3.2.1 Solution Test
A solution test was used to determine the efficacy of the decontaminants in the absence of
coupon materials. Sixty microliters of decontamination solution were added to a vial containing
1 uL of neat Lewisite (test solutions). The test solutions and positive controls (Lewisite that was
not decontaminated) were analyzed for L-l and (a derivatized sample) for derL-1 after a 15 min
reaction time. The results of a 15 minute solution test are shown in Table 15 where efficacy
indicates how much of the applied mass was not recovered, e.g., was decontaminated. Efficacies
for the 15 min reaction time with water were 97 %, measured as L-l, and 72 %, measured as
derL-1. Most of the L-l positive control was observed to be present by both the cool on-column
GC/MS measurement of L-l (76 %) and by GC/MS measurement of derL-1 (91 %). Contact
with water for 15 minutes reduced the measured L-l and derL-1 by 97 % (only 3 % of the mass
applied measured in the extract) and 78 % (only 22 % of the mass applied was measured in the
extract), respectively. These results suggest that the L-l in contact with water was converted to
CVAA as well as to unidentified degradation products. (CVAA was detected as derL-1, but not
in the L-l measurement).
No L-l (for hydrogen peroxide) or derL-1 (for hydrogen peroxide solution, bleach solution, or
DF 200 solution) were detected after 15 min contact. (Because hydrolysis in water effectively
converted all L-l to CVAA or other degradation product, analysis for L-l was not performed for
bleach and DF200 decontamination.)
23
-------
Table 15. Solution Decontamination Test (15 min Reaction Time)
Reaction
Time, min
No
Exposure
15
15
15
15
Decontamination
Method
Positive Control
Water Test
Hydrogen Peroxide
Test
Bleach Test
DF200 Test
Nominal
Mass of
Lewisite
Applied,
MS
1,880
1,880
1,880
1,880
1,880
Mean Total
Calculated
Mass of L-l,
MS
(%RSD)
1,427(11)
40(15)
<25
~
~
Mean
Efficacy,
% (L-l)
24
97
>98
~
~
Mean Total
Calculated
Mass of der
L-l, "g
(%RSD)
1,717(5)
480 (20)
<25
<25
<25
Mean
Efficacy, %
(derL-1)
9
72
>98
>98
>98
~ No test performed.
Based on peak area, no efficacy was observed for water against L-2, measured as L-2 or derL-2
in the solution test. In contrast, no L-2 or derL-2 were detected in the hydrogen peroxide, bleach,
or DF200 solution tests.
3.2.2 Efficiency Results Using Water to Decontaminate Lewisite on Glass or Wood
Water (deionized) was evaluated as a decontamination method for Lewisite as L-l on glass or
wood. The mean mass of L-l recovered from coupons and corresponding calculated
decontamination efficacies after a 30 min reaction time are summarized in Table 16. After 60
min (sum of 30 min delay time between application and start of decontamination plus 30 min
decontamination reaction time), 24 % of L-l was recovered from the glass positive controls.
Very low levels of L-l were recovered from glass (92 % mean efficacy) after a 30 min reaction
time with water. No L-l was recovered from wood, both from positive control coupons and from
test coupons after the 30 min reaction time with water.
Table 16. Water Decontamination Efficacy (measuring L-l)
Reaction
Time,
min
30
30
Material
Glass
Wood
Mass of
Lewisite
applied,
P&
1,580
1,580
Mean Positive
Controls Total
Mass of L-l,
P&
(%RSD)
373 (9)
<25T
Mean Test
Coupons
Total Mass
ofL-l,jig
(%RSD)
28
<25T
Mean
Positive
Controls
Recovery,
%
24
<2
Mean
Test
Coupons
Recovery,
%
2
<2
Mean
Efficacy,
%
(p value)
92
(<0.05)
*
*Cannot be determined because no L-l was recovered from positive control coupons.
value reported was based on the lowest calibration standard for a particular data set.
24
-------
Water (deionized) was evaluated as a decontamination method for Lewisite analyzed as derL-1
that includes both derivatized L-l and CVAA extracted from glass or wood after
decontamination. The mean mass of derL-1 recovered from coupons and corresponding
calculated decontamination efficacies after a 30 min or 60 min reaction time are summarized in
Table 17. Significant efficacy was observed at both 30 min and 60 min reaction times on both
glass (mean efficacy 31 % and 53 %, respectively) and wood (mean efficacy 81 % and 86 %,
respectively). Greater efficacy was observed at the longer reaction time for both materials.
However, the mass of derL-1 recovered from glass and wood was much greater than the mass of
L-l without derivatization, suggesting that L-l was converted to CVAA that was detected after
derivatization.
Table 17. Water Decontamination Efficacy (measuring derL-1)
Reaction
time,
min
30
30
60
60
Material
Glass
Wood
Glass
Wood
derL-1
Mass of
Lewisite
applied,
^g
1,490
1,490
1,610
1,610
Mean Positive
Controls Total
Mass of derL-1,
^g
(%RSD)
797 (3)
800 (8)
737 (5)
660 (10)
Mean Test
Coupons
Total
Mass of
Der L-l,
^g
(%RSD)
546 (9)
154 (23)
344(17)
92 (44)
Mean
Positive
Controls
Recovery,
%
53
54
46
41
Mean
Test
Coupons
Recovery,
%
37
10
21
6
IV/fp on
Efficacy,
%
(p value)
31
(<0.05)
81
(<0.05)
53
(<0.05)
86
(<0.05)
L-2 was detected on glass, but not on wood after 30 min reaction time with water.
DerL-2 was not detected on glass or wood positive control or test coupons after a 30 min
reaction time with water.
No L-l, derL-1, L-2, or derL-2 was found on any laboratory blank or procedural blank coupon.
CVAOA was detected (<25 jig) on two glass positive control coupons and not detected on the
third. CVAOA was detected (<25 jig) on all five glass test coupons at 30 minutes. CVAOA was
detected (>25 jig ) on all three wood positive control coupons and on four of five wood test
coupons after a 30 min reaction time with water; CVAOA was detected at <25 jig on the fifth
wood test coupon. CVAOA was not detected on glass or wood coupons after a 60 min reaction
time with water and was only detected at <250 jig on one of the three wood positive control
coupons.
25
-------
3.2.3 Efficiency Results Using Hydrogen Peroxide to Decontaminate Lewisite on Glass or
Wood
Hydrogen peroxide (3 %) was evaluated as a decontamination method for Lewisite as L-l on
glass and wood. The mean mass of L-l recovered from coupons and corresponding calculated
decontamination efficacies after a 30 min reaction time are summarized in Table 18. L-l was not
recovered after a 30 min reaction time with hydrogen peroxide. Mean efficacy was >95 %. No
L-l was found on any laboratory blank or procedural blank coupon.
Table 18. Hydrogen Peroxide Decontamination Efficacy Measured as L-l
Reaction
time, min
30
30
Material
Glass
Wood
Mass of
Lewisite
applied,
1,580
1,580
IX/IpOfl
Positive
Controls
Total Mass
of
T,_1 IIQ
*-* Lt> Hfe
(%RSD)
473 (19)
<100T
Mean Test
Coupons
Total
Mass of
L-l, jig
(%RSD)
<25
<100T
Mean
Positive
Controls
Recovery,
30
<6
Mean Test
Coupons
Recovery,
<2
<6
Mean
Efficacy,
(p value)
>94
(<0.05)
*
* Cannot be determined.
^The value reported was based on the lowest calibration standard for a particular data set.
No L-2 was detected on wood test coupons in the hydrogen peroxide 30 min or 60 min reaction
time decontamination tests. (No derL-2 was detected on the wood positive controls.)
Hydrogen peroxide (3 %) was evaluated as a decontamination method for Lewisite analyzed as
derL-1 that includes both derivatized L-l and CVAA extracted from glass or wood after
decontamination. The mean mass of derL-1 recovered from coupons and corresponding
calculated decontamination efficacies after a 30 min or 60 min reaction time are summarized in
Table 19. Significant efficacy was observed with 30 min reaction time on both glass (derL-1 not
detected; >96 % mean efficacy) and wood (derL-1 not detected on positive controls or test
coupons).
After the 30 min and 60 min reaction times with hydrogen peroxide, chromatographic peaks
were not detected for derL-2 in the glass extraction sample.
No derL-1 was found on any laboratory blank or procedural blank coupon.
In the 30 min hydrogen peroxide decontamination testing, qualitative analysis showed the
product CVAOA was detected at <25 jig on the glass positive control coupons. CVAOA
recovered from coupons was >25 jig on all three wood positive control coupons and >25 jig on
all glass and wood test coupons after a 30 min contact with hydrogen peroxide. In the 60 min
hydrogen peroxide decontamination testing, CVAOA was not detected on the glass and wood
positive control coupons. CVAOA was >250 jig on four of five glass test coupons, detected on
26
-------
one glass coupon at <250 jig, and detected at <250 jig on all wood test coupons after 60 min
decontamination with hydrogen peroxide.
Table 19. Hydrogen Peroxide Decontamination Efficacy (measuring derL-1)
Reaction
time, min
30
30
60
60
Material
Glass
Wood
Glass
Wood
derL-1
Mass of
Lewisite
applied,
MS
1,490
1,490
1,610
1,610
Mean
Positive
Controls
Total
Mass of
derL-1, ug
(%RSD)
637(11)
470(18)
490 (7)
387 (20)
Mean Test
Coupons
Total
Mass of
derL-1, ug
(%RSD)
<25
<25
<25
<25
Mean
Positive
Controls
Recovery,
%
43
32
30
24
Mean
Test
Coupons
Recovery,
%
<2
<2
<2
<2
Mean
Efficacy,
%
(p value)
>96 (<0.05)
>95 (<0.05)
>95 (<0.05)
>94 (<0.05)
3.2.4 Efficiency Results Using Bleach to Decontaminate Lewisite on Glass or Wood
Bleach (8.7 % sodium hypochlorite solution by redox titration) was evaluated as a
decontamination method for Lewisite analyzed as derL-1 that includes both derivatized L-l and
CVAA extracted from glass or wood after decontamination. The mean mass of derL-1 recovered
from coupons and corresponding calculated decontamination efficacies after a 30 min or 60 min
reaction time are summarized in Table 20. Significant efficacy was observed with 30 min
reaction time on both glass (derL-1 not detected; >96 % mean efficacy) and wood (94 % mean
efficacy) and for the 60 min reaction time on wood (derL-1 not detected; >97 % mean efficacy).
Glass was not included in the 60 min reaction time evaluation because no derL-1 was detected
after the 30 min reaction time.
No derL-2 was detected on positive control or test coupons after bleach decontamination of glass
and wood after a 30 reaction time or wood after a 60 min reaction time.
No derL-1 or derL-2 was found on any laboratory blank or procedural blank coupon.
Since no derL-1 was detected in any of the test coupon abstracts following bleach
decontamination, no efforts were made to measure the LI or L2 amount by cool on-column
GC/MS.
27
-------
Table 20. Bleach Decontamination Efficacy (measuring derL-1)
Reaction
time, min
30
30
60
Material
Glass
Wood
Wood
derL-1
Mass of
Lewisite
applied, ug
1,570
1,570
1,420
Mean
Positive
Controls
Total Mass
of derL-1,
(%RSD)
670 (24)
593 (22)
723 (4)
Mean
Coupons
Total
Mass of
derL-1,
(%RSD)
<25
36 (104)
<25
Mean
Positive
Controls
Recovery,
43
38
51
Mean
Test
Coupons
Recovery,
<2
2
<2
Mean
Efficacy,
(p value)
>96
(<0.05)
94
(<0.05)
>97
(<0.05)
CVAOA was detected at <250 jig for glass and wood positive control coupons after 30 min
decontamination with bleach. CVAOA was also detected at <250 jig for all five wood test
coupons and four of five glass test coupons for the 30 min reaction time. No CVAOA was found
after 60 min on wood positive control coupons. After 60 min of bleach decontamination <250 jig
CVAOA was present on one wood test coupon; no CVAOA was detected on four coupons.
Analysis for CVAOA after the 60 min of bleach decontamination of glass was not conducted.
3.2.5 Efficiency Results Using DF200 to Decontaminate Lewisite on Glass or Wood
DF200 was evaluated as a decontamination method for Lewisite extracted from glass or wood
after decontamination and analyzed as derL-1. The mean mass of derL-1 recovered from
coupons and corresponding calculated decontamination efficacies after a 30 min or 60 min
reaction time are summarized in Table 21. Significant efficacy was observed with 30 min
reaction time on both glass (derL-1 not detected; >96 % mean efficacy) and wood (87 % mean
efficacy) and for the 60 min reaction time on wood (derL-1 not detected; >97 % mean efficacy).
Glass was not included in the 60 min reaction time evaluation because no L-l was detected after
the 30 min reaction time.
No derL-1 was found on any laboratory blank or procedural blank coupon.
No derL-2 was detected on positive control or test coupons after DF200 decontamination of glass
and wood after a 30 reaction time or wood after a 60 min reaction time.
Since no derL-1 was detected in any of the test coupon abstracts following DF200
decontamination, no efforts were made to measure the LI or L2 amount by cool on-column
GC/MS.
28
-------
Table 21. DF200 Decontamination Efficacy (measuring derL-1)
Reaction
Time,
min
30
30
60
Material
Glass
Wood
Wood
derL-1
Mass of
Lewisite
applied,
1,570
1,570
1,420
Mean
Positive
Controls
Total Mass
of derL-1,
(%RSD)
593(18)
770 (6)
743 (9)
Mean Test
Coupons
Total Mass
of derL-1,
(%RSD)
<25
102(61)
<25
Mean
Positive
Controls
Recovery,
38
49
52
Mean
Test
Coupons
Recovery,
<2
7
<2
Mean
Efficacy,
(p value)
>96
(<0.05)
87 (<0.05)
>97
(<0.05)
In the 30 min DF200 decontamination testing, qualitative analysis showed the by-product
CVAOA was <250 |ig/mL on the glass and wood positive control and test coupons, except that
one glass test coupon had CVAOA >250 |ig/mL. In the 60 min DF200 decontamination testing,
qualitative analysis showed the product CVAOA was not detected on the wood positive control
coupons, but was detected at <250 |ig/mL on four of five test coupons. (No glass coupons were
included in the 60 min DF200 decontamination testing.)
3.3 Arsenic Removal Results
Coupons that were spiked with Lewisite and decontaminated with water or with hydrogen
peroxide were subsequently sprayed with either water or with LeadOff and wiped with a wetted
gauze pad (2 inch x 2 inch gauze sponge). As shown in Table 22, arsenic extracted with nitric
acid and analyzed with GFAA from coupons was substantially reduced on glass after spraying
with water (85 % - 92 % mean efficacies) or LeadOff (92 %-98 % mean efficacies) and wiping.
Neither removal procedure (water nor LeadOff) was efficacious at removing arsenic from wood.
High variability in the amounts of residual arsenic on the un-wiped coupons was noted and in
some cases with wood, the amount of arsenic recovered after wiping was greater than the amount
of arsenic detected before wiping. The cause of this anomaly was not investigated, but the
presence of residual arsenic may have been related to conditions that varied between the day the
unwiped coupons were prepared and decontaminated and the day the wiped coupons were
prepared and decontaminated.
29
-------
Table 22. Arsenic Removal Efficiencies Using Water and LeadOff
Decontaminant,
Material, and
Coupon Type from
Lewisite
Decontamination
Test
Water, Glass
Positive Control
Water, Glass Test
Coupon
Water, Wood
Positive Control
Water, Wood Test
Coupon
Hydrogen Peroxide,
Glass Positive
Control
Hydrogen Peroxide,
Glass Test Coupon
Hydrogen Peroxide,
Wood Positive
Control
Hydrogen Peroxide,
Wood Test Coupon
No Removal
Total Mass of
Arsenic, \ig
(%RSD)
251 (17)
425(11)
263 (55)
308 (35)
251 (17)
417 (22)
263 (55)
421 (21)
Water Wiping
Total Mass of
Arsenic, \ig
(%RSD)
29 (14)
66 (53)
448 (10)
461 (5)
29 (70)
35 (52)
527(17)
499(18)
Water
Efficacy,
%
88
85
0*
0*
90
92
0*
0*
LeadOff
Wiping
Total Mass of
Arsenic, \ig
(%RSD)
16(80)
13 (20)
452(19)
419(13)
21 (78)
6.5 (76)
444(17)
518(4)
LeadOff
Efficacy,
%
94
97
0*
0*
92
98
0*
0*
*Recovery after wiping was greater than recovery before wiping.
3.4 Observations of Damage to Coupons
The decontamination treatment resulted in no obvious visible change to any of the coupons.
Example photographs before and after the decontamination treatment are shown in Figure 5.
30
-------
Blank Water Hydrogen Bleach DF200
Peroxide
a4 tju uJ VU wJ
0 min
30 min
31
-------
Blank
Water
Hydrogen
Peroxide
Bleach
DF200
60 min
Figure 5. Photographs of coupons before and after decontamination treatment.
32
-------
4.0 Quality Assurance/Quality Control
4.1 Control of Monitoring and Measuring Devices
Quality control (QC) requirements and results are shown in Table 23.
Table 23. Data Quality Objectives and Results for the Test Measurements
Parameter
Measurement
Method
QC Requirement
Results
Temperature,
°C
NIST*-traceable
thermometer
Compare against calibrated
thermometer once before testing, agree
±1°C
Accuracy of thermometer
was acceptable
Relative
Humidity, %
NIST*-traceable
hygrometer
Compare against calibrated hygrometer
once before testing, agree ±10 % (full
scale)
Accuracy of thermometer
was acceptable
Time, sec
Timer/data logger
Compare once before testing; agree
±2 sec/hour
Accuracy of laboratory
clock was acceptable
Volume, uL
Calibrated pipette
Check pipettes for accuracy and
repeatability one time before use by
determining the mass of water
delivered. The pipette was acceptable if
the range of observed masses for five
droplets was ±10 % of expected.
Received with a
calibration certificate
from manufacturer
Agent on
Positive
Control,
ug/mL
Extraction, GC
The mean percent recovery for a known
quantity of each analyte added to a test
coupon used to gauge recovery must
fall within the range of 40 %-120 % and
have a coefficient of variation of <30 %
between replicates
All were within the
acceptable range for
percent recovery using
hexane.
Agent on
Laboratory
Blank, ug/mL
Extraction, GC
Laboratory blanks should have less than
1 % of the amount of derL-1 compared
to that found on positive controls
All laboratory blanks
were blank (no agent
detected)
Agent on
Procedural
Blank, ug/mL
Extraction, GC
Procedural blanks should have less than
10 % of the amount of derL-1 compared
to that found on positive controls
All were within the
acceptable range
Arsenic on
Positive
Control,
ug/mL
Extraction,
GFAA
The mean recovery of arsenic must be
40 %-120 % and have a coefficient of
variation of <30 % between replicates
33
All were within the
acceptable range for mean
recovery using Certifier 6
and EPA 200.9. Observed
range 87.9 % - 99.7 %.
The coefficient of
variance was with the
range of acceptance.
-------
Parameter
Measurement
Method
QC Requirement
Results
Observed range <1%-11
Arsenic on
Laboratory
Blank, ug/mL
Extraction,
GFAA
Laboratory blanks should have less than
1 % of the amount of arsenic compared
to that found on positive controls
In method development,
metal and sealed concrete
were shown to have
background levels of
arsenic greater than 1 %
of the arsenic recovered
from positive controls.
Only glass and wood,
which had no arsenic
above the detection limit
(<2.5 ng/mL), were used
in arsenic removal testing.
Arsenic on
Procedural
Blank, ug/mL
Extraction,
GFAA
Procedural blanks should have less than
10 % of the amount of arsenic
compared to that found on positive
controls
All were within the
acceptable range
*NIST is the National Institute of Standards and Technology.
4.2 Equipment Calibrations
The instrumentation used to determine L-l, derL-1, CVAOA, and arsenic is identified in Section
2.5.1. The required analytical equipment was maintained and operated according to the quality
requirements and documentation of the BBRC. All equipment was calibrated at the time of use
and at the frequency specified in Table 24. The LC/MS equipment was not explicitly calibrated
as the CVAOA analysis was semi-qualitatively only.
Table 24. Equipment Calibration Schedule
Equipment
Calibrated Pipette
Calibrated
Hygrometer/Thermometer
GC/MS
GFAA
Frequency
Prior to testing and every six months thereafter
Prior to testing and annually thereafter
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
A calibration curve was analyzed and saved prior to testing. Calibration
verification standards were run at the beginning of each sample batch,
every six samples, and at the end of a batch of samples
34
-------
For both GC/MS and GFAA spectrometry, independently prepared continuing calibration
verification (CCV) standards were analyzed prior to sample analysis, following at least every six
samples and at the end of each batch of samples. Two or more CCV concentrations were used,
one of which was equal to the low calibration standard and the other(s) within the calibration
range. CCV response within 25 % (for the low standard) or 15 % of nominal concentration was
acceptable. Samples analyzed prior to or following CC Vs that were outside of acceptance limits
were re-analyzed, except that the low CCVs for direct measurement of L-l and L-2 sometimes
failed and were not repeated. In those cases, the lowest acceptable calibration provided the
lowest value for the calibration curve. (See Section 4.7 for a discussion of this deviation from the
test/QA plan.)
At least a five-point calibration was used for each batch of samples for analysis for L-l with a
lower level of approximately 2.5 |ig/mL and an upper range of approximately 150 |ig/mL. The
GC/MS calibration curves met the following performance requirements:
• r2 greater than 0.99
• % bias for the lowest standard less than 25 %
• % bias for the remaining standards less than 15 %.
Standards do not exist for L-2 and L-3 so only a qualitative analysis of these species of Lewisite
was performed. They were reported as a ratio to L-l.
At least a five-point calibration was used for each batch of samples for analysis for derL-1 with a
lower level of-2.5 |ig/mL and an upper range of-60 |ig/mL. The GC/MS calibration curves met
the following performance requirements:
• r2 greater than 0.99
• % bias for the lowest standard less than 25 %
• % bias for the remaining standards less than 15 %.
Standards do not exist for derL-2 so only a qualitative analysis of this species of Lewisite was
performed. They were reported as a ratio to derL-1.
At least a five-point calibration curve for arsenic was used with a lower calibration level of-2.5
nanograms (ng)/mL and an upper range of-50 ng/mL.
The GFAA calibration curves met the following performance requirements:
• r2 greater than 0.99
• % bias for the lowest standard less than 25 %
• % bias for the remaining standards less than 15 %.
The calibration curve was verified each day of use with the analysis of two calibration standards,
one of which was equal to the low calibration level and the other within the calibration range.
Independently prepared CCV standards were analyzed prior to sample analysis, following at
least every six samples and at the end of each batch of samples. Two or more CCV
35
-------
concentrations were used, one of which was equal to the low calibration standard and the other(s)
within the calibration range. CCV response within 25 % (for the low standard) or 15 % of
nominal concentration was acceptable. Samples analyzed prior to or following CCVs that were
outside of acceptance limits were re-analyzed.
4.3 Technical Systems A udit
The Quality Assurance (QA) Manager performed a Technical Systems Audit (TSA) during the
performance of the decontamination testing. The purpose of the TSA was to ensure that testing
was performed in accordance with the test/QA plan. In the audit, the QA Manager reviewed the
sampling and analysis methods used, compared actual test procedures to those specified in the
test/QA plan and Amendment 1, and reviewed handling procedures. The QA Manager prepared a
report, the findings of which were addressed either by modifications to the test procedures or by
documentation in the test records. TSA results are summarized in Table 25.
Table 25. TSA Results
Reference
Finding
Corrective Action
Amendment 1, Table Al
While Table Al in the amendment does specify
that GC/MS operating conditions may be
modified by the analyst as needed to optimize
performance, it does not really cover the changes
needed for analysis of derivatized samples.
A formal deviation was
prepared to reflect the
changes needed for
GC/MS of derivatized
Lewisite.
4.4 Performance Evaluation A udits
A performance evaluation (PE) audit was conducted for temperature (±1 °C), relative humidity
(±10 %), and time (±1 sec/min). Results are shown in Table 26.
Table 26. PE Results
Parameter
Temperature
Relative
Humidity
Time
Audit Procedure
Compare to independent
thermometer value
Compare to independent
hygrometer value
Compare time to
independent clock
Expected Tolerance
±1°C
±10 %
±1 sec/min
Results
All were within the
acceptable range
(±0.9 °C)
All were within the
acceptable range
(±5 %)
All were within the
acceptable range
(±0 sec/min)
36
-------
4.5 Data Quality Audit
The QA Manager audited at least 10 % of the investigation data and traced the data from initial
acquisition, through reduction and statistical comparisons, to final reporting. All data analysis
calculations were checked.
4.6 Amendments
One amendment was requested to evaluate alternative analyses for Lewisite degradation
products: derivatization and analysis by GC/MS and use of LC/MS. Decisions on
decontamination methods, weathering time (time between contamination and decontamination),
changes in coupons used, and minor edits were also captured in this amendment.
4.7 Deviations
CCV Deviation. Some low calibration standards failed to satisfy the 15 % criterion and low
CCVs failed to satisfy the 25 % criterion for the cool on-column GC/MS analyses. The cool on-
column inlet allows for an aqueous sample to be directly deposited onto the capillary column,
minimizing heat-induced degradation of Lewisite and enabling L-l and L-2 to be measured.
However, direct sample injection degrades the column phase more quickly resulting in poor
chromatography. Study samples were numerous, resulting in frequent reduction in analyte
sensitivity and costly instrument maintenance. The arsenic component of these samples also
contributes to loss of instrument performance. The ion source becomes dirty over time as sample
components accumulate, resulting in deteriorating performance accelerated by the cool on-
column injections.
The impact of the CCV deviation on study data was minimal. Changes in calibration range were
noted on all affected analyses. Multiple CCVs were included in every analysis and only the
lower end of the calibration range required adjustment. Detectable results below the verified
calibration range were noted as "less than" the lowest calibration concentration. All study
samples were also derivatized for additional GC/MS analysis that was not impacted by
calibration challenges.
Analytical Changes to Address MDL of DerL-1 Rather than L-l. Method detection limits were
determined for derivatized L-l rather than L-l. Because the decontamination methods were all
aqueous, hydrolysis of L-l would be expected in all cases. Therefore, the use of derivatization,
which detects and measures both L-l and its similarly harmful hydrolysis product CVAA, was
considered to be the more useful analytical approach. Therefore the MDL was determined for the
derivatized products using GC/MS rather than L-l using cool on-column GC/MS.
The impact of the deviation was to improve the usefulness of results compared to measuring only
changes in L-l and L-2 with decontamination.
Internal Standard Deviation. An internal standard was not included in analysis standards or
sample extractions; an alternative method for calculating Lewisite concentration that did not
include measurement of internal standards was used. In the cool on-column method development
the instrument response for the internal standard varied widely for analysis of both samples and
37
-------
calibration standards. A suitable internal standard for the derivatization of Lewisite was also
unknown. Additional methods development would be required to determine a suitable internal
standard for use with both Lewisite analysis methods that would substantially add to the project
scope. Therefore an internal standard was not used.
An alternative method was used to calculate Lewisite concentration that did not include
measurement of an internal standard concentration. Lewisite concentration in a coupon extract or
positive control solution sample was determined by Equation 7.
As = a'Cs 2 + b'Cs + c (7)
where:
As = Area of the target analyte peak in the sample
Cs = Concentration of the target analyte in the sample
a, b, c = Constants
Procedures in the QAPP provided other control measures that minimized the impact of not
having an effective internal standard. Dilutions of Lewisite were prepared on every test day to
verify acceptable purity of the neat Lewisite. These dose confirmation samples were prepared
using the same positive displacement pipette and a volumetric flask to further aid in identifying
any potential equipment failures. All results were acceptable (±25 % of expected).
Each test also included multiple positive control samples for each building material. Positive
controls undergo the same manipulation as test samples providing a reference on which
decontamination efficacy was established.
38
-------
5.0 Summary
The objective of this evaluation was to develop/demonstrate methods and apply the methods to
determine the neutralization efficacies of various readily-available, liquid-based methods for
Lewisite decontamination.
Because decontamination of Lewisite generates arsenical compounds, physical removal of
arsenic may be required to adequately remediate a contaminated site or facility. Here, the amount
of residual arsenic remaining on building materials after decontamination and spraying with
water or LeadOff followed by wiping with wetted gauze was evaluated.
Method development was used to determine extraction efficiencies for L-l extracted from four
materials (sealed concrete, wood flooring, galvanized metal, and glass). Three solvents were
evaluated: acetone, hexane, and toluene. Efficiencies varied by material. The best L-l recovery
efficiencies were from glass: 81 % for acetone, 65 % for hexane, and 100 % for toluene.
Recoveries from sealed concrete were very low, ranging from 0 % for acetone to 4 % for
toluene.
Because L-l hydrolyzes in the presence of water, method demonstration was included to show
that derivatization could be used during analysis to determine the mass of derL-1 (the product of
derivatization of both L-l and CVAA). Extraction efficiencies for derL-1 from glass and wood
were 77 % and 43 % respectively. The method detection limit analyzing derL-1 using GC/MS
was determined to be 1.1 ug/mL from glass and 1.7 ug/mL from wood. (Glass and wood were
selected for subsequent decontamination testing; derL-1 extraction efficiencies were not
determined from sealed concrete and metal.)
Extraction efficiencies for arsenic were also determined in method development. Arsenic
recoveries ranged from 85 % from galvanized metal to 110 % from sealed concrete using the
EPA 200.9 extraction method and 89 % from glass to 100 % from galvanized metal using the
Certifier 6 extraction method. Background levels of arsenic were detected in the negative
controls for sealed concrete using both extraction methods; from wood flooring using EPA
200.9; and from galvanized metal using Certifier 65. The MDL using the Certifier 65 method was
3 ng/mL or less from solution, glass, wood, and sealed concrete.
Because of the chemical properties of these various vesicant compounds, multiple methods of
analysis were required. L-l, L-2, and L-3 have low solubility in water and are more volatile than
CVAA and CVAOA which are water soluble, thus complicating extraction and analysis of
Lewisite and degradation products. Method development identified hexane as a single extraction
for subsequent analysis of L-l, L-2, and L-3 (analyzed using a cool on-column inlet with GC/MS
as well as by derivatization and GC/MS) and CVAA and BCVAA (using derivatization and
GC/MS). Acetone was used to extract CVAOA that was analyzed using LC/MS.
The amount of L-l recovered from glass positive controls after a combined 60 min interaction
time of Lewisite on glass was less than 30 % while for wood, no L-l was extracted (below
detection limit). This decrease can be attributed to a combination of evaporation, degradation,
and absorption of L-l.
39
-------
Decontamination efficacy was evaluated for four decontaminants: water, hydrogen peroxide (3
%), bleach (8.7 % hypochlorite), and DF200. The results of mixing the decontaminant with
Lewisite (no coupons) was summarized in Table 27. L-l recovered from coupons was at or
below 2 % after 15 min contact with water or hydrogen peroxide. However, 22 % of derL-1 was
recovered, suggesting that a significant amount of L-l had been converted to CVAA in the
presence of water. An important finding is that evaluating the decontamination efficacy against
L-l alone may lead to an inaccurate conclusion that vesicant properties have been eliminated.
Decontamination products with vesicant properties, such as CVAA, may still be present.
Decontamination with hydrogen peroxide, bleach, and DF200 was >98 % efficacious (measured
as of derL-1).
Table 27. Efficacy (% reduction) of Decontaminants Analyzed as L-l or derL-1
Reaction time,
min
No Exposure
15
15
15
15
Decontamination
Method
Positive Control
Water Test
Hydrogen Peroxide Test
Bleach Test
DF200 Test
Mean Efficacy, %
(L-l)
[76 % of applied mass
recovered]
97
>98
~
~
Mean Efficacy, %
(derL-1)
[91 % of applied mass
recovered]
72
>98
>98
>98
The results of decontamination by various methods for 30 min or 60 min reaction times are
shown in Table 28. While all methods showed efficacy, water exhibited the lowest efficacy at 30
min and no additional efficacy was observed with the longer (60 min) reaction time. No derL-1
was detected after a 30 or 60 min reaction time with hydrogen peroxide applied to either glass or
wood. A small amount of derL-1 was detected after a 30 min reaction time for the derL-1 on
wood (but not glass) with both bleach and DF200; after a 60 min reaction time with bleach or
DF200 no derL-1 was detected.
40
-------
Table 28. Mean Efficacy of Decontamination Methods Analyzed as DerL-1
Reaction Time, min
30
30
60
60
30
30
60
60
30
30
60
30
30
60
Material
Glass
Wood
Glass
Wood
Glass
Wood
Glass
Wood
Glass
Wood
Wood
Glass
Wood
Wood
Decontamination Method
Water
Water
Water
Water
Hydrogen Peroxide
Hydrogen Peroxide
Hydrogen Peroxide
Hydrogen Peroxide
Bleach
Bleach
Bleach
DF200
DF200
DF200
Mean Efficacy, % (p value)
31 (<0.05)
81 (<0.05)
53 (<0.05)
86 (<0.05)
>96 (<0.05)
>95 (<0.05)
>95 (<0.05)
>94 (<0.05)
>96 (<0.05)
94 (<0.05)
>97 (<0.05)
>96 (<0.05)
87 (<0.05)
>97 (<0.05)
The amount of L-2 present in samples was defined as the percentage of L-2 to L-l
chromatographic peak area. The percent of L-2 relative to L-l was 53 % for glass positive
control coupons in the 30 min water decontamination test and substantially higher for
corresponding test coupons. DerL-2 was not detected for 30 min decontamination testing. No
derL-2 was found in positive controls coupons, but 4.5 % (relative to derL-1) was detected in 60
min water decontamination glass test coupons and 6.4 % in wood test coupons. No L-2 or derL-2
was detected after decontamination with hydrogen peroxide, bleach, or DF200.
Tables 29 and 30 provide a summary of the results for L-l and L-2 as determined using both the
cool on-column approach and the derivatization during analysis of the Lewisite. All four
methods demonstrated efficacy. However, water was not effective at reducing the Lewisite
below the limits of detection. Hydrogen peroxide, bleach, and DF200 decontamination were able
to reduce both derL-1 and derL-2 below the limits of detection.
41
-------
Table 29. Summary of Efficacy Results for L-l and DerL-1
Form of Agent Analyzed and Decontaminant
L-l conversion by water
L-l conversion by hydrogen peroxide (3 %)
DerL-1 conversion by water
DerL-1 conversion by hydrogen peroxide (3 %)
DerL-1 conversion by bleach (8.7 % hypochlorite)
DerL-1 conversion by DF200
Efficacy on Building Materials
30 min Reaction Time
Glass
Hcin
30 min Reaction Time
Glass
Wood
Wood
Wood
60 min Reaction Time
Glass
Wood
Not tested
Wood
Not tested
*: Insufficient amount of L-l recovered from positive controls after 30 min.
Key:
Efficacy less than 87 % for agent in specified form, e.g., L-l or derL-1.
Agent detected on some of the test coupons in specified form with efficacy greater than 87 %.
No agent detected in specified form and efficacy greater than 94 %.
42
-------
Table 30. Summary of Efficacy Results for L-2 and DerL-2
Form of Agent Analyzed and Decontaminant
Detection on Building Materials
Reaction Time:
30 min
60 min
L-2 conversion by water
Glass
L-2 conversion by hydrogen peroxide (3 %)
Not Tested
Not Tested
Not Tested
Not Tested
Glass
DerL-2 conversion by water
Glass
Wood (Detected on Two
of Five Coupons)
DerL-2 conversion by hydrogen peroxide (3 %)
DerL-2 conversion by bleach (8.7 % hypochlorite)
DerL-2 conversion by DF200
Key:
Detected in specified form, e.g., L-2 or derL-2
No agent detected in specified form
Not Tested
Results for analysis of coupon extracts seeking to detect CVAOA are summarized in Table 31.
Analysis for CVAOA using LC/MS showed that a brief (60 min) contact with wood appears to
convert some L-l into relatively large quantities of CVAOA (>250 ng). After a one hour reaction
time with water, CVAOA was not detected on glass or wood. The dynamics of CVAOA
formation and degradation were not obvious from these results. In the presence of hydrogen
peroxide for 30 or 60 minutes, amounts of CVAOA > 250 ng may be formed.
43
-------
Table 31. Results of Qualitative Analysis for CVAOA
Reaction time,
min
30
30
60
60
30
30
60
60
30
30
60
30
Material
Glass
Wood
Glass
Wood
Glass
Wood
Glass
Wood
Glass
Wood
Wood
Glass
Decontamination
Method
Water
Water
Water
Water
Hydrogen Peroxide
Hydrogen Peroxide
Hydrogen Peroxide
Hydrogen Peroxide
Bleach
Bleach
Bleach
DF200
CVAOA
Positive Controls,
n = 3
Non-detect: 1
Detect, < 25 (ig: 2
Detect, > 25 (ig: 0
Non-detect: 0
Detect, < 25 (ig: 0
Detect, >25 (ig: 3
Non-detect: 3
Detect, < 250 (ig: 0
Detect, > 250 (ig: 0
Non-detect: 2
Detect, < 250 (ig: 1
Detect, > 250 (ig: 0
Non-detect: 0
Detect, < 25 (ig: 3
Detect, > 25 (ig: 0
Non-detect: 0
Detect, < 25 (ig: 0
Detect, > 25 (ig: 3
Non-detect: 3
Detect, < 250 (ig: 0
Detect, >250 (ig: 0
Non-detect: 3
Detect, < 250 (ig: 0
Detect, >250 (ig: 0
Non-detect: 0
Detect, < 250 (ig: 3
Detect, > 250 (ig: 0
Non-detect: 0
Detect, < 250 (ig: 3
Detect, >250 (ig: 0
Non-detect: 3
Detect, < 250 (ig: 0
Detect, > 250 (ig: 0
Non-detect: 0
Detect, < 250 (ig: 3
Detect, >250 (ig: 0
CVAOA
Test Coupons,
n=5
Non-detect: 0
Detect, < 25 (ig: 5
Detect, > 25 (ig: 0
Non-detect: 0
Detect, < 25 (ig: 1
Detect, > 25 (ig: 4
Non-detect: 5
Detect, < 250 (ig: 0
Detect, > 250 (ig: 0
Non-detect: 5
Detect, < 250 (ig: 0
Detect, > 250 (ig: 0
Non-detect: 0
Detect, < 25 (ig: 0
Detect, > 25 (ig: 5
Non-detect: 0
Detect, < 25 (ig: 0
Detect, >25 (ig: 5
Non-detect: 0
Detect, < 250 (ig: 1
Detect, >250 (ig: 4
Non-detect: 0
Detect, < 250 (ig: 5
Detect, >250 (ig: 0
Non-detect: 1
Detect, < 250 (ig: 4
Detect, >250 (ig: 0
Non-detect: 0
Detect, < 250 (ig: 5
Detect, >250 (ig: 0
Non-detect: 4
Detect, < 250 (ig: 1
Detect, >250 (ig: 0
Non-detect: 0
Detect, < 250 (ig: 4
Detect, > 250 (ig: 1
44
-------
Reaction time,
min
30
60
Material
Wood
Wood
Decontamination
Method
DF200
DF200
CVAOA
Positive Controls,
n = 3
Non-detect: 0
Detect, < 250 (ig: 3
Detect, > 250 (ig: 0
Non-detect: 3
Detect, < 250 (ig: 0
Detect, > 250 (ig: 0
CVAOA
Test Coupons, n=5
Non-detect: 0
Detect, < 250 (ig: 5
Detect, > 250 (ig: 0
Non-detect: 1
Detect, < 250 (ig: 4
Detect, > 250 (ig: 0
Removal of residual arsenic after decontamination with water and hydrogen peroxide was
evaluated by wiping with a gauze pad wetted with either water of LeadOff As shown in Table
32, spraying the coupon and wiping with gauze wetted with water and LeadOff were efficacious
removing arsenic from glass. LeadOff removal efficiencies from glass were, in each case,
slightly better than using water. Removal of arsenic from wood was ineffective with either water
spray or wiping or LeadOff spray and wiping. It was assumed that the arsenic soaks into the
wood where it was not readily removed by wiping.
Table 32. Summary of Arsenic Removal Efficiency Using Gauze Wetted with Water or LeadOff
Decontaminant, Material, and Coupon
Type
Water, Glass Positive Control
Water, Glass Test Coupon
Water, Wood Positive Control
Water, Wood Test Coupon
Hydrogen Peroxide, Glass Positive Control
Hydrogen Peroxide, Glass Test Coupon
Hydrogen Peroxide, Wood Positive Control
Hydrogen Peroxide, Wood Test Coupon
Water Efficacy, %
88
84
0*
0*
89
92
0*
0*
LeadOff Efficacy, %
94
97
0*
0*
92
98
0*
0*
*Recovery after wiping was greater than recovery before wiping.
In summary, hydrogen peroxide, bleach, and DF200 all showed significant efficacy against
Lewisite and associated vesicant decontamination byproducts. Wiping glass (and possibly other
smooth surfaces) with gauze wetted with water or LeadOff appears efficacious. Wiping wood
(and possibly other rough and/or porous surfaces) may not be efficacious for arsenic removal.
45
-------
6.0 References
1. M. Goldman and J. C. Dacre. (1989). Lewisite: its chemistry, toxicology, and biological
effects. Reviews of Environmental Contamination and Toxicology 110: 75-115.
2. 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.
3. K. K. Kroening, R. N. Easter, D. D. Richardson, S. A. Willison, J. A. Caruso. (2011).
Analysis of Chemical Warfare Degradation Products. West Sussex, UK: John Wiley & Sons,
p36-42.
4. N. Munro, S. S. Talmage, G. D. Griffin, L. C. Waters, A. P. Watson, J. F. King, V.
Hauschild. (1999). The sources, fate, and toxicity of chemical warfare agent degradation
products. Environmental Health Perspectives 107(12):933-974.
5. M.B. De La Calle, T. Linsinger, H. Emteborg, J. Charoud-Got, and I Verbist. 2010. Report of
the seventh interlaboratory comparison organized by the European Union - Reference
Laboratory for Heavy Metals in Feed and Food. IMEP-107: Total and inorganic As in rice.
https://irmm.irc.ec.europa.eu/EURLs/EURL heavy metals/inteiiaboratory comparisons/ime
p-107/Documents/eur24314en.pdf , Accessed October 7, 2012.
6. U.S. Envionmental Protection Agency Selected Analytical Methods for Environmental
Remediation and Recovery (SAM), http://www.epa.gov/sam/index.htm. Accessed May 1,
2014.
7. B. Muir, B. J. Slater, D. B. Cooper, and C.M. Timperley. 2004. 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 (2), 313-20.
8. S. K. Hanaoka, T. Nomura, and T. Wada. (2006). Determination of Mustard and Lewisite
Related Compounds in Abandoned Chemical Weapons (Yellow Shells) from Sources in
China and Japan. Journal of Chromatography A , 1101 (1-2), 268-77.
9. J.T. Creed, T.D. Martin, and J.W. O'Dell. 1994. Method 200.9. Determination of Trace
Elements by Stabilized Temperature Graphite Furnace Atomic Absorption. Revision 2.2.
EMMC Version.
http://water.epa.gov/scitech/methods/cwa/bioindicators/upload/2007 07 10 methods metho
d 200 9.pdf. Accessed April 3, 2014.
46
-------
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
------- |