United States
Environmental Protection
Agency
Decontamination of Toxic Industrial
Chemicals and Chemical Warfare
Agents on Building Materials Using
Chlorine Dioxide Fumigant and
Liquid Oxidant Technologies
TECHNOLOGY INVESTIGATION REPORT
Office of Research and Development
National Homeland Security Research Center
-------
-------
EPA/600/R-09/012 | Febraary 2009 www.epa.gov/ord
TECHNOLOGY INVESTIGATION REPORT
Decontamination of Toxic Industrial
Chemicals and Chemical Warfare Agents
On Building Materials Using Chlorine
Dioxide Fumigant and Liquid Oxidant
Technologies
By
James Rogers, Timothy Hayes, Donald Kenny, Ian MacGregor, Karen Tracy,
Robert Krile, Marcia Nishioka, Michael Taylor, Karen Riggs, and Harry Stone
Battelle
505 King Avenue
Columbus, OH 43201
Shawn Ryan
Task Order Project Officer
National Homeland Security Research Center
Office of Research and Development
U.S. Environmental Protection Agency
Mail Code E343-06
Research Triangle Park, NC 27711
-------
ERRATA Sheet
For the document: Decontamination of Toxic Industrial Chemicals and Chemical Warfare Agents
On Building Materials Using Chlorine Dioxide Fumigant and Liquid Oxidant Technologies
(EPA/600/R-09/012) February 2009
Updated 07/2009 with the following modifications:
• Table 4-13 (Page 27): The numbers in the rightmost column were made bold to correspond with
footnote c [Decontamination efficacy shown in bold indicates a statistically significant difference
in recovery with and without decontamination (p < 0.05)].
• Table 4-19 (Page 34): The numbers in the rightmost column were made bold to correspond
with footnote a [Calculated using Eq (8); values given in bold are statistically significant].
• Page 31 (Left Column, 2nd paragraph): 100% bleach (6% hypochlorite) was modified to read
10% bleach (0.6% hypochlorite).
-------
Notice
The U.S. Environmental Protection Agency (EPA), through the Office of Research and
Development's (ORD) National Homeland Security Research Center (NHSRC), funded and
managed this technology investigation through a Blanket Purchase Agreement under General
Services Administration contract number GS23F0011L-3 with Battelle. This report has been peer
and administratively reviewed and has been approved for publication as an EPA document. Mention
of trade names or commercial products does not constitute endorsement or recommendation for use
of a specific product.
-------
Foreword
The EPA is charged by Congress with protecting the nation's air, water, and land resources. Under
a mandate of national environmental laws, the Agency strives to formulate and implement actions
leading to a compatible balance between human activities and the ability of natural systems to support
and nurture life. To meet this mandate, ORD provides data and science support that can be used
to solve environmental problems and to build the scientific knowledge base needed to manage our
ecological resources wisely, to understand how pollutants affect our health, and to prevent or reduce
environmental risks.
In September 2002, EPA announced the formation of the NHSRC. The NHSRC is part of the ORD; it
manages, coordinates, and supports a variety of research and technical assistance efforts. These efforts
are designed to provide appropriate, affordable, effective, and validated technologies and methods for
addressing risks posed by chemical, biological, and radiological terrorist attacks. Research focuses on
enhancing our ability to detect, contain, and clean up in the event of such attacks.
NHSRC's team of world renowned scientists and engineers is dedicated to understanding the terrorist
threat, communicating the risks, and mitigating the results of attacks. Guided by the roadmap set forth
in EPA's Strategic Plan for Homeland Security, NHSRC ensures rapid production and distribution of
security-related products.
The NHSRC has created the Technology Testing and Evaluation Program (TTEP) in an effort to
provide reliable information regarding the performance of homeland security related technologies.
TTEP provides independent, quality assured performance information that is useful to decision makers
in purchasing or applying the tested technologies. It provides potential users with unbiased, third-
party information that can supplement vendor-provided information. Stakeholder involvement ensures
that user needs and perspectives are incorporated into the test design so that useful performance
information is produced for each of the tested technologies. The technology categories of interest
include detection and monitoring, water treatment, air purification, decontamination, and computer
modeling tools for use by those responsible for protecting buildings, drinking water supplies and
infrastructure, and for decontaminating structures and the outdoor environment.
The investigation reported herein was conducted by Battelle, under the direction of NHSRC, as
part of TTEP. Information on NHSRC and TTEP can be found at http://www.epa. gov/nhsrc.
-------
Acknowledgments
The authors wish to acknowledge the support of all those who helped plan and conduct the
investigation, analyze the data, and prepare this report. We also would like to thank Lawrence
Kaelin (EPA National Decontamination Team), David Mickunas (EPA Environmental Response
Team), Donna Getty (Lockheed-Martin), Emily Snyder (EPA/ORD/NHSRC), and Joseph Wood
(EPA/ORD/NHSRC) for reviewing this report.
-------
Contents
Notice v
Foreword vi
Acknowledgments vii
Abbreviations/Acronyms xiii
Executive Summary xiv
1.0 Introduction 1
1.1 Objectives 1
1.2 Approach 1
1.3 Experimental Design 2
1.4 Definitions of Coupon Treatments 3
2.0 Methods 5
2.1 ClO2Fumigant Decontamination Technology Used with TICs and CWAs 5
2.1.1 Sabre Technical Services C1O2 Generator 5
2.2.2 Measurement in the Test Chamber Atmosphere 5
2.2 Liquid Decontamination Technologies Used with CWAs 5
2.2.1 Liquid Decontamination Solutions Tested 5
2.2.2 Liquid Decontamination Test System 6
2.3 TICs 6
2.3.1 Test Chamber for Fumigant Decontamination Tests 6
2.3.2 Building Materials 7
2.3.3 Sequence of Testing 7
2.3.4 Source of TICs 7
2.3.5 Application of TICs to Test Coupons 8
2.3.6 Extraction Method for TICs from Test Coupons 8
2.3.7 Air Sampling and Analysis of TICs Using Tenax® Sorbent Air Sampling Tubes 9
2.3.8 Analysis Method for TICs 9
2.3.9 Calculation of TIC Recovery Efficiency, Percent Recovery, and
Decontamination Efficacy 10
2.4 CWAs 11
2.4.1 Test Chamber for Fumigant Decontamination Tests 11
2.4.2 Test Chamber for Liquid Decontamination Tests 12
2.4.3 Building Materials 12
2.4.4 Sequence of Testing 12
2.4.5 CWAs and Surrogate Recovery Standard 13
2.4.6 Application of CWAs to Test Coupons 13
2.4.7 Extraction Method for CWAs 14
2.4.8 Analysis Method for CWAs 14
2.4.9 Calculation of CWA Recovery Efficiency, Percent Recovery, and
Decontamination Efficacy 14
2.5 Qualitative Evaluation of the Impact of CIO Fumigation on Building Materials 14
-------
3.0 Quality Assurance/Quality Control 15
S.lPEAudit 15
3.2 Technical Systems Audit 15
3.3 Data Quality Audit 16
3.4 QA/QC Reporting 16
3.5 Deviations from Test/QAPlan 16
4.0 Results 17
4.1 Results for Fumigant C1O2 Decontamination of TICs 17
4.1.1 Analytical Method Development Results 17
4.1.2 Environmental Conditions during Decontamination Tests 17
4.1.3 Recovery over Time of TICs on Building Materials With and Without Sabre
C1O2 Fumigant Decontamination 18
4.1.4 Statistical Analysis of Recovery Trends and Decontamination Efficacy 19
4.1.5 TICs on Laboratory, Handling, and Procedural Blank Coupons 22
4.1.6 TICs in Chamber Air 23
4.1.7 Detection of Oxidized Malathion Product on Coupons 23
4.1.8 Condition of ClO2-Treated Coupons after Six Months 24
4.2 Results for Fumigant C1O2 Decontamination of CWAs 24
4.2.1 Analytical Method Results 24
4.2.2 Recovery over Time of CWAs on Building Materials With and Without Sabre
C1O2 Fumigant Decontamination 24
4.2.3 Statistical Analysis of Recovery Trends and Fumigant Decontamination Efficacy 26
4.2.4 CWAs on Laboratory and Procedural Blank Coupons 30
4.2.5 Concentration of C1O2 in Test Chambers 30
4.3 Results for Liquid Decontamination of CWAs 31
4.3.1 Recovery of CWAs from Liquid Decontamination Solutions 31
4.3.2 Recovery over Time of CWAs on Building Materials With and Without
Liquid Bleach Decontamination 31
4.3.3 Statistical Analysis of Recovery Trends and Liquid Decontamination Efficacy 34
4.3.4 CWAs on Laboratory and Procedural Blank Coupons During Liquid
Decontamination Tests 35
5.0 Summary 37
6.0 References 39
7.0 Appendix 41
-------
Figures
Figure 2-1. Frontal (left) and Overhead (right) Views of Test Chamber Used for
Decontamination Test of TICs 7
Figure 4-1. Statistically Modeled Percent Recovery Data for DMMP on Coupons
With and Without Sabre C1O2 Fumigant Decontamination 21
Figure 4-2. Statistically Modeled Percent Recovery Data for Malathion on Coupons
With and Without Sabre C1O2 Fumigant Decontamination 21
Figure 4-3. Percent Oxidation of Malathion to Maloxon by Coupon Type and C1O2 Dosage 24
Figure 4-4. GB Recovery (%) on Coupons With and Without Sabre C1O2 Fumigant
Decontamination 28
Figure 4-5. TGD Recovery (%) on Coupons With and Without Sabre C1O2 Fumigant
Decontamination 28
Figure 4-6. VX Recovery (%) on Coupons With and Without Sabre C1O2 Fumigant
Decontamination 29
Figure 4-7. Statistical Analysis Results of VX Decontamination with Liquid CIO 35
-------
Tables
Table ES-1. Mean Percent Recovery of TICs and CWAs on Building Materials With and Without .
C1O2 Fumigation (Normalized to Recovery at Time 0) xiv
Table ES-2. Percent Recovery of CWAs Spiked Directly into Decontamination Solutions
With and Without Neutralization xiv
Table ES-3. Percent Recovery of CWAs from Building Materials Following
Various Treatments xv
Table 1-1. Building Materials Used in Decontamination Investigation 2
Table 1-2. Target Parameters for Decontamination Testing 3
Table 1-3. Coupon Treatments 3
Table 1 -4. Selected TICs and CWAs with Analytical Measurement Parameters 4
Table 2-1. Types of Coupons Used in the TIC Fumigation Investigation 7
Table 2-2. Source of TICs and PE Audit Standards 8
Table 2-3. Solvent Evaporation Times for TICs Spiked on Various Building Materials 8
Table 2-4. Extraction and Concentration Techniques Used for TICs 8
Table 2-5. GC and MS Conditions for Analysis of DMMP in Air Samples 9
Table 2-6. GC and MS Conditions for TIC Analyses 9
Table 2-7. GC Retention Times and Monitored Ions for TIC Analyses 10
Table 2-8. Building Material Test Coupon Characteristics for CWA Decontamination Tests 12
Table 2-9. Source of CWAs and SRS 13
Table 2-10. Mass of CWA Applied to Building Material Coupons 13
Table 2-11. GC/FPD Conditions for CWA Analyses 14
Table 2-12. GC Retention Times for CWA Analyses 14
Table 3-1. TIC PE Audit Results 15
Table 3-2. CWA PE Audit Results 15
Table 4-1. Extraction Efficiencies of TICs and Matched SRSs from Building Materials 17
Table 4-2. MDLs for TICs 17
Table 4-3. Mean Temperature, RH, C1O2 Concentration, and Air Velocity during TIC
Decontamination Tests 18
Table 4-4. Comparison of Mean Percent Recovery of DMMP with Different Chamber
Air Exchange Rates 18
Table 4-5. Mean Recovery of TICs from Building Materials over Time With and
Without Sabre C1O2 Fumigant Decontamination 19
Table 4-6. Statistically Modeled, Eq (11), Percent Recovery of TICs With and
Without Sabre C1O2 Fumigant Decontamination and Decontamination Efficacy 20
Table 4-7. Mean TIC Levels on Laboratory, Handling, and Procedural Blank Coupons 22
Table 4-8. Mean Concentration of TICs in Test Chamber Air 23
Table 4-9. Mean Mass of Malathion Oxidized to Maloxonby Coupon Type and C1O2 CT 23
Table 4-10. Recovery of CWAs and SRS from Building Materials 25
Table 4-11.MDLs for CWAs 25
Table 4-12. Average Recovery of CWAs from Building Materials over Time With and
Without Sabre CIO Fumigant Decontamination 25
-------
Table 4-13. Statistically Modeled, Eq (11), Percent Recovery of CWAs With and
Without Sabre C1O2 Fumigant Decontamination and Decontamination Efficacy 27
Table 4-14. Comparison of Mean CWA Levels on Laboratory and Procedural Blank Coupons 30
Table 4-15. Mean Concentration of C1O2 in Test Chamber for CWA Decontamination Tests 30
Table 4-16. Mean Recovery of CWAs and SRS from Liquid Decontamination Solutions 31
Table 4-17. Mean Recovery of CWAs from Building Materials After Various
Treatments as Percent of Mass Applied 32
Table 4-18. Mean Recovery of CWAs from Building Materials After Various
Treatments as Percent of TO Recovery 33
Table 4-19. Statistically Modeled Percent, Eq (11), Recovery of VX With and Without
Liquid C1O2 Decontamination and Decontamination Efficacy
(Compared to Recovery from Acidified Water) 34
Table 4-20. CWA Levels on Laboratory and Procedural Blank Coupons with Bleach
Decontamination 35
Table 4-21. VX Levels onProcedural Blank Coupons withLiquid C1O2Decontamination 36
Table 5-1. Summarization of Percent Recovery With and Without Decontamination
Technologies, and Decontamination Efficacy for TICs and CWAs 37
Table 5-2. Mean Recovery of CWAs Directly Spiked into Liquid Decontamination Solutions 38
Table A-l. Normalized Mean Recovery of TICs on Building Materials over Time With
and Without Sabre C1O2 Fumigant Decontamination 41
Table A-2. Normalized Mean Recovery of CWAs on Building Materials over Time With
and Without Sabre C1O2 Fumigant Decontamination 42
Table A-3. Normalized Mean Recovery of VX on Building Materials With and
Without Liquid CIO Decontamination 42
-------
Abbreviations/Acronyms
ANOVA analysis of variance N
BBRC Battelle Biomedical Research Center ND
CIO" hypochlorite ion NT
C1O2 chlorine dioxide NHSRC
cm centimeter(s) NIST
CWA(s) chemical warfare agent(s) ORD
CT concentration x contact time PE
DEEP diethyl ethylphosphonate ppm
DIMP diisopropyl methylphosphonate QA
DMMP dimethyl methylphosphonate QC
EPA U.S. Environmental Protection Agency QMP
Eq equation RH
FPD flame photometric detection SD
ft feet sec
g gram(s) SRS(s)
GB sarin STS
GC gas chromatograph; gas chromatography TBP
GC/FPD gas chromatography/flame photometric detection TGD
GC/MS gas chromatography/mass spectrometry TIC(s)
GD soman TD
h hour(s) TSA
IS internal standard TTEP
KD Kuderna-Danish
HVAC heating, ventilation, and air conditioning
L liter(s)
m meter(s)
M molarity; moles/liter
MDL method detection limit
MFC mass flow controller
ug microgram(s)
uL microliter(s)
um micrometer(s)
m meter(s)
mg milligram(s)
min minute(s)
mL milliliter(s)
mm millimeter(s)
MS mass spectrometer
m/z mass-to-charge ratio (unitless)
n number of samples
normal
not detected
not tested
National Homeland Security Research Center
National Institute of Standards and Technology
Office of Research and Development
performance evaluation
parts per million
quality assurance
quality control
quality management plan
relative humidity
standard deviation
second(s)
surrogate recovery standard(s)
sodium thiosulfate
tributyl phosphate
thickened soman
toxic industrial chemical(s)
thermal desorption
technical systems audit
Technology Testing and Evaluation Program
-------
Executive Summary
The purpose of this effort was to determine the effectiveness
of several decontamination technologies against both
toxic industrial chemicals (TICs) and chemical warfare
agents (CWAs) on standard building materials was
recently completed. Prior to the test design for this effort,
an assessment of the state of the knowledge for TIC and
CWA decontamination on complex building materials
was performed. Following this assessment, combinations
of TICs, CWAs, building materials, decontamination
technologies and conditions were selected for the study. The
first phase of the testing included a study of the persistence
of the TICs and CWAs on the materials at typical ambient
conditions1'1. Decontamination studies were only performed
with combinations of agents and materials in which
persistence was determined sufficient in order to observe
an impact of the decontamination treatment. The TICs that
were used in this testing included malathion and dimethyl
methylphosphonate (DMMP). These TICs were chosen for a
combination of two reason: (1) for their properties as TICs,
and (2) their potential as simulants for VX and sarin (GB),
respectively, due to the similarities in key structural elements.
The CWAs that were tested included sarin, thickened soman
(TGD), and VX. The technologies investigated included
fumigation with chlorine dioxide (C1O2) and soaking in liquid
oxidants.
Fumigation at 80% relative humidity (RH) and 3000 parts
per million (ppm) C1O2 used a commercially available system
manufactured by Sabre, Inc. This Sabre technology was
tested against:
• Malathion on carpet and laminate materials
• DMMP on carpet and ceiling tile
• GB on carpet
• TGD on carpet, laminate, and metal ductwork
• VX on carpet, laminate, and metal ductwork.
The liquid oxidants included diluted commercial bleach
(diluted 1:10 with water; 5000 ppm hypochlorite ion [CIO])
and an aqueous solution of 3000 ppm C1O2. Aqueous bleach
was tested against:
• GB on carpet
• TGD on carpet, laminate, and metal ductwork
• VX on carpet, laminate, and metal ductwork.
The aqueous C1O2 solution was tested against VX on carpet,
laminate, and metal ductwork.
This decontamination research addressed the following
questions:
• What is the percent recovery of TICs and CWAs exposed
to the specified treatments?
• What is the relative recovery (decontamination efficacy)
of TICs and CWAs in the presence of the decontamination
technology compared to a similarly matched control?
• Is there visible surface damage to building materials as a
result of the decontamination technology?
The above matrix of test agent and building material coupons
reflects the results of the persistence study that preceded
this decontamination study. In that prior testing, persistence
was found to vary widely, from as low as 0% GB recovered
from laminate after five minutes (min) up to 85% malathion
recovered from carpet after seven days.[1] While there was
evidence for complex interactions between test compound
and substrate, persistence could mostly be predicted and
explained on the basis of vapor pressure and hydrolysis rate
of the compounds, and the type of the building material.
Hence, only combinations of chemicals and materials in
which the agent was persistent enough to determine the effect
of the decontamination were used, and some substitutions
in materials were made in some cases to provide longer
persistence (e.g., for DMMP, ceiling tile was used rather than
laminate).
Very similar bench-scale testing approaches were used for
these technology investigations as were utilized for the
persistence testing. This included liquid spiking of a known
amount of an individual TIC or CWA onto replicate coupons
of the building material (five identical coupons each either
5 square centimeters [cm2] for TICs or 10 cm2 for CWAs
were utilized). These coupons were allowed to contact the
decontamination technology for a fixed amount of time up to
7 hours (h) in a sealed chamber under controlled conditions
of temperature, RH, and air movement. Each coupon was
then extracted to measure the amount of TIC or CWA that
remained on the coupons. These tests used spikes of 500
micrograms (ug) of TICs onto 5 cm2 coupons, or 1000 ug of
CWAs onto 10 cm2 coupons, so as to approximate a surface
loading of 1 g nr2, which is assumed to be a worst-case
contamination scenario.
The persistence of these chemicals under conventional
environmental conditions is highly variable, and in some
cases the chemicals have limited persistence. For these
reasons, the fumigant decontamination tests were matched
with positive control tests of the chemical on the same type
of building material coupon for the same duration and at the
same temperature as the decontamination test, so that the
effectiveness of the decontamination technology could be
accurately differentiated from other loss mechanisms such as
volatilization and decomposition unrelated to the fumigation.
Fumigation with C1O2 resulted in statistically significant
efficacy against malathion on both carpet and laminate at
1,3, and 7 h exposures; a small, but statistically significant
efficacy (7.5%) was observed against DMMP only on ceiling
tile at 7 h. Fumigation with C1O2 resulted in statistically
significant efficacy >99% against VX on carpet, laminate,
and ductwork after a Ih exposure; a smaller statistically
significant efficacy (34% - 62%) was observed against GB
and TGD on carpet after a 1 h exposure. No statistically
-------
significant efficacy of C1O2 fumigation was observed against
TGB on laminate or ductwork.
As shown in Tables ES-1 and ES-2, VX was reduced to
non-detectable levels on all materials with all three of
the decontamination technologies tested. When VX was
exposed to high RH only (as a control for the fumigant
C1O2 decontamination test), the mean recovery was high.
indicating that the vapor phase C1O2 itself was effective at
decontaminating the VX.
Liquid technologies were screened for efficacy in preliminary
testing in which the CWA was placed into decontamination
technology and neutralized decontamination technology.
The CWA was extracted from the solution and analyzed
to determine the mass of CWA in the extract. The results
of the tests, shown in Table ES-2, demonstrated that.
except for GD in bleach, the difference in mean recovery
between the decontamination technology and neutralized
decontamination technology was negligible. VX, which
is persistent on building materials exposed to air, was not
recovered from neutralized bleach or neutralized C1O2
solutions when extracted within 15 seconds (sec); causes
were not determined but could include hydrolysis of the
VX or an ineffective recovery method. In bleach, none of
the three CWAs were recovered. Only small amounts of GB
(1%) were recovered after 15 sec in neutralized bleach. There
was little or no difference between the mean recovery of any
CWA after 15 sec in neutralized C1O2 or after 1 h exposed to
C1O2 decontamination; a large percentage (58%) of GD was
recovered from the C1O2 solution.
Table ES-3 summarizes the mean percent recovery of
the TICs and CWAs on different building materials
following liquid decontamination treatments. The bleach
decontamination tests were matched with positive controls
exposed to air. The positive controls used with liquid bleach
were placed in empty vials. The results in Table ES-3 show
that in all cases, when compared to untreated coupons.
the CWA recovered from coupons soaked in bleach were
significantly lower and in most cases not detected, after
Table ES-1. Mean Percent Recovery of TICs and CWAs on Building Materials
With and Without CI02 Fumigation (Normalized to Recovery at Time 0)
TIC/CWA
DMMP
Malathion
GB
TGD
VX
Material
Carpet
Ceiling tile
Carpet
Laminate
Carpet
Carpet
Laminate
Ductwork
Carpet
Laminate
Ductwork
Time (n)
7 h (n = 5)
7 h (n = 5)
7 h (n = 5)
7 h (n = 5)
4 h (n = 5)
2 h (n = 5)
2 h (n = 5)
2 h (n = 5)
1 h (n = 5)
1 h (n = 5)
1 h (n = 5)
Mean Recovery, % of TO ± SD
Without With CI02
Decontamination Fumigation
16±3
12 ± 1
87 ± 1
94 ± 10
5. 7 ±0.2
57 ±6
1.3 ± 1.6
5.9 ±2. 6
105 ±31
86 ±8
101±3
23 ±4
8.7 ±0.3
24 ±2
0.4 ±0.2
3.3 ± 1.0
36 ± 17
0.10±0.11
7.4 ±9.0
ND, <0.7a
ND, <0.7
ND, <0.7
a) ND = not detected; method detection limit converted to equivalent recovery value
Table ES-2. Mean Percent Recovery of CWAs Spiked Directly
into Decontamination Solutions With and Without Neutralization
Solution - Hold Time Before SRS
Addition and Hexane Extraction
Bleach - 1 h
Neutralized bleach - 15 sec
CI02- 1 h
Neutralized CI02 - 15 sec
CWA Mean Recovery,
GB
ND, <0.1
1 ±0
8±2
8± 1
GD
ND, <0.1
41 ± 10
58 ±5
56 ± 10
VX
ND, <0.7
ND, <0.7
ND, <0.7
ND, <0.7
ND = not detected
-------
exposures of 10 to 30 min. The positive controls used
with liquid bleach were placed in empty vials, so that
apparent decontamination with this solution could not be
distinguished from other potential loss processes such as
aqueous hydrolysis. The results shown in Tables ES-2 and
ES-3 suggest that bleach is effective for decontamination
of GD/TGD. However, the low recoveries of GB and no
recovery of VX from the neutralized bleach require caution
in interpreting these results; based on these data, apparent
effectiveness cannot be distinguished from other explanations
of low recoveries such as aqueous hydrolysis or ineffective
recovery methods. Further research is needed to isolate the
cause.
The tests with liquid C1O2 decontamination of VX (which is
unstable in alkaline solutions) were matched with positive
controls in acidified water to attempt to reduce the hydrolysis
effects of water for comparison with the decontamination
effects of C1O2. For VX on carpet and laminate, a statistically
significant reduction of 86% and 63% was achieved after
10 min exposure to liquid C1O2; efficacy against VX on
ductwork was indeterminate.
The C1O2 fumigation was shown to be highly effective
against VX. In some but not all cases, dependent on the type
of coupon, C1O2 resulted in statistically significant reductions
in GB and TGD.
The TGD levels on the three materials were reduced to very
low levels (<1% of the initial mass applied to the coupon)
using liquid bleach; fumigant C1O2 was not as effective as
liquid bleach in attacking TGD on carpet and ductwork.
Liquid bleach was effective in removing TGD from the
surfaces of the building materials. There was no attempt to
measure TGD in the bleach solution. TGD may have been
hydrolyzed by the aqueous bleach or effectively solubilized
by it.
Similarly, GB was effectively removed from the carpet by
liquid bleach; however, the investigation did not clarify
whether this removal was due to volatilization, hydrolysis.
solubilization, or a combination of these processes. No
decontamination efficacy was observed for liquid C1O2
against GB or GD in the solution testing; efficacy against VX
was indeterminate.
No damage or visible change to any of the materials was
observed comparing extracted laboratory blank coupons
(not exposed to decontamination) to extracted procedural
blank coupons (exposed to decontamination) directly after
decontamination treatment. Materials exposed to the liquid
decontaminants were not evaluated at subsequent time
post-decontamination. For C1O2 fumigation, the coupons
were relatively unchanged after three months, and the carpet
showed some very minor "bleaching" after six months.
This work demonstrated that there are very simple and
effective methods for removing toxic chemicals from the
surfaces of building materials. In some cases the results
from this investigation could not be used to establish the
mechanisms and processes involved in removal of the
chemicals from the surfaces because measurement of the
residual chemical on the building material surface did not
indicate whether the chemical was transferred intact to the
aqueous decontamination solution or whether it was degraded
or hydrolyzed to another chemical. These types of questions
may best be answered by direct analyses (e.g., no solvent
extraction) of the decontamination solutions to verify whether
intact toxic chemicals or degradation products are present.
Insofar as this program was designed to evaluate residuals
on surfaces after decontamination treatments, the goals were
achieved, and reliable technologies for decontamination were
identified.
Table ES-3. Mean Percent Recovery of CWAs from Building Materials
Following Various Treatments
GB
Carpet
Mean Recovery, % of Mass Recovered at Time 0 ± SD
10 min
Without Decontamination
(in air in a sealed vial)
93 ±7
With Bleach
Decontamination
ND,
-------
1.0
Introduction
The EPA's NHSRC is helping to protect human health and the
environment from adverse impacts resulting from intentional
acts of terror. With an emphasis on decontamination and
consequence management, water infrastructure protection,
and threat and consequence assessment, NHRSC is working
to develop tools and information that will help detect
the intentional introduction of chemical, biological or
radiological contaminants in buildings or water systems,
contain these contaminants, decontaminate buildings and/or
water systems, and facilitate the disposal of material resulting
from cleanups.
NHSRC's TTEP works in partnership with recognized testing
organizations; with stakeholder groups consisting of buyers,
vendor organizations, and permitters; and with the voluntary
participation of individual technology developers in carrying
out performance tests on homeland security technologies.
The program evaluates the performance of innovative
homeland security technologies by developing test plans that
are responsive to the needs of stakeholders, conducting tests,
collecting and analyzing data, and preparing peer-reviewed
reports. All investigations are conducted in accordance with
rigorous quality assurance (QA) protocols to ensure that data
of known and high quality are generated and that the results
are defensible. TTEP provides high-quality information that
is useful to decision makers in purchasing or applying the
tested technologies. It provides potential users with unbiased,
third-party information that can supplement vendor-provided
information. Stakeholder involvement ensures that user needs
and perspectives are incorporated into the test design so that
useful performance information is produced for each of the
tested technologies.
1.1 Objectives
This testing was conducted to measure the effectiveness
of different decontamination technologies against two
representative TICs and three representative CWAs on a
range of indoor building materials. The recovery (extractable
mass) of the chemicals with and without the decontamination
agent was ascertained initially. The effectiveness of the
decontamination technology was assessed as the statistically
significant relative recovery of TIC or CWA from building
material coupons after a specified time period with or without
contact with the decontamination technology. This approach
controls for losses due to normal environmental processes
such as volatilization.
This decontamination research addressed the following
questions:
• What is the relative recovery of TICs and CWAs in the
presence of the decontamination technology compared to
a similarly matched control?
• What is the percent reduction of TICs and CWAs exposed
to various treatments?
• What is the decontamination efficacy of the
decontamination technology for removal of TICs and
CWAs?
• Is there visible surface damage to building materials as a
result of the decontamination technology?
1.2 Approach
The general approach developed and utilized for
decontamination testing was to apply a known amount of
each TIC or CWA to each of several test coupons of the
same building material (replicate coupons, identical in size
and shape) and allow these spiked test coupons to age under
controlled environmental conditions of temperature and RH,
either with the decontamination agent or without this agent.
At specified intervals, replicate test coupons were extracted
and the extracts were analyzed to determine the amount of
the TIC or CWA that remained on the test coupon at that
specific time.
The overall approach developed and applied for
decontamination testing of fumigant and liquid
decontamination technologies was generally the same;
however, the fumigant technology testing was conducted
in a sealed chamber whereas the liquid decontamination
was conducted in individual sealed vials. The fumigant
technology was applied to both TICs and CWAs; the
liquid decontamination was applied to only the CWAs.
The specific details for the methodologies used for
decontamination testing of TICs and CWAs are
described in detail in Section 2.0.
-------
1.3 Experimental Design
The chemicals that were selected for use in this investigation
include:
TICs
• Malathion
• DMMP
CWAs
• GB
• TGD
• VX.
Table 1-1 specifies the building materials used in this
decontamination investigation which included industrial
grade carpet, decorative laminate, galvanized metal
ductwork, and ceiling tile. Building materials were cut into
coupons of small, defined size and the TICs and CWAs were
applied at a rate equivalent to 1 g nr2, which is representative
of a potential indoor contamination scenario. The coupons to
which TICs were applied were approximately 5 cm2 and the
coupons to which CWAs were applied were approximately
10 cm2. The sizes were chosen so as to take advantage of
the available area in the test chambers and to optimize the
spiking volume of chemical being applied to the coupons.
All decontamination testing with TICs was carried out
in standard chemical laboratories at Battelle. Due to
the stringent controls needed for working with CWAs.
decontamination tests for CWAs were carried out at one
of Battelle's certified chemical surety facilities (Battelle
Biomedical Research Center [BBRC]) near West Jefferson.
Ohio. Special test chamber equipment and protocols were
prepared and utilized for conducting this investigation. For
fumigant decontamination and its associated controls, tests
were conducted with coupons inside specially fabricated test
chambers which allowed for controls on temperature, RH, air
flow over the coupons, and air exchange rate in the chamber.
In these tests, the decontamination of each chemical (TIC or
CWA) was investigated separately; however, the behavior of
a given chemical was investigated on all building material
types simultaneously. For decontamination tests using
liquids, the test coupons and associated controls were
placed in sealed vials containing the decontamination liquid.
Table 1-2 (on page 3) presents a summary of the matrix of
building materials and chemicals, together with the target test
chamber conditions and fumigant and liquid decontamination
technologies. The materials and chemicals selected for testing
were based on the previous persistence study and preliminary
solution testing of liquid decontamination technologies.
Only the material/chemical combinations with chemical
persistence sufficient to determine the decontamination
effects were included in the matrix. For DMMP, ceiling tile
was used rather than laminate because the persistence study
showed less than 1% of the spiked DMMP was recovered
from laminate after one day. Similarly, decontamination
of GB was evaluated only on carpet because GB was not
detected on laminate and ductwork in as little as 15 min after
spiking. GB and TGD were not evaluated with liquid C1O2
because preliminary solution testing showed no efficacy
compared to the control solution. Therefore, only VX on
coupons was investigated with the liquid C1O2; spiked
coupons were placed in water, acidified with acetic acid to
reduce hydrolysis, as a control.
The temperature and RH inside the test chambers were
monitored and recorded. There were five replicate coupons
of each building material type analyzed at each time point.
for each TIC or CWA. For TICs testing only, air velocity
over the coupons was measured. To ensure that air was
passed uniformly across the TIC coupons, two fans were
used that produced an air velocity of 400 ft min'1 over the
coupons. A smaller test chamber was used in the investigation
of fumigation of CWAs. Uniform gas distribution was not
anticipated to be a problem, so fans were not used in this
test chamber.
Table 1-1. Building Materials Used in Decontamination Investigation
Material
Decorative laminate
Industrial grade
carpet
Galvanized metal
ductwork
Ceiling tile
Lot, Batch, or Observation ^f^
Laminate/Formica/White Matte Finish
EcoTek 6; Style #M7978, color #910;
Carpet Corp of America
Industry HVAC standard 24 gauge
galvanized steel; Adept Products Inc
Armstrong 954, Classic Fine Textured
Solid Surface Design
Shaw Industries, Inc
Accurate Fabrication
Armstrong
-------
1.4 Definitions of Coupon Treatments
The types of test and control coupons used in this
investigation are described in Table 1-3. Test coupons and
positive control coupons are spiked with TICs or CWAs.
Laboratory, handling, and procedural blank coupons are
not spiked with TICs or CWAs. Test and procedural blank
coupons are exposed to the decontamination treatment.
Positive control coupons and laboratory and handling blanks
are not exposed to the decontamination treatment.
Table 1-4 (on page 4) shows surrogate recovery standard
(SRS) and internal standard (IS) compounds that were
utilized in the quantitative chemical analyses (See Section
2.0). This table summarizes the analysis method for extracts
of building materials and the sampling and analysis methods
employed in measuring the chemicals in the air over the
building materials during decontamination.
Table 1-2. Target Parameters for Decontamination Testing
Chemical Building Temperature Decontamination Air Flow Sampling Points
Materials and RHa Agent Over Coupons in Time
TIC
Malathion
DMMP
Carpet
Laminate
Carpet
Ceiling Tile
24°C,
80% RH
24°C,
80% RH
Fumigant CI02
at 3000 ppm
Fumigant CI02
at 3000 ppm
400ft min1
400ft min1
0, 1, 3, 7 h
0, 1,3, 7 h
CWA
GB
GB
TGD
TGD
VX
VX
VX
Carpet
Carpet
Carpet
Laminate
Ductwork
Carpet
Laminate
Ductwork
Carpet
Laminate
Ductwork
Carpet
Laminate
Ductwork
Carpet
Laminate
Ductwork
22°C,
80% RH
22°C,
NA
22°C,
80% RH
22°C,
NA
22°C,
80% RH
22°C,
NA
22°C,
NA
Fumigant CI02
at 3000 ppm
Liquid bleachb
at 5000 ppm CIO
Fumigant CI02
at 3000 ppm
Liquid bleach
at 5000 ppm CIO
Fumigant CI02
at 3000 ppm
Liquid bleach
at 5000 ppm CIO
Liquid CI02
at 3000 ppm
Oft min1
Oft min1
Oft min1
Oft min1
Oft min1
Oft min1
Oft min1
0, l,4h
0, 10, 20, 30 min
0, 1, 2 h
0, 10, 20, 30 min
0, l,4h
0, 10, 20, 30 min
0, 10, 20, 30 min
a) RH is not applicable (NA) to testing in which liquid decontamination technologies is used
b) Commercial bleach diluted 1:10 with water
Table 1-3. Coupon Treatments
Positive control
coupon
Test coupon
Laboratory blank
Handling blank
Procedural blank
Building material coupon spiked with TIC or CWA that is not exposed to
the decontamination treatment; analyzed together with test coupons at
the designated time interval
Building material coupon spiked with TIC or CWA and exposed to the
decontamination treatment for the designated time interval
Building material coupon that is loaded into an extraction vial before all
other test activities
Building material coupon that is exposed to the fume hood atmosphere
during sample spiking, and then loaded into an extraction vial
Building material coupon, with no TIC or CWA spike, that is exposed to
the decontamination or control treatment for the designated time interval
-------
Table 1-4. Selected TICs and CWAs with Analytical Measurement Parameters
SRS
IS
Extraction
Analysis
Air sample collection
Air sample analysis
Fenchlorphos
DBBC
Sonication
GC/MS6
Tenax® sorbent
Extract & C/MS
DEEP3
DIMPd
Sonication
GC/MS
Tenax® sorbent
TD-GC/MSs
TBPb
DIMP
Shake/stand
GC/FPD'
Not collected
Not applicable
TBP
DIMP
Shake/stand
GC/FPD
Not collected
Not applicable
TBP
DIMP
Shake/stand
GC/FPD
Not collected
Not applicable
a) DEEP - diethyl ethylphosphonate
b) TBP - tributyl phosphate
c) DBB - dibromobiphenyl
d) DIMP - diisopropyl methylphosphonate
e) GC/MS - gas chromatography/mass spectrometry in the multiple ion detection mode
f) GC/FPD - gas chromatography/flame photometric detection
g) TD-GC/MS - thermal desorption GC/MS of analytes from Tenax®
-------
2.0
Methods
2.1 CI02 Fumigant Decontamination
Technology Used with TICs and CWAs
2.1.1 Sabre Technical Services CI02 Generator
The Sabre C1O2 generator consisted of a 20.3 cm x 20.3 cm
base onto which a 15.2 cm x 15.2 cm, 91.4 cm high sparging
column was mounted. A 19 L container with 15 L of the
C1O2 decontamination solution was placed at the base of the
sparging column. The C1O2 decontamination solution (an
aqueous solution consisting of 3 g L'1 of C1O2 plus 1000 ppm
of sodium chlorite) was prepared just prior to use in each test.
Using a peristaltic pump, the C1O2 decontamination solution
was pumped into the sparging column. Air from the test
chamber was also pumped into and through this column and
this air was used to sparge the C1O2 solution so as to form
a fine mist; this air stream re-entered the test chamber with
C1O2 at the desired concentration. Variation in the flow rate of
liquid and air into the sparging chamber was used to establish
the desired C1O2 concentration in the test chamber. Liquid
from the reservoir of C1O2 decontamination solution was
initially introduced into the sparging column at 60 mL min"1.
When the desired C1O2 concentration in the test chamber was
achieved, the liquid introduction into the sparging column
was decreased to 0 mL min"1 until the concentration dropped;
at that time the system was restarted at brief intervals to
maintain the desired concentration. The spent liquid exiting
the sparging column was collected in a reservoir. The air
from the chamber was recirculated into and out of the
sparging column. At the end of the decontamination test the
C1O2 in the system was destroyed by pumping the exhaust
air through a scrubber containing a sodium thiosulfate (STS)
trap. The test chambers are described in greater detail in
Sections 2.3.land 2.4.1.
2.1.2 Measurement in the Test Chamber Atmosphere
The concentration of C1O2 in the test chamber was monitored
before beginning a test and periodically during the test
using a titration method. Use of this method assumes, as has
previously been demonstrated using the Sabre generation
method, that C1O2 is the only chlorine species in the gas
being sampled. For this titration method, air in the chamber
was sampled using an impinger (at a rate of 1 L min"1
for 2 min) containing 15 mL of 5% potassium iodide in
phosphate buffer (pH 7.0). Under these conditions C1O2
oxidizes iodide in the phosphate buffer solution to iodine
and C1O2 gas is reduced to the chlorite ion (C1O2~) which
dissolves in solution. The molecular iodine that is produced
appears yellow-brown in aqueous solution. After sampling
the chamber atmosphere, the impinger solution is acidified
and the chlorite is allowed to react further with the iodide
ion, forming additional iodine and reducing the chlorite to
chloride. The total resulting iodine is then reduced to iodide
by titrating with standard 0.1 M (molarity, moles per liter)
STS. The titration endpoint is determined when the color of
the solution changes from yellow-brown to colorless. The
volume (mL) of STS needed to achieve the color change
(brown to colorless) is proportional to the amount of iodine
generated, which is proportional to the gas-phase C1O2
concentration in the air that passed through the impinger.
Using the formula below, the concentration of C1O2 (in parts
per million [ppm] volume in air) is calculated as
C1O3. ppm = —:
SR.Lmni' .TTime.mm 5
where V is the volume of STS titrant (mL, which is 10"3
L); M is the molarity of STS titrant (g divided by formula
weight) which for STS is equal to its normality, N); SR is
the sampling rate through the impinger (L min"1); Time is the
sampling time (min); 1/5 is the stoichiometric ratio of 1 mole
(mol) C1O2 reacting with 5 mol STS; 24.45 is the ideal gas
constant, L mol"1, at 25°C, 1 atm; and 1000 is the conversion
factor (L nr3). The formula weight of an ion species is equal
to its concentration in ppm at 10"3 M; therefore the results
from this equation are equivalent to ppm. The calculation
of concentration of C1O2 is limited to two significant
figures (e.g., 3100 ppm). The allowable variation in the
C1O2 concentration (±10%) encompasses the <2% variation
introduced by differences in test barometric pressure and
temperature conditions from ideal gas constant conditions
(latmand25°C).
2.2 Liquid Decontamination Technologies
Used with CWAs
Two aqueous solutions of oxidants were selected for testing.
One oxidant solution was diluted liquid bleach (sodium
hypochlorite) and the other was aqueous C1O2.
2.2.1 Liquid Decontamination Solutions Tested
The diluted bleach was prepared fresh daily using Clorox®
brand bleach that was less than 3 months old (based on
the code on the bottle). Clorox® bleach, 5%-6% sodium
hypochlorite, was diluted ten fold with deionized water to
give a nominal concentration of 5000 ppm CIO" oxidant.
(The concentration of CIO" in the bleach was not measured.)
The C1O2 was prepared fresh daily at a nominal concentration
of 3000-3500 ppm. This was prepared by adding 36 mL
of 6 N hydrochloric acid, 105 mL of Clorox® bleach, 100 mL
of 25% aqueous sodium chlorite to 3000 mL of deionized
water. The aqueous C1O2 solution was stored in a dark bottle.
The pH was checked to ascertain that it was between 4.5 and
7.5 units.
(1)
-------
An STS titration method was used to measure the
concentration of the C1O2 in the liquid decontamination
solutions. For this titration method, 5.0 mL of phosphate
buffer (pH 7.0) and a 5.0 mL sample of the C1O2 solution
was added to potassium iodide solution (150 mL water into
which 1.0 g of potassium iodide was dissolved). Under these
conditions C1O2 oxidizes iodide in the phosphate buffer
solution to iodine and C1O2 gas is reduced to the chlorite
ion which dissolves in solution. The molecular iodine that
is produced appears yellow-brown in aqueous solution.
The total resulting iodine is then reduced to iodide by
titrating with standard 0.1 M STS. The titration endpoint
is determined when the color of the solution changes from
yellow-brown to colorless and is recorded as "A". To the
solution, 5.0 mL of 6 N hydrochloric acid is added and the
titration is repeated; the result is recorded as "B".
The concentrations (mg/L) of C1O2 and chlorite in the
solution are calculated using Equations 2 and 3.
CIO,.we L"1 =
A. ml .A" N. woli's L ,T 6 ~. 450 tug wolf'
VjuL
(2)
A = mL of titrant to endpoint before acidification
N = normality of STS titrant (moles = equivalents)
V = mL of sample
67,450 = equivalent weigh (g mole'1) of CIO. x 1000 mg g'1
Chlorite.mz L1 = ~
B. ml - 4A. mil .T Ar moles L'1 ,T 16.S60 tug mole"1
J'.tiil (3)
B = mL of titrant to endpoint after acidification
4 A = proportion of chlorite measurement due to C1O2
16,860 = equivalent weight (g mole"1) of chlorite x
1000 mg g'1
2.2.2 Liquid Decontamination Test System
The liquid decontamination tests were carried out in 40
mL glass vials. A coupon of specified building material
was spiked with the indicated CWA and placed into the
vial which contained 10 mL of decontamination solution.
The cap was placed securely on each vial, and the vial
was laid horizontally so that the decontamination solution
fully covered the test coupon. The liquid decontamination
investigation is described further in Section 2.4.2.
2.3 TICs
2.3.1 Test Chamber for Fumigant Decontamination Tests
A customized test chamber consisting of fabricated and
off-the-shelf equipment and components was assembled and
used to carry out the investigation of the decontamination
of TICs. The 448 L test chamber (Labconco) is shown in
Figure 2-1 (on page 6). The temperature in the chamber was
maintained between 23 °C and 25°C. Hydrocarbon-free zero
air was supplied to the test chamber by a zero air generator
(AADCO). To achieve the desired RH in the chamber
at the start of a test, mass flow controllers (MFC; Sierra
Instruments) admitted both dry air and humid air in known
proportions. One MFC admitted moisture-free air to the test
chamber at a rate of 4.25 L miff1. The second MFC admitted
3.25 L miff1 of dry air through a 10 L miff1 RH generator
(Fuel Cell Technologies). The relative amounts of the two
air streams were adjusted to reach 80% ± 5% RH inside the
test chamber. A small 8-cm fan (Papst Model 8412), mounted
in the upper left side of the chamber, ensured that the test
chamber atmosphere remained well-mixed.
Temperature and RH were either recorded in real-time every
minute (for six of eight tests with TICs) with a National
Institute of Standards and Technology (NIST)-traceable
thermo-hygrometers (Control Company Model 4185), or
in near real-time roughly every 15 to 20 min (for two of
the eight tests) with a hygrometer (Control Company model
number 15551-126).
The building material coupons were placed on a custom-
fabricated polycarbonate carousel that was mounted inside
the test chamber, as shown in Figure 2-1. Two 8-cm fans
were positioned in a straight line along the carousel diameter
so as to pass air directly above the coupon surfaces. The
carousel completed one full rotation each minute. The
operation of the carousel was controlled to ensure that air
was passed across all coupons as uniformly as possible for
the duration of each 7 h test. Each of the two fans produced
an air velocity of 400 ft miff1 over the coupons as measured
by anemometers (TSI model 8455) placed downstream of
each of the two carousel fans.
To maintain one air exchange h'1 in the test chamber during
the without decontamination test with malathion, 7.5 L miff1
of hydrocarbon-free air (containing less than approximately
0.1 ppm total hydrocarbons) was admitted to the chamber.
For all other tests, no air was admitted to the chamber during
testing; maintaining essentially static conditions inside
the test chamber allowed for better stability in the C1O2
concentration. The flow rate of the air sampling performed
using Tenax® adsorbent cartridges to determine the gas-
phase TIC concentrations resulted in small air exchange
rates—only 0.03 to 0.05 air exchanges h'1 during 1 or 3 h
of sampling. Additional description of the air sampling is
provided in Section 2.3.7. Because the control tests without
fumigant C1O2 were conducted at a higher air exchange rate
(1 air exchange h'1), the control test for the volatile DMMP
was repeated with the minimal air exchange rate. As there
was no loss of malathion from the control coupons at the high
air exchange rate, this control was not repeated at a lower air
exchange rate.
-------
2.3.2 Building Materials
The test coupons included both porous (ceiling tile and
industrial grade carpet) and non-porous (decorative laminate)
surfaces. Test coupons were cut to 3.5 cm length x 1.5 cm
width (5.25 cm2) from larger pieces of stock material. Test
coupons were each visually inspected prior to being used in
any test. Coupons with anomalies on the application surface
were not used.
2.3.3 Sequence of Testing
The test sequence for each TIC and type of building material
was (1) the control condition (without C1O2 decontamination)
followed by (2) decontamination at 3000 ppm C1O2. Each
test included 25 coupons described in Table 2-1. At each
time point (1, 3, and 7 h), five test coupons, and one each of
the procedural blank, handling blank, and positive control
coupon were extracted and analyzed to determine the mass of
the TIC in the extract. At each time point during the control
condition tests (1, 3, and 7 h), five positive control coupons,
and one each of the procedural blank, handling blank,
and handling positive control coupon were extracted and
analyzed to determine the mass of the TIC in the extract. The
laboratory blank coupon was analyzed with the 1 h samples.
The same regimen was used for the tests done both with and
without decontamination. The test chamber was thoroughly
ventilated after a test with C1O2 before starting a control test
where there was no C1O2 being admitted to the chamber.
2.3.4 Source of TICs
The source, lot number and purity of the TICs used for the
decontamination tests are listed in Table 2-2 (on page 8)
(upper section of table); these parameters are also listed in
Table 2-2 (lower section) for the secondary source material
used in the QA performance evaluation (PE) audit.
Figure 2-1. Frontal (left) and Overhead (right) Views of Test Chamber Used for Decontamination Test of TICs
Table 2-1. Types of Coupons Used in the TIC Fumigation Investigation
Test coupon
(n = 15)
Procedural
blank (n = 3)
Handling blank
(n=3)
Handling
positive control
(n=3)
Laboratory blank
(n = l)
Spiked with TIC and exposed to chamber conditions
(with or without CI02)
Not spiked with TIC, and exposed to chamber
conditions (with or without CI02)
Not spiked with TIC, exposed in the fume hood during
the spiking process, then placed into sealed vials
Spiked with TIC and sealed in vials outside
of test chamber
Not spiked with TIC and sealed in vial outside
of test chamber
Coupons per Time
Five removed from chamber after each of
1, 3, and 7 h exposure times
One removed from chamber after each of
1, 3, and 7 h exposure times
One removed from vial and analyzed
along with test coupons at each time
point (1, 3, and 7 h)
One removed from vial and analyzed
along with test coupons at each time
point (1, 3, and 7 h)
One removed from vial and analyzed at
1 h time point
n = number of samples
-------
Table 2-2. Source of TICs and PE Audit Standards
Chemical
mcniuidiiuici/ . .. . ruiny ui
~ ,. K, Lot Number _ ' ..
Supplier Name Concentration
Concentration
Materials used for recovery and efficacy tests
Malathion
DMMP
Chem Service, Inc.
Sigma-Aldrich
343-1 10B
10110EA
99.2%
97%
10 mg ml_ !
in acetone
10 mg ml_ !
in acetone
Materials used for QA performance evaluation audit
Malathion
DMMP
Chem Service, Inc.
Chem Service, Inc
332-16B
08113TC
98%
97%
10 mg ml_ !
in acetone
10 mg ml_ !
in acetone
2.3.5 Application of TICs to Test Coupons
For both analytical method recovery testing and
decontamination testing, the test and positive control
coupons were spiked with 50 |jL of 10 mg per mL
(mg ml/1) of individual TICs to achieve a loading of 500
|j£ (~1 g m2). The addition of 500 ug of a TIC to carpet
laminate, or ceiling tile coupon is equivalent to 0.5 mg per
5.3 cm2, or about 1 mg per 10 cm2 or 1 g nr2. The spike of
each TIC was delivered from a variable volume pipettor
(Eppendorf) onto each test coupon in a laboratory fume
hood separate from the test chamber.
One laboratory blank coupon per test was not exposed
to TICs or to the laboratory atmosphere in which the test
chamber resides. Instead, when the coupons were retrieved
from storage, these blank coupons were placed immediately
into an airtight vial for subsequent extraction. Additional
laboratory blank coupons that were designated as handling
laboratory blanks for the different sampling intervals were
exposed to the laboratory air during the spiking of samples.
These blanks were loaded into vials when other coupons
were loaded into the chamber. Then, when coupons were
retrieved from the chamber for analysis, a handling blank
coupon was also retrieved for analysis. All other coupons
retrieved from storage were placed in the fume hood where
the test coupons and positive controls were spiked. The
process handling blank coupons were not spiked with TICs.
but were in the hood during the spiking and handling of the
test coupons. Subsequently, the process blank coupons were
in the chamber during the decontamination test.
For the analytical method recovery tests, the TIC and
SRS solutions were spiked onto the coupons just prior to
extraction. A short drying time was used to allow the solvent
to evaporate before extraction (Table 2-3). Similarly, for
decontamination tests, the coupons were placed in the
laboratory fume hood and spiked with the appropriate TIC
solution. The solvent was allowed to evaporate before
the coupons were placed in the test chamber. The solvent
evaporation times, listed in Table 2-3, were selected on the
basis of the TIC and coupon type. The test chamber was
already equilibrated at the appropriate temperature and RH
when the coupons were added. For the decontamination tests.
the SRS was not spiked onto each coupon until just before
analytical extraction.
Table 2-3. Solvent Evaporation Times for TICs Spiked
on Various Building Materials
Malathion
Malathion
DMMP
Material Evaporation Time (min)
Carpet
Laminate
Carpet, ceiling tile
30
3
1
2.3.6 Extraction Method for TICs from Test Coupons
For extraction, each coupon was placed into a 22 mL amber
glass vial and then spiked with 25 microliters (uL) of a
10 ug uL'1 solution of the appropriate SRS (to deliver 250
ug) for extraction. A 20 mL aliquot of acetone was added
to each vial; the vial was sealed with a screw-cap lid and
ultrasonicated for 30 min in an ultrasonic bath (Branson
5510). The extract was decanted to either a 200 mL TurboVap
tube or a 25 mL Kuderna-Danish (KD) tube with attached
125 mL reservoir. Carpet samples were extracted with three
replicate aliquots of acetone; extracts were combined before
concentration. Laminate and ceiling tile coupons required
only one extraction cycle. The number of extraction cycles
and concentration technique used for each TIC and building
material combination are listed in Table 2-4. Extracts were
concentrated to a final volume of 5 mL and spiked with 25 uL
of a 10 ug uL'1 solution of the appropriate IS to give a 50 ug
Table 2-4. Extraction and Concentration Techniques Used for TICs
Coupon Type
Carpet
Carpet
Laminate
Ceiling tile
TIC
Malathion
DMMP
Malathion
DMMP
txtraction
Technique
Sonication
Sonication
Sonication
Sonication
Extraction
3x 20 mL
3x 20 mL
1 x 20 mL
1 x 20 mL
concentration
Technique
TurboVap
KD
TurboVap
KD
-------
mL"1 concentration. A 1 mL aliquot was then filtered through
a disposable syringe filter (GD/X; Whatman) prior to the gas
chromatography/mass spectrometry (GC/MS) analysis.
2.3.7 Air Sampling and Analysis of TICs Using Tenax®
Sorbent Air Sampling Tubes
Sampling of the chamber air to assess vapor phase
concentrations of the TICs was accomplished using active
sampling onto the Tenax® TA™ sorbent. The air sampling
cartridges consisted of stainless steel thermal desorption
tubes (0.25 inch outside diameter x 3.5 inch length) packed
with 200 mg of 60/80 mesh Tenax® TA polymer (Markes
International, Ltd). Sampling was performed using calibrated
mass flow controllers. For malathion, sampling rates were
500 mL min"1 for control conditions and 200 mL min"1 for test
conditions, both for 3 h. The sampling rates were lowered
to reduce the air exchange rate in order to stabilize the C1O2
concentration during decontamination testing. For DMMP, all
sampling was performed at 100 mL min"1 for 1 h. The longer
sampling times and higher sampling rates for malathion were
used because it was expected that chamber air concentrations
of malathion would be considerably lower than those of
DMMP due to vapor pressure differences.
At the conclusion of sampling during malathion tests, the
Tenax® air cartridges were spiked with 100 ug of the SRS
(fenchlorphos) and extracted with 2 mL of acetone; after
addition of the IS, the 2 mL volume was analyzed without
further concentration. The GC/MS analysis for the air
extracts was performed as described in Section 2.3.8 for
coupon extracts.
At the conclusion of sampling during DMMP tests.
the Tenax® air cartridges were spiked with diisopropyl
methylphosphonate (DIMP) as an IS and analyzed using
thermal desorption (TD) GC/MS technique (using a Markes
Unity Thermal Desorption unit interfaced to an Agilent
6890/5973 GC/MS). Before desorption, a tube leak check
was performed to ensure quantitative sample transfer. After
the leak check, the tube was flushed with helium for one
min at 40 mL min"1, with the effluent sent to vent. Following
the purge, analytes were desorbed from the sorbent tube at
20 mL min'1 for 10 min at 300°C; analytes were collected
on the focusing trap (Markes General Purpose Graphitized
Carbon trap) which was held at 25°C. The analytes were then
desorbed from the focusing trap at 300°C for five min using
the maximum heating rate (>60°C sec"1) and a flow rate of
40 mL min'1. A portion of this flow was directed to the GC/
MS. Minor modification of the tube desorption flow rate was
used in various sets of analyses to optimize the percentage
of analyte transfer from sorbent tube to GC/MS for analysis.
In each set, however, samples and standards were analyzed
under identical conditions. The GC/MS conditions for
analysis of vapor phase DMMP are given in Table 2-5.
2.3.8 Analysis Method for TICs
Coupon extracts were analyzed using GC/MS in multiple
ion monitoring mode on an Agilent 6890/5973 GC/MS. Data
collection, reduction, and analysis were performed using
Agilent Chemstation software, version B.02.05. The GC and
MS conditions used for analyses of the two different TICs are
listed in Table 2-6.
Table 2-5. GC and MS Conditions for Analysis of DMMP in Air Samples
Parameter
GC column
Temperature program
Transfer line temperature
MS source temperature
Quadruple temperature
Flow rate
Ions monitored
^•JCTj
DB-5ms; 60 meters (m) x 0.32 mm inside diameter x
0.25 urn film thickness (Agilent)
50°C (2 min); 50°C-140°C @ 6°C mirr1; 140°C-250°C
@ 20°C min-1; hold 0.5 min (23 min run time)
280°C
230°C
150°C
1.5 ml min1 at 50°C
m/z 94, 79 for DMMP and m/z 97, 123 for DIMP
Table 2-6. GC and MS Conditions for TIC Analyses
GC column3
Inlet liner
Temperature program
for malathion
Temperature program
for DMMP
GC injection
Transfer line temperature
MS source temperature
Quadruple temperature
DB-1701; 30 m x 0.25 mm ID x 0.15 urn film thickness; J&W Scientific
Siltek double gooseneck
100°C (2 min); 100°C-180°C @ 10°C min-1; 180°C-220°C @ 5°C min-1;
220°C-260 °C @ 20°C min ! (20 min run time)
50°C (2 min); 50°C-95°C @ 3°C min !; 95°C-250°C @ 20°C min !;
hold 2.25 min (27 min run time)
1 uLsplitlessat 280°C
280°C
230°C
150°C
a) In all cases helium was the carrier gas: 0.8 ml mirr1 flow for malathion; 1 ml mirr1 for DMMP
-------
Table 2-7. GC Retention Times and Monitored Ions
for TIC Analyses
Chemical
Malathion
SRS
IS
DMMP
SRS
IS
Ions Monito
Time, min QuantifiCation Ion
17.2
15.5
16.5
9.0
17.0
15.6
^^B^^^|
285
312
94
111
97
Qualifier Ion
^Eg^B
125
152
79
93
123
Two ions were monitored for each TIC, SRS, and IS. The primary ion was used for quantification
and the secondary ion was used for qualitative confirmation of identification. Criteria for
identification of an analyte included the correct GC retention time ±0.02 min, chromatographically
co-maximized primary and secondary ions and correct ratio between the relative intensities of the
primary and secondary ions. The monitored ions and GC retention times are listed in Table 2-7.
The quantification was performed using the IS method.[2] The IS was present at the same
concentration in all samples and standards. The 11-point calibration curve spanned the range of
0.1-150 ug ml/1. This concentration range is equivalent to 0.1% to 150% recovery of the mass
applied to coupons.
The full calibration curve was analyzed at the start of each analysis set. Samples were then analyzed
with the 20 ug ml/1 standard run after every five samples as a continuing check on the calibration.
If the calculated concentration of the continuing calibration standard showed a variance more than
20% beyond its true concentration, the cause of the problem was investigated and the five samples
before and after this standard were reanalyzed. Calibration curves were constructed using a quadratic
least-squares regression analysis routine with the weighting scaled by the inverse of the analyte
concentration. Typically, the calibration data could be fitted to a single curve for malathion and
DMMP. However, due to the wide calibration range, occasionally two separate calibration curves
(one with high values and one with low values) were needed to define the malathion calibration data.
2.3.9 Calculation of TIC Recovery Efficiency, Percent Recovery,
and Decontamination Efficacy
The analytical method performance recovery efficiency for both a TIC and its matched SRS extracted
from a spiked coupon was determined according to the following formula:
,,, , , _ „,,., ,, Concentration, us!mL) Extract Volume, mL ,nm (A\
Analytical Method Recovery Efficiency; % = — ; • 100% W
MUM Applied. //?
Using results from Equation 4 for SRS and TIC recoveries, the analytical method recovery ratio
between the mean analytical method SRS recovery and the mean analytic method TIC recovery was
determined using the following equation:
SRS /TIC Recovery Ratio =
Mean Analytical Method SRS Recovery,'
Mean Analytical Method TIC Recovery. °
(5)
The raw recovery value for the individual TIC or SRS from a building material coupon during a
test was calculated by multiplying the mass (ug ml/1) determined in the analytical method by the
extract volume (mL). The raw recovery values were adjusted using the SRS/TIC recovery ratio from
Equation 5 to generate the normalized TIC recovery values according to the following equation:
,. ._„-,,- - Raw TIC recovery, jt.
Normalized TIC Recovery, % = =—^-
Raw SRS recovery, n
SRS .••' TIC Recovery Ratio • 100%
(6)
-------
The mean normalized TIC recovery from Equation 6, at each sampling interval, was divided by the
mean normalized TIC recovery at Time 0 of the test to obtain the percent recovery at a given time
reported in the results:
_. . Alean Normalized RecoveiT citTime ,.°o ..,, ,_,
Recoveiy fir Tniu>,,vo = '- • 100°o (7)
Mean Normalized Recowiy fit Time a, °o
The decontamination efficacy of a decontamination technology was calculated as the ratio of
the mean recoveries extracted from a coupon at a given time (from Equation 7) for the test
(decontaminated) and control using the following equation:
[ Recovery of TIC at Time With Decontamination, % ]
Decontamination Efficacy at Time,., %= 1- xlOO%
^ Recovery of TIC at Times Without Decontamination,°o)
(8)
The SRS recovery in each particular sample was used to correct for potential differences in
application of the analytical method to each individual sample. The SRS/TIC recovery ratio,
calculated using Equation 5, used data from the recovery measurements of these compounds in the
method performance tests. This ratio was used to make an adjustment for the fact that an isotopically-
labeled chemical equivalent (e.g., deuterium - or CIS-labeled) of each TIC was not available for
use as the SRS. In many analytical methods, a labeled version of an analyte is used as the SRS; in
that case the analyte and SRS are assumed to be recovered through an analytical method to the same
extent. Where an isotopically-labeled version of an analyte is not available, an SRS is chosen to be as
similar as possible to a given analyte so as to minimize the potential for differential loss mechanisms.
However, there may be differences in recovery between an analyte and SRS, and this difference
requires a minor adjustment based on differential recoveries.
To convert a Tenax® sorbent sample extract concentration for malathion to an air concentration of
malathion in the test chamber, the following equation was used:
, _3 Concentration in Extract, u%inU ExltnclToLiiiL , ,3 T _3 ,n,.
Concentration in Air, t/giii = ^j-^ 10 L in (?)
SatnplitigRate, Latin ScttnplitigTiiiie, min
To convert an amount of DMMP measured on a Tenax® sorbent tube to a DMMP air concentration in
the test chamber, the following equation was used:
. ,, _3 Ailtoi int Desorhed fi'ont Tube, tig 3 -3/1™
Concentrationni Air, iigin = ^j-^ — 10 L in (10)
Sampling Rate. L niin Sfinip/iiigTitiie, min
2.4 CWAs
2.4.1 Test Chamber for Fumigant Decontamination Tests
Two test chambers were constructed for testing of each CWA. One chamber housed the control
coupons that were not exposed to C1O2; the second chamber was interfaced to the Sabre C1O2
generator, so that the coupons housed in this chamber could be decontaminated. Each chamber
consisted of a specially fabricated polycarbonate (Lexan®) housing with removable custom built
shelves made of 26 gauge cold rolled steel. The chamber had dimensions of 26 cm x 29 cm x 27
cm, or 20.4 L. A new polycarbonate chamber with shelves was used for each CWA tested. The
temperature in the chambers was that of the ambient air in the laboratory. An ultrasonic fog generator
was used to establish -100% RH air in a separate fogging chamber. The humidified air from the
fogging chamber was pumped through a water trap and then into the two test chambers until the
test chambers reached 80% RH. Each chamber was then sealed for the test. The test chamber
was interfaced with the Sabre C1O2 generator, which was described earlier (Section 2.1.1). At the
conclusion of each test, the chambers were decontaminated and decommissioned according to
Department of the Army regulations and BBRC standard operating procedures.
The temperature and humidity in the chamber were monitored continuously (at 30 min intervals)
by the heating, ventilation, and air conditioning (HVAC) system. The HVAC readings were verified
twice daily using a calibrated NIST-traceable thermometer/hygrometer with accuracy of ±1°C
for temperature and ±5% for RH. All of the readings taken in the laboratory indicated that the
temperature and RH were constant throughout the test periods.
-------
2.4.2 Test Chamber for Liquid Decontamination Tests
As described earlier (Section 2.2.2), the testing for
decontamination of building material coupons with liquid
solutions was carried out in individual 40 mL vials. Each
spiked or blank coupon was placed in a vial and 10 mL of
decontamination solution was added. The vial was capped
securely laid horizontally so that the coupon was bathed
completely in the decontamination solution during the test
period. The vials were periodically shaken for the first few
hours to extract the CWA from the coupon and the coupons
were allowed to stand in the extraction solution overnight. At
the completion of the specified decontamination period, the
coupon was removed and shaken lightly to remove as much
aqueous liquid as possible; the SRS was added to the coupon
immediately and placed into a 40 mL vial containing 10 mL
of extraction solvent. The vials were again shaken a few
times before the coupon was removed.
Neutralization of the decontamination solution was not used
in the decontamination efficacy tests because the preliminary
solution tests (Section 2.2.2) suggested that loss of CWA
recovery occurred in neutralized solution, except possibly
for GD. Rapid extraction was preferred to inclusion of a
neutralization step. Therefore, the coupons were removed
from the decontamination solution, shaken lightly to remove
excess liquid, spiked with the SRS, and placed directly into
the extraction solution.
2.4.3 Building Materials
The building materials that were utilized in testing for
decontamination efficiency for fumigant and liquid
decontamination technologies are listed in Table 2-8; these
materials included porous, adsorptive and non-porous surface
types. Test coupons were cut to the sizes indicated in Table
2-8 from larger pieces of stock material.
Table 2-8. Building Material Test Coupon
Characteristics for CWA Decontamination Tests
Material
Laminate
Carpet
Ductwork
Approximate Coupon Surface, Material
Length x Width, cm Preparation
6.5x 1.5 (9. 75 cm2)
6.5x 1.5 (9. 75 cm2)
6.5x 1.5 (9. 75 cm2)
Clean with
acetone
None
None
2.4.4 Sequence of Testing
For the fumigant testing with C1O2 generated by the Sabre
system, all building material types were tested concurrently
in a given test. However, since the test (i.e., fumigant) and
control chambers were both sealed to maintain the RH, only
enough coupons were loaded into the identical chambers for
a single time increment of decontamination. For example.
when testing decontamination of VX, the identical chambers
were each loaded with five spiked carpet coupons, five
spiked laminate coupons, five spiked ductwork coupons, and
one procedural blank of each material type. The chambers
were sealed and the decontamination and control tests were
then carried out for 1 h. After unloading the chambers and
ventilating them, an identical set of coupons were prepared.
loaded into the chambers, and this time the test was
conducted for 2 or 4 h, depending on CWA.
To select appropriate conditions for the liquid
decontamination tests, several preliminary tests were
conducted to determine whether neutralization of the liquid
decontamination agent was needed prior to extraction of
the CWA with the organic solvent. To carry out this test.
each CWA was spiked into both the full strength liquid
decontamination (bleach or C1O2) and the neutralized liquid
decontamination solution. The CWA was held in the 10 mL
decontamination solution for 1 h prior to addition of the
SRS and the 10 mL of hexane for extraction; in contrast, the
CWA was held for 15 sec in the neutralized decontamination
solution prior to addition of the SRS and the 10 mL of hexane
for extraction. The neutralized solution consisted of the
decontamination liquid and STS equivalent to 10% more
than that required to neutralize the oxidant; this solution was
then adjusted to pH 7.0 by the addition of acetic acid before
extraction with hexane.
For the tests of decontamination with liquid bleach, testing
was done sequentially with individual combinations of
CWA and building material because of the very short
decontamination times being tested (10-30 min). The
sequence of testing included first, the analysis of Time 0
positive controls without decontamination and laboratory
blanks of a given building material and CWA for verification
of mass applied. This group of samples included five positive
control coupons of the building material spiked with 1 mg
of the CWA, and one coupon of that material that was a
laboratory blank. These coupons were spiked (as needed)
and immediately extracted in hexane. Then, that same
combination of building material and CWA was tested for
different durations with and without decontamination. Each
test grouping included, for instance, five carpet coupons
spiked with 1 mg of CWA and loaded into vials containing
10 mL of diluted bleach (i.e., with decontamination), five
carpet coupons spiked with 1 mg of CWA and placed
in empty vials (i.e., without decontamination), and one
procedural blank carpet coupon that was placed in a vial
with diluted bleach. This sequence was completed for all
decontamination test durations with carpet, then laminate and
then ductwork first for VX, then repeated again with TGD.
and then repeated a third time with GB (here, tested only on
carpet coupons because GB was not persistent on laminate or
ductwork). The only deviation to the test/QA plan was that
positive control coupons (i.e., without decontamination) for
VX were not prepared for the 10 and 20 min decontamination
durations; these coupons were only prepared for the 30 min
test duration due to the very low vapor pressure and expected
stability of VX over 10 and 20 min test periods.
In each test combination of CWA and building material, the
ten spiked building material coupons were loaded into vials.
alternating between a vial containing the decontamination
solution and a control vial that was empty, as a control for
bleach, or containing acidified water, as a control for C1O2
decontamination solution. This alteration between test and
-------
control vials was done to minimize the effect of loading
time. The same alternating sequence was used in retrieving
coupons from vials, spiking them with SRS and loading them
into the hexane-containing extraction vials — once again
to minimize the effect of time required for spiking SRS and
loading the coupon into an extraction vial. This was deemed
potentially important for the more volatile CWAs (GB
and TGD).
Decontamination tests using C1O2 decontamination solution
were conducted only for VX because the results of the
solution tests suggested that decontamination of GB and GD
with C1O2 solution was not effective. Solution testing results
with VX were inconclusive, possibly due to hydrolysis in the
aqueous solution. For decontamination with C1O2 solution.
the same sequence as described above for bleach was used.
This included preparing only one coupon type at a time, and
preparing these for only one decontamination duration.
For decontamination of VX with C1O2 solution, the positive
control samples (spiked without decontamination) were
handled differently. VX is quite unstable in alkaline solutions
and this was a possible explanation for the inability to
recover VX from neutralized C1O2 decontamination solution
in the solution tests. Therefore, rather than load the positive
control coupons into empty vials for the test duration, they
were loaded into vials containing 10 mL of acidified (pH
4.5-<7) water; VX is quite unstable in alkaline solutions. The
recovery of VX from acidified water enables the hydrolysis
effects of water to be differentiated from the decontamination
effects of the C1CX.
2.4.5 CWAs and Surrogate Recovery Standard
The source, lot number, and purity of the CWAs used in this
investigation are listed in Table 2-9.
Polymethyl methacrylate was added, 5% on a weightvolume
basis, as a thickening agent for GD. Typically, 5 mL of TGD
was prepared in a batch.
2.4.6 Application of CWAs to Test Coupons
For both the control tests (with no C1O2 admitted to the
chamber) and the decontamination tests with C1O2, the
coupons were spiked with the individual CWA to achieve
a loading of approximately 1 g nr2. All building materials
were spiked with 1 uL of neat CWA to deliver a mass of
approximately 1 mg to one 10 cm2 surface of each coupon.
A 50 uL repeating dispenser pipette (Hamilton) that delivers
50 equal volumes per syringe load was used to apply the
CWA to the test coupons. Concurrently with spiking the test
coupons, 1 uL of each CWA was spiked directly into 10 mL
of the extraction solvent and this solution was analyzed to
measure the mass of CWA applied to the building materials
(i.e., a confirmation spike). The mass of CWA applied to test
coupons, as determined from analyses of the confirmation
spikes, are listed in Table 2-10.
For these decontamination tests, the coupon spiking was
completed within approximately 30 sec, and coupons were
loaded directly into the test chambers after spiking. Drying
time was not needed since there was no solvent involved.
Table 2-9. Source of CWAs and SRS
Chemical
Manufacturer/Supplier
Purity or Concentration
Concentration as ADD! led
Materials used in this investigation
GB
GD
VX
TBP (SRS)
U.S. Army
U.S. Army
U.S. Army
Sigma-Aldrich
96
94
70
99
Neat
95% neat
Neat
Neat
Standard Analytical Reference Material used to confirm CWA purity
GB
GD
VX
Institute of Chemical Defense
Institute of Chemical Defense
Institute of Chemical Defense
1 mg ml_ !
1 mg ml_ !
1 mg ml_ !
NA
NA
NA
Table 2-10. Mass of CWA Applied to Building Material Coupons
GB
TGD
VX
Mass of CWA A
FurnJEfant CIO_
916 ug
880 ug
833 ug
pplied for Decontamination Tests
Liauid Bleach Liauid CIO.
830 ug
1530 or 1570 ug,
depending on day
840 ug
NTa
NT
950 ug
a) NT = not tested
-------
2.4.7 Extraction Method for CWAs
For extraction of building material coupons, 1 uL of the SRS
(1 uL = 1 mg) was first applied as neat material to the coupon
just prior to extraction. The coupon was then loaded into a
40 mL sample extraction vial and a 10 mL aliquot of hexane
containing the IS at 100 ug ml/1 was added. The vial was
shaken briefly and then the building material was allowed
to stand in the solvent overnight (-14-16 h) for passive
extraction. The vials were shaken again before removal of the
analysis aliquot.
2.4.8 Analysis Method for CWAs
Sample extracts were analyzed using gas chromatography
with flame photometric detection (GC/FPD) on an Agilent
6890 GC. Data collection, reduction, and analysis were
performed using Agilent Chemstation software, Version
B.02.05. The GC conditions used for analyses of the three
different CWAs are listed in Table 2-11.
The GC retention times were monitored for each CWA, SRS.
and IS. Identification of an analyte entailed matching the
correct GC retention time ±0.02 min. The GC retention times
are listed in Table 2-12. The quantification was performed
using the IS method.[2] The 9-point calibration curve
spanned the range of 0.24-190 ug ml/1. This concentration
range is equivalent to 0.24% to 190% recovery of the mass
applied in this investigation. The IS was present at the same
concentration in all samples and standards.
2.4.9 Calculation of CWA Recovery Efficiency, Percent
Recovery, and Decontamination Efficacy
The calculations of recovery efficiency (Equation 4).
percent recovery (Equation 7), and decontamination
efficacy (Equation 8) were carried out using the same
equations listed and described in Section 2.3.9 for the TICs.
The only exception here for the CWAs was that SRS and
relative extraction efficiency corrections were not made in
calculations for the data from the liquid decontamination
tests. In these data, for unknown reasons, the SRS recoveries
from the coupons were not stable. The SRS recoveries were
measured, but were not useful for comparison.
2.5 Qualitative Evaluation of the Impact of CI02
Fumigation on Building Materials
A total of eight laboratory blanks (building materials not
exposed to C1O2) and 24 procedural blanks (building
materials exposed to C1O2) were maintained under controlled
ambient temperature and RH conditions for assessment of
long-term changes in appearance or structure due to C1O2
fumigation. For this work, two polycarbonate chambers,
one 30 L and the other 15 L, were fabricated, and connected
via a cylindrical flow-through unit. The front end of the
larger chamber was connected to the humidification system
described earlier; the back end of the smaller chamber was
exhausted to ambient. Ambient air, typically at 24°C and
40% RH, was directed through these chambers with a flow
of one air exchange h'1. The laboratory blank coupons were
placed in the first chamber and the procedural blank coupons
exposed to CIO were placed in the second chamber.
Table 2-11. GC/FPD Conditions for CWA Analyses
Table 2-12. GC Retention
Times for CWA Analyses
rarameier
GC column for GBa
Temperature program for GB
GC column forTGD3
Temperature program for TGD
GC column for VXa
Temperature program for VX
GC injection
Detector temperature
Hydrogen flow
Oxidizer flow
Makeup gas flow
DB-5; 25 m x 0.32 mm inside diameter x 0.52 urn
film thickness; Agilent
55°C (1 min); 55°C-100°C @ 10°C/min; 100°C
-250°C @ 25°C/min (11.5 min run time)
Rtx-5; 30 m x 0.32 mm inside diameter x 0.50 |jm
film thickness; Restek
40°C (1 min); 40°C-100°C @ 10°C/min; 100°C
-250°C @ 30°C/min (12 min run time)
DB-5; 25 M x 0.32 mm ID x 0.52 urn film
thickness; Agilent
55°C (1 min); 55°C-100°C @ 10°C/min; 100°C
-300°C @ 25°C/min (13.5 min run time)
1 uLsplitlessat250°C
250°C
70 ml mirv1
Air at 90 ml min !
Nitrogen at 15 ml mirv1
GC Retention
Time (min)
GB
SRS
IS
TGD isomer 1
TGD isomer 2
SRS
IS
VX
SRS
IS
a) In all cases helium was the carrier gas with a flow rate of 1.7 ml min"1
3.49
11.6
6.92
6.62
6.67
11.8
6.92
6.16
5.74
2.05
-------
3.0
Quality Assurance/Quality Control
QA/quality control (QC) procedures were performed in
accordance with the TTEP Quality Management Plan
(QMP)Pi, the test/QA planM, Amendment 2 for C1O2
decontamination of TICs (March 16, 2006), Amendment 6
for C1O2 decontamination of CWAs (March 21, 2007), and
Amendment 8 for liquid technologies for decontamination
of CWAs (July 6, 2007). QA/QC procedures are summarized
below.
3.1 PE Audit
A PE audit was conducted to assess the quality of the GC/
MS results obtained during these tests. For the two TICs, this
PE audit was performed by diluting and analyzing standards
obtained from a secondary source. The secondary source
standards were diluted to 100 (ig ml/1 and analyzed using
a calibration curve constructed from the primary source
standards. The results of this analysis are given in Table 3-1.
The target tolerance was a difference less than 25%; results
were well within the target tolerance.
A PE audit was conducted to assess the quality of the
CWA results obtained during these tests. For the three
CWAs, this PE audit was performed by comparing purity
of stock solution results against a second source (Standard
Analytical Reference Material). In addition, PE audits
were performed for the chamber conditions that affected
results (time, temperature, RH and flow). The results of
these analyses are given in Table 3-2, except for time.
Time comparisons were made four times over the span of
40 min on April 27, 2007 and May 31, 2007. Time was
compared to the U.S. government's time provided by both
NIST/U.S. Naval Observatory (Military counterpart to NIST)
and the Department of Commerce (www.time.gov). Each
measurement was exact to the second, resulting in a percent
difference of 0.0% for both days. The target tolerance was a
difference less than 10%; results were well within the target
tolerance.
3.2 Technical Systems Audit
The Battelle QA Manager and his designee conducted a
technical systems audit (TSA) to ensure that the tests were
being performed in accordance with the QMP, test/QA plan.
Amendment 2 for C1O2 decontamination of TICs (March 16.
2006), Amendment 6 for C1O2 decontamination of CWAs
(March 21, 2007), and Amendment 8 for liquid technologies
for decontamination of CWAs (July 6, 2007). As part of the
audit, the Battelle QA Manager and his designee reviewed
the reference sampling and analysis methods used, compared
actual test procedures with those specified in the test/QA
plan, and reviewed data acquisition and handling procedures.
No significant findings were noted in this audit that might
impact the quality of the investigation results. The records
concerning the TSA are permanently stored with the Battelle
QA Manager.
Table 3-1. TIC PE Audit Results
Malathion
DMMP
^^^B
10/27/2006
10/27/2006
Standard
Concentration
0.100 mg ml1
0.100 mg ml1
Measured
Result
0.087 mg mL1
0.104 mg ml1
^^^1^
-13
4.0
Table 3-2. CWA PE Audit Results
CWA or Parameter Date of Audit Standard Measured Result % Difference
GB
TGD
TGD
VX
MFC3
Temperature
RH
Temperature
RH
6/06/2007
5/03/2007
5/11/2007
5/02/2007
5/16-18/2007
4/27/2007
4/27/2007
5/31/2007
5/31/2007
0.110 mg ml1
0.110 mg ml1
0.110 mg ml1
0.070 mg mL1
1.00 L min1
19.3°CC
36.3%c
22.1°CC
38.8%c
0.114 mg ml1
0.106 mg ml1
0.103 mg ml1
0.109 mg ml_lc
0.111 mg ml_lc
0.073 mg mL1
1.01 Lminlb
18.8°CC
37.8%c
22.4°CC
41%c
-3.6
3.6
6.4
0.9
-0.9
-4.3
-0.6
2.6
-4.1
-1.4
-5.8
a) Mass Flow Controller
b) Average of six results, three for 5/16/2007 and three for 5/18/2007
c) Average of four results
-------
3.3 Data Quality Audit
At least 10% of the data acquired during the investigation
was audited. Battelle's QA Manager traced the data from
the initial acquisition through reduction to final reporting, to
ensure the integrity of the reported results. All calculations
performed on the data undergoing the audit were checked.
3.4 QA/QC Reporting
Each assessment and audit was documented in accordance
with the test/QA plan and QMP. For this investigation, no
significant findings were noted in any assessment or audit,
and no follow-up correction action was necessary. Copies of
the TSA and assessment reports were distributed to the EPA
QA Manager and Battelle staff.
QA/QC procedures were performed in accordance with the
Program QMP and the test/QA plan for this investigation.
3.5 Deviations from Test/QA Plan
The test/QA plan envisioned use of a 317 L test chamber for
testing decontamination of TICs. When the coupon carousel
and equipment would not fit into this sized chamber, a 448
L chamber was substituted. This change did not impact the
investigation.
-------
4.0
Results
4.1 Results for Fumigant CIO Decontamination
of TICs
4.1.1 Analytical Method Development Results
The analytical methods were tested initially to verify the
extraction efficiencies of the TICs and their matched SRSs
from the building materials that were to be used in the
decontamination tests. These recoveries are listed in
Table 4-1.
The mean of the extraction efficiencies for the TICs and
SRSs were in a range of 73% to 97%. This range of recovery
was within the acceptance range specified in Appendix A3 of
the test/QA plan (40% to 120%) and all values were within
the acceptance level of 3 SD of the mean. The SRS/TIC
values (0.91-1.11) show that the SRS recovery correlated
well with the respective TIC recovery. These results show
that the method development resulted in extraction and
analysis methods for the TICs that were sufficient to achieve
the data quality objectives of the test/QA plan (inclusive of
the requirements in Appendix A3).
The approximate method detection limits (MDLs) for the
TICs are listed in Table 4-2. Each MDL was estimated based
on the signal of the lowest level calibration standard, the
signal to noise ratio for this concentration and the peak area
that can be integrated reliably for any signal.
4.1.2 Environmental Conditions during
Decontamination Tests
The temperature, RH, C1O2 concentration, and air velocity
over the coupons during testing were measured as specified
in the test/QA plan.141 The averages for temperature, RH.
C1O2 concentration, and air velocity over the coupons during
testing, in control and test chambers, are shown in Table 4-3
(on page 18). Temperature and RH were controlled to 24° C
± 2°C and 80% ± 10% RH. The target C1O2 concentration
was 3000 ppm. Differences in air velocities among the tests
were observed. The mean of the average velocities for all of
the tests was 117 ft mirr1; all mean velocities were within 2
SD of the mean of the averages. The differences in velocity
within and among tests may be due to small changes in the
operation of the fans or positioning of the anemometers.
Originally, one air exchange Iv1 was planned for the
persistence testing and the C1O2 fumigation testing. However.
at one exchange Iv1, control of the C1O2 concentrations was
problematic. Therefore the air exchange rate was reduced to
<0.1 air exchange Iv1 for the C1O2 fumigation testing.
Due to the high volatility of DMMP, there was a concern
that the difference in air exchange between the control tests
(one air exchange Iv1) and the decontamination tests (0.1 air
exchange Iv1) would present a variable that would interfere
with the analysis of data and trends. For this reason, the tests
Table 4-1. Extraction Efficiencies of TICs and Matched SRSs
from Building Materials
Material
Carpet
Laminate
Carpet
Ceiling tile
Recovery from Bulk
Malathion
80 ±3
80 ± 19
DMMP
87 ±3
93 ±3
Jmg Material, % ± SDa
Fenchlorphos
73 ±5
89 ±4
DIMP
87 ±7
97 ±3
SRS/TIC
Recovery Ratiob
0.90
1.1
0.99
1.1
a) Calculated using Equation (Eq) (4)
b) Calculated using Eq (5); carried out using 3 significant figures on TIC and SRS
recoveries
Table 4-2. MDLs for TICs
Malathion
DMMP
0.01 ug ml_]
In solution
0.01 ug ml_]
On coupon
In air
0.05 ug
0.26 ug m:
0.05 ug
0.70 ug nr
-------
for DMMP (without C1O2 decontamination) were conducted
a second time at the low air exchange rate to determine
whether air exchange has any significant effect on recovery.
The recovery of DMMP on carpet and ceiling tile with
the two different air exchange rates, shown in Table 4-4.
demonstrates that there is a small difference after 1 h, but
that difference largely dissipates by the conclusion of the 7 h
test. Statistical analysis of these data showed that the small
differences on the ceiling tile were statistically significant.
For DMMP, the recoveries at the low air exchange rate were
used as the basis for comparison with the recoveries after
decontamination.
There was not a significant loss of malathion from the test
coupons over a 7 h at one air exchange h'1. Therefore, the
persistence testing was not repeated at the lower air exchange
rate. For malathion, the recoveries at one air exchange h'1
were used as the basis for comparison with the recoveries
after decontamination.
4.1.3 Recovery over Time of TICs on Building Materials
With and Without Sabre CI02 Fumigant Decontamination
Individual tests, both with and without the Sabre C1O2
fumigant decontamination technology, were conducted for
each building material to assess the recovery of the given
TIC on that building material. As noted in Section 1.3
(Experimental Design), the decontamination testing covered
a 7 h time span, and five building material coupons were
removed from the chamber for testing after 1, 3, and 7 h. In
addition, analyses of five spiked building material coupons
that were extracted immediately after spiking (Time 0) were
used to ascertain the baseline recovery values. The recoveries
of the spiked TICs from the coupons at Time 0 (0 h) and
at subsequent times (1, 3, and 7 h) with and without the
application of the Sabre C1O2 decontamination process are
listed in Table 4-5 (on page 19).
Table 4-3. Mean Temperature, RH, CIO Concentration, and Air Velocity during TIC Decontamination Tests
DMMP
Malathion
Carpet
Ceiling tile
Carpet
Laminate
Control
Test
Control
Test
Control
Test
Control
Test
Mean During the Test
Temp, °C ± SD RH, % ± SD CI02, ppm ± SD (n) Air Velocity, ft min -1 ± SD
24.3 ±0.1
23. 9 ±0.1
24.2 ±0.2
25.1 ±0.8
23. 7 ±0.1
24.4 ± .03
22. 9 ±0.2
23.0 ± 1.1
80 ±4
78 ±3
82 ±3
83 ±2
84 ± 1
77 ±2
87 ±2
79 ±3
0
3300 ± 600
(n = 17)
0
3010 ± 120
(n = 21)
0
3050 ± 200
(n = 21)
0
3010 ± 180
(n = 21)
102 ± 11
108 ±0
136 ± 18
121 ±3
108 ± 16
140 ± 30
116± 11
105 ± 13
Table 4-4. Comparison of Mean Percent Recovery of DMMP with Different
Chamber Air Exchange Rates
TIC
DMMP
Material
Carpet
Ceiling tile
Time (n)
1 h (n = 5)
3 h (n = 5)
7 h (n = 5)
1 h (n = 5)
3 h (n = 5)
7 h (n = 5)
0.1 Air Exchange h'1 1 Air Exchange h *
33 ±2
24 ±3
16 ±3
39 ±4
28 ± 1
12 ± 1
40 ±5
26 ±3
18 ±2
32 ± 1
19 ±2
9.4 ±0.4
-------
As shown in Table 4-5, the amount of DMMP on carpet
and ceiling tile decreased quite substantially both with
and without the decontamination treatment. At this point.
it is not clear whether the significant decrease over time
was due to volatilization of DMMP from the coupons or a
combination of degradation and hydrolysis due to the high
RH in the chamber. Statistical analysis of the data was used
to determine whether there was a statistically significant
difference attributable to the C1O2 decontamination treatment.
For malathion, it was readily evident that the
decontamination with fumigant C1O2 resulted in a significant
decrease in the amount of the TIC remaining on the coupons.
Without decontamination, most of the malathion remained on
the coupon; and with decontamination less than 1% remained
on the laminate and approximately 25% remained on the
carpet after 7 h.
The recoveries of DMMP and malathion on these coupons.
normalized to the recoveries at 0 h, are presented in
Appendix A, Table A-l of this report. These normalized data
were used as the input to the statistical analyses described
below. The statistical analyses are more informative than the
normalized recovery data, in that they provide insight into
whether there is a statistical difference between recoveries
with and without decontamination. Therefore, to avoid
confusion by the presentation of somewhat repetitive, but not
identical data here in the body of the report, the normalized
recovery data are presented in the appendix.
4.1.4 Statistical Analysis of Recovery Trends and
Decontamination Efficacy
Statistical analysis was performed using recovery values
(i.e., measured recovery results normalized to the Time
0 recovery) to evaluate whether there were statically
significant differences in the TICs recovered from treated
coupons compared to controls. For each TIC and material
combination, an analysis of variance (ANOVA) model was fit
to the data. The form of the model was:
Y = n + treatment + time( + (treatment * time() + e
(11)
where Y is the recovery for the rth coupon (/'=! to 5) in
treatment group j (with or without decontamination) at time
t (1, 3, or 7 h), \a is an overall constant, and e.., is the random
error left unexplained by the model, assumed to be normally
distributed with mean 0 and variance o2.
The model included a fixed effect for type of treatment (with
or without decontamination) and time (number of hours
after application). A treatment by time interaction effect
was also included if significant. For the TIC analysis, the
interaction term was significant for all TIC and material
combinations except for DMMP on carpet. The model above
was fit in S AS® v9. 1 using PROC GLM. Model diagnostics
were evaluated to assure that there were no outliers, that the
residuals were approximately normally distributed about the
effect means, and that the assumption of consistent variance
Table 4-5. Mean Recovery of TICs from Building Materials over Time
With and Without Sabre CI02 Fumigant Decontamination
Mean Recovery, % ± SDa
TIC Material Time (n) Without With
decontamination decontamination
DMMP
Malathion
Carpet
Ceiling tile
Carpet
Laminate
Oh(n=3)
1 h (n = 5)
3 h (n = 5)
7 h (n = 5)
Oh(n =3)
1 h (n = 5)
3 h (n = 5)
7 h (n = 5)
Oh(n =3)
1 h (n = 5)
3 h (n = 5)
7 h (n = 5)
Oh(n=3)
1 h (n = 5)
3 h (n = 5)
7 h (n = 5)
97 ±4
32 ±2
24 ±3
15±3
102 ±2
40 ±4
28 ± 1
13 ± 1
112± 15
100 ±3
102 ±3
98 ± 1
125 ± 1
126 ± 15
115±7
118± 13
84±9b
36 ±5
24 ±5
19 ±3
105 ± lb
31 ±2
16 ± 1
9.1 ±0.4
95±3b
72 ±3
49 ±4
23 ±2
105 ± 16b
30 ±20
3.4 ± 1.6
0.4 ±0.2
a) Calculated using Eq (6)
b) 0 h coupons were positive controls that were not exposed to fumigation
-------
for the groups was appropriate. For DMMP on ceiling tile
and malathion on laminate, the variability in recovery from
replicate coupon at a given treatment and variability in time
means was too large to meet the homogeneous error variance
assumption of the model above. Instead, these data were fit
to a heterogeneous variance model with the same effects but
with separate variance terms for each treatment group and
time. This model was fit in PROC MIXED.
From the model output, mean recovery and corresponding
confidence intervals were calculated for each treatment and
time. The difference (and corresponding confidence interval)
in mean recovery between the two treatments (with and
without decontamination) was calculated at each time point.
The three time point comparisons between the treatment
groups were evaluated at joint 95% confidence. Using a
Bonferroni approach, this resulted in each separate time point
comparison being made at 98.3% confidence. The exception
here was for DMMP on carpet, where the lack of a significant
interaction term meant that a single estimated comparison
(at 95% confidence) of with and without decontamination
applied to all three time points.
Table 4-6 shows the statistically-modeled, Eq (11), recovery
results and the resulting calculation of decontamination
efficacy. For the ANOVA modeling, the recovery data.
shown in Table 4-6, were normalized relative to the
recovery from the 0 h positive control, i.e., the recovery
for 0 h positive controls is assumed to be 100%. The
decontamination efficacy values are calculated with reference
to the equivalent recovery in the humid air control chamber
tests so as to account for the efficacy due solely to the
fumigant C1O2. Those results where there was a statistically
significant difference between decontamination and without
decontamination are highlighted in bold.
Recoveries of DMMP from carpet coupons with and without
exposure to the Sabre C1O2 fumigant decontamination
steadily declined over the 7 h of testing. There were no
statistically significant differences between recoveries of
DMMP from carpet coupons without decontamination
compared to the recoveries from coupons with
decontamination. No decontamination efficacy was observed
for C1O2 fumigation of DMMP on carpet.
Recoveries of DMMP from ceiling tile coupons with
and without exposure to the Sabre C1O2 fumigant
decontamination steadily declined over the 7 h of testing. The
Sabre decontamination recoveries were slightly lower than
the recoveries without decontamination, and the difference
was statistically significant at each test point. Although
statistically significant, the small differences may not be of
practical significance.
For malathion on carpet, the Sabre C1O2 fumigant
decontamination resulted in a steady decline from 76% at 1
h to 24% recovery at 7 h. Coupons without decontamination
showed very little reduction in recovery over the 7 h of
testing, with recoveries being in excess of 90% over this time
period. The Sabre decontamination recovery was therefore
statistically significantly less than the recovery without
decontamination for all three time periods.
Table 4-6. Statistically Modeled, Eq (11), Percent Recovery of TICs With and Without Sabre CI02
Fumigant Decontamination and Decontamination Efficacy
TIC
DMMP
Malathion
Material
Carpetb
Ceiling tiled
Carpetd
Laminated
Time
1 h
3 h
7 h
1 h
3 h
7 h
1 h
3 h
7 h
1 h
3 h
7 h
Recovery, % (Confidence Interval) Decontamination
Without decontamination With decontamination Efficacy, %a
35 (33-38)
23 (20-25)
15(13-18)
39 (30-48)
28 (25-30)
12 (10-15)
89 (86-93)
90 (87-94)
87 (83-91)
100(80-121)
91 (82-101)
94(76-112)
42 (40-45)
30 (27-32)
22 (20-25)
30 (26-34)
15(14-16)
8.7(8.1-9.3)
76 (73-80)
51 (48-55)
24 (20-27)
28 (-4.7-61)
3.3 (0.5-6.0)
0.4(0.1-0.7)
No efficacy0
No efficacy
No efficacy
23% '
46%
28%
15%
43%
72%
72%
96%
99.6%
a) Calculated using Eq (8)
b) Intervals are 95% confidence
c) No decontamination efficacy demonstrated; recovery with decontamination greater than recovery without
decontamination
d) Intervals are 98.33% confidence to control error rate at 5% for all three time point comparisons of treatment to control
with the TIC and material combination
e) Decontamination efficacy shown in bold indicates a statistically significant difference in recovery with and without
decontamination (p < 0.05)
-------
For malathion on laminate, the Sabre C1O2 fumigant
decontamination resulted in a rapid decline from 28% at 1 h
to 0.4% recovery at 7 h. Coupons without decontamination
showed very little reduction in recovery over the 7 h of
testing, with recoveries in excess of 90% over this time
period. The Sabre decontamination recovery was therefore
statistically significantly less than the recovery without
decontamination for all three time periods. These trends are
shown graphically in Figures 4-1 and 4-2 for DMMP and
malathion, respectively.
60 -i
The Sabre C1O2 fumigation was shown to be more effective
at removing malathion than DMMP. Sabre C1O2 fumigation
exhibits a significant and potentially useful level of
decontamination efficacy against malathion on both carpet
and laminate. There were differences between carpet and
laminate in the C1O2 decontamination efficacy against
malathion. The results show that Sabre C1O2 exhibits little
or no potentially useful decontamination efficacy against
DMMP on either carpet or ceiling tile.
Carpet Ceiling Tile Carpet Ceiling Tile Carpet Ceiling Tile
1h 3h 7h
S Denotes persistance statistically significantly less with Sabre CI02fumigation decontamination
Figure 4-1. Statistically Modeled Percent Recovery Data for DMMP on Coupons With and
Without Sabre CI02 Fumigant Decontamination (Error Bars Show 95% Confidence Interval)
100
Carpet Laminate
•1 h
Carpet Laminate
3h
Q With Decon
D
Without Decon
S Denotes recovery statistically significantly less with Sabre CI02 fumigant decontamination
Figure 4-2. Statistically Modeled Percent Recovery Data for Malathion on Coupons With and
Without Sabre CI02 Fumigant Decontamination (Error Bars Show 95% Confidence Interval)
-------
4.1.5 TICs on Laboratory, Handling, and Procedural
Blank Coupons
There were three different types of blank coupons employed
in the investigation of the Sabre decontamination technology.
The laboratory blank coupon of each material was taken from
storage directly to an extraction vial and was not exposed to
the fume hood where the TICs were spiked. The handling
blanks were placed in the fume hood during sample spiking
and were then loaded into vials and sealed. At the designated
times when coupons were removed from the chamber, one
of these handling blanks was also removed from its sealed
vial for extraction. The procedural blanks were also placed in
the fume hood during sample spiking. These blank coupons
were then placed into the test chambers along with the spiked
samples, exposed to the fumigation treatment, and removed
at the designated times with deliberately spiked coupons.
The levels of the TICs on the laboratory, handling, and
procedural blank coupons are listed in Table 4-7. Due to
similarity in levels on the laboratory and handling blanks.
these data were averaged. Because the procedural blanks
show specific trends in the transfer and deposition of material
to clean surfaces in the chamber, the procedural blank
data are presented individually for the experiments with or
without decontamination, and also individually for each
sampling time.
As described previously, the laboratory and handling coupons
were never in the chamber, while the procedural coupons
were blank coupons that were placed in the chamber at
the beginning of an experiment. Any level of TIC found
on a procedural blank coupon above the handling coupon
blank level had to have arisen from redistribution of TICs
in the chamber. The levels of DMMP recovered from the
procedural blanks were approximately 10 fold higher than
the background levels recovered from the laboratory and
handling blanks. This appears to arise from volatilization of
DMMP from spiked coupons and redeposition onto initially
clean surfaces in the test or control chamber. Higher levels of
DMMP were recovered from ceiling tile (12.0 |j,g) than carpet
(3.41 |j,g) after being exposed to the fumigation treatment for
7 h. The recovery of DMMP from procedural blanks did not
impact recovery calculations. However, the contamination
was recorded and analyzed as an experimental result. No
corrective action was deemed necessary.
For malathion, which is considerably less volatile than
DMMP, the malathion detected on the procedural blanks was
within 2 SD of the mean mass measured on the handling and
laboratory blanks (not significantly different).
Because there was only one procedural blank coupon at
each sampling interval, it is difficult to determine whether
there was a statistical difference in the levels found on
these coupons for coupons collected with and without the
C1O2 decontamination — with exception of DMMP on the
ceiling tile coupons. The mass of DMMP recovered from
procedural blank coupons subjected to decontamination were
approximately two to three times lower than the amounts
recovered from similar procedural blank coupons that were
not subjected to decontamination.
Table 4-7. Mean TIC Levels on Laboratory, Handling, and Procedural
Blank Coupons
TIC Material Time Mean Mass on Lab/Handling Blanks, ug
DMMP
Carpet
Ceiling tile
0-7 h
0-7 h
0.20 ± 0.34 (n= 8)
0.16±0.04(n =8)
Malathion
Carpet
0-7 h
0.65 ± 0.43 (n= 8)
Laminate
0-7 h
0.06±0.06(n =8)
Mean Mass on Procedural Blanks, ug
Vithout
)econtamination Decontamination
DMMP
Carpet
Ceiling tile
1 h
2.14
3h
2.79
7 h
3.32
1 h
17.0
3 h
26.7
7h
31.87
1.81
2.35
3.41
6.22
11.7
12.0
1 h
1.37
0.97
Carpet
3h
1.24
1.00
Malathion
7 h
1.20
0.97
1 h
ND, <0.05a
0.07
Laminate
3 h
ND, <0.05
0.22
7 h
ND, <0.05
0.25
a) ND = not detected; less than stated MDL
-------
4.1.6 TICs in Chamber Air
The concentrations of the TICs in the vapor phase of the test
chamber under near-static conditions are listed in Table 4-8.
This air concentration is also listed as a percentage of mass
applied onto the coupons.
The relatively high air level of DMMP and the relatively
low air level of malathion, in the chambers that did not
undergo decontamination may be a reflection of the relative
vapor pressures of the two TICs. In the decontamination
chambers, though, it was evident that the fumigant C1O2
decreased the amount of both TICs in air. (The higher
DMMP concentrations are observed even though the air
exchange for malathion was high [one air exchange Iv1
conditions] compared to the static air conditions for DMMP.)
This effect was more pronounced for DMMP since the
vapor phase levels were detectable both with and without
decontamination. It was interesting to note that with DMMP.
fumigant C1O2 had a greater decontamination effect on
vapor phase material than on the residues on the coupons.
The recovery of DMMP on the coupons was nearly identical
whether exposed to fumigant C1O2 or not. These air data
demonstrated that the fumigant appeared to markedly lower
the gas phase levels of DMMP. One caveat applied to the air
data collected during decontamination testing is that there
were no preliminary recovery tests carried out to determine
the stability of the TICs on Tenax® in the presence of vapor
phase C1O2. Thus, these air levels may represent only a
portion of the amount present in the chamber.
4.1.7 Detection of Oxidized Malathion Product on
Coupons
To evaluate the usefulness of a decontamination procedure
against chemical compounds, it is important to determine
whether toxic by-products are produced. The extracts of
selected carpet and laminate samples were analyzed for the
presence of one potential oxidation product of malathion.
malathion oxon (or maloxon). In maloxon, the P=S portion of
the thiophosphate group on malathion is replaced with P=O.
so that the molecule becomes an organophosphate, rather
than a thiophosphate. Organophosphates are generally more
toxic than the corresponding thiophosphate.
One of each of the three positive controls (spiked with
malathion, but not decontaminated) for each coupon type
(carpet and laminate) was removed from the chamber
after 1,3, and 7 h and analyzed. Recoveries of maloxon
were assumed to be the same as for malathion. Moreover.
calibration curves were prepared from maloxon standards and
other QC criteria were similar to those applied to the TICs.
Maloxon was not detected on any of the positive controls.
In addition, three carpet and three laminate test coupon
extracts were similarly analyzed (one each from the 1, 3, and
7 h decontamination tests). For both coupon types, maloxon
was detected at levels corresponding to malathion to maloxon
conversion percentages that increased in linear proportion to
the log of the concentration multiplied by contact time (CT)
of C1O2 to which the coupons were exposed. The results are
presented in Table 4-9 and Figure 4-3 (on page 24).
Table 4-8. Mean Concentration of TICs in Test Chamber Air
Mean Air Concentration, ug m3 (% of Mass Applied)3 b
TIC Material Time
Without Decontamination (n = 2) With Decontamination (n = 2)
DMMP
Malathion
Carpet
Ceiling tile
Carpet
Laminate
0-1 h
0-1 h
1-3 h
1-3 h
1740 ± 13(10%)
1430 ± 253 (9%)
0.42 ± 0.00 (<0.01%)c
0.42 ± 0.00 (<0.01%)c
ND, <3.6 (n = 1) (0.7%)
484 (n = 1) (2.9%)
ND, <0.26 (<0.01%)
ND, <0.26 (<0.01%)
a) Air concentration expressed as percentage of total amount of TIC in the chamber if there had been no hydrolysis or
degradation of the TIC
b) Calculated using Eq (10)
c) This sample was collected from the test chamber under one air exchange tr1 conditions; samples were not collected
from the test chamber under near-static conditions
Table 4-9. Mean Mass of Malathion Oxidized to Maloxon by Coupon Type and CIO CT
_ T Decontamination CI02 CT, Maloxon Malathion % Malathion
Time, h ppm-h Recovered, ug Converted, ug Oxidized to Maloxon
Carpet
Laminate
1
3
7
1
3
7
3366
9494
21,343
2926
8941
21,076
78
143
205
215
297
338
82
150
215
226
312
355
16
30
43
45
62
71
-------
rj
O
o
0
uu
Hn
~7n
fu
60
A n
4U
qn_
9fl-
^-* I
y = 1 3.24Ln(x) - 59.802 ^~~^ " Latr
R2= 0.9864 i^^"^ + Car
^*^ Log
..-+
...-•" y= 14.335Ln(x)- 100.38
...-•'' R2= 0.997
imate
set
(Laminate)
(Carpet)
*""
i
1000 10,000
100,0
: Dosage, ppm-h
Figure 4-3. Percent Oxidation of Malathion to Maloxon by Coupon Type and CI02 Dosage
4.1.8 Condition of CICyTreated Coupons after Six
Months
To assess the effect of fumigant C1O2 on long-term stability
and integrity of building materials, eight laboratory blanks
(not exposed to C1O2) and 24 procedural blanks (exposed to
C1O2) were maintained under ambient temperature and RH
conditions in open vials for extended periods of time. After
three months, and then again after six months, the coupons
were inspected to assess any changes in appearance or
structural integrity. The coupons were relatively unchanged
after three months, and the carpet showed some minor
"bleaching" after six months.
4.2 Results for Fumigant CI02
of CWAs
Decontamination
4.2.1 Analytical Method Results
The recoveries of the CWAs spiked onto the various coupon
materials were not replicated after the initial assessment. The
analytical method recoveries listed in Table 4-10 (on page
25) are those that were determined prior to decontamination
tests, and are presented here as a baseline against which the
subsequent decontamination recovery data can be compared.
The approximate MDLs for the CWAs are listed in Table
4-11 (on page 25). Each MDL was estimated based on the
peak area of the lowest calibration standard, the signal to
noise ratio for this concentration, and the peak area that
could be integrated reliably for any peak with the data
system.
4.2.2 Recovery over Time of CWAs on Building
Materials With and Without Sabre CI02 Fumigant
Decontamination
Side-by-side chambers were installed in the laboratory for
these tests, with one chamber (test) having fumigant C1O2
decontamination and the matching chamber (control) having
no C1O2. The side-by-side design was adopted because of the
high volatility of GB and TGD; with this design, any slight
variation in laboratory temperature would be eliminated as a
variable when assessing decontamination efficacy. A single
spiked coupon of each type was analyzed directly after
spiking to establish the baseline recovery. The recoveries of
the CWAs from coupons at initiation (0 h) and at subsequent
times (1 h and 2 or 4 h, depending on CWA) are listed in
Table 4-12 (on page 25). The recoveries at initiation (0 h)
for TGD from carpet, laminate, and ductwork, and these
recoveries of VX from carpet and ductwork do not agree
very well with the recoveries shown in Table 4-10 (on page
25) (analytical method recoveries). In theory, these two
values should have been very similar. At this time, there is
no apparent reason for the difference, and the difference is
simply noted here.
In many cases there were measurements made close to the
detection limit. When there was at least one detected sample
out of the five replicates, the median value of the sample
set is reported along with the range, where a non-detectable
value is reported as less than the MDL. When all five samples
had non-detectable values, the percent recovery was reported
as less than the MDL. In addition, the number of non-
detected samples in each set of five replicates is also
noted in Table 4-12 (on page 25).
-------
Table 4-10. Mean Recovery of CWAs and SRS from Building Materials Table 4-11. MDLs for CWAs
Evaporation . Mean Recovery from ^ SRS/CWA
Carpet
Carpet
Laminate
Metal ductwork
Carpet
Laminate
Metal ductwork
^^Q
7 min
5 min
5 min
5 min
5 min
5 min
5 min
GB
91 ± 12
TGD
88 ± 18
97 ±8
98 ± 11
VX
113 ±9
107 ±6
110±6
TBP
87 ± 14
TBP
98 ± 11
89 ±9
88 ± 10
TBP
103 ±21
93 ± 14
94 ± 15
•jjfSM
0.96
1.11
0.92
0.90
0.91
0.87
0.85
a) Calculated using Eq (4)
b) Calculated using Eq (5)
Table 4-12. Average Recovery of CWAs from Building Materials over Time With
and Without Sabre CI02 Fumigant Decontamination
GB
TGD
VX
Carpet
Carpet
Laminate
Ductwork
Carpet
Laminate
Ductwork
Oh (n = 1)
1 h (n = 5)
4 h (n = 5)
Oh(n = 1)
1 h (n = 5)
2 h (n = 5)
Oh(n = 1)
1 h (n = 5)
2 h (n = 5)
Oh (n = 1)
1 h (n = 5)
2 h (n = 5)
Oh(n = 1)
1 h (n = 5)
4 h (n = 5)
Oh (n = 1)
1 h (n = 5)
4 h (n = 5)
Oh (n = 1)
1 h (n = 5)
4 h (n = 5)
Without Decontamination With Decontamination
87
5.1 ±0.9 (0)c
4. 9 ±0.2 (0)
52
40 ±8. 7 (0)
30 ±3. 2 (0)
89
1.3 (<0.1 -3.6) (l)e
0.4(<0.1-3.2) (2)e
73
16 ±5.5 (0)
4.3± 1.9 (0)
74
77 ± 23 (0)
72 ±7.1 (0)
94
81 ±7.6 (0)
88 ±5.1 (0)
84
84 ±2. 6 (0)
85 ± 7.6 (0)
NAb
3.6± 1.8 (0)c
2. 8 ±0.9 (0)
NA
15 ±2. 7 (0)
19 ±8.8 (0)
NA
5. 4 ±6.1 (0)
<0.1 (0.1-0.3)(4)e
NA
10 ±7. 9 (0)
5. 4 ±6. 5(0)
NA
<0.7 (5)d
<0.7 (5)d
NA
<0.7 (5)d
<0.7 (5)d
NA
<0.7 (5)d
<0.7 (5)d
a) Calculated using Eq (6) except where non-detects occur, and only the median and range of
values is shown
b) NA = not applicable; Time 0 coupons did not have decontamination
c) The number of non-detects is shown in parenthesis
d) All coupons less than stated MDL; value shown is MDL as percentage of mass applied
e) Non-detects occurred - only the median and range of values is shown
-------
Measurable amounts of all CWAs were removed from
all building materials with the fumigant C1O2 treatment.
However, for both GB and TGD, there was also a substantial
decrease in the amount found on coupons that did not
undergo C1O2 treatment. In these cases, statistical analysis
of the data was the only way to accurately assess whether
the slight differences could be attributable to the C1O2
treatment. Other physical processes such as volatilization
and redistribution to the chamber and/or aqueous hydrolysis
could account for the losses, as both chambers were operated
under conditions of 80% RH. The case for VX is much more
straight-forward. The VX persisted on the coupons which
did not have decontamination treatment, in spite of the high
RH of the control chamber; in contrast, VX was not detected
on the coupons that were treated with the fumigant C1O2
process.
The recoveries of GB, TGD, and VX, normalized to the
recoveries at 0 h, are presented in Appendix A, Table A-2
of this report. These normalized data were used as the input
to the statistical analyses described below. The statistical
analyses are more informative than the normalized recovery
data, in that they provide insight into whether there is a
statistical difference between recoveries with and without
decontamination; therefore, to avoid presentation of
somewhat repetitive data here in the body of the report, the
normalized recovery data are presented in the appendix.
4.2.3 Statistical Analysis of Recovery Trends and
Fumigant Decontamination Efficacy
Statistical analysis was performed using recovery values
(i.e., measured recovery results normalized to the Time 0
recovery) to evaluate whether there were statically significant
differences in the CWA recovered from treated coupons
compared to controls. Because the recoveries are normalized
to the Time 0 recovery value, it is possible to obtain recovery
values that are greater than 100%. Values above 100% were
left as calculated, though they should be interpreted in the
context they were created, and not truly greater than 100%.
For CWA and material combinations with detectable
recoveries on coupons with and without decontamination,
an ANOVA model was fit to the data. The model included
a fixed effect for type of treatment (with or without
decontamination) and time (number of h after application).
A treatment by time interaction effect was also included if
significant. The model was fit in SAS® v9.1 using PROC
MIXED. Model diagnostics were evaluated to assure that
there were no outliers, that the residuals were approximately
normally distributed about the effect means, and that the
assumption of consistent variance for the groups was
appropriate.
From the model output, mean recovery and corresponding
confidence intervals were calculated for each treatment and
time. The difference (and corresponding confidence interval)
in mean recovery between the two treatments (with and
without decontamination) was calculated at each time point.
The two time point comparisons between the treatment
groups were evaluated at joint 95% confidence, so each
separate comparison is essentially approximately 97.5%
confidence.
For CWA and material combinations with non-detects, e.g.,
TGD on laminate, the ANOVA approach is not appropriate
since the data are censored and may violate the assumptions
of normality and constant variance. Instead, a less constrained
non-parametric analysis was performed. Consistent with this
type of analysis, median (rather than mean) recovery was
reported for each treatment and time point. In cases where the
median value was a non-detect (i.e., three or more of the five
replicate coupons were non-detects), the results are reported
as "ND
-------
Table 4-13. Statistically Modeled, Eq (11), Percent Recovery of CWAs With and Without
Sabre CIO Fumigant Decontamination and Decontamination Efficacy
CWA Material Time
GB
TGD
VX
Carpetb
Carpetb
Laminated
Ductworkb
Carpetd
Laminated
Ductworkd
1 h
4h
1 h
2h
1 h
2 h
1 h
4h
1 h
4h
1 h
4h
1 h
4h
Recovery, % (Confidence Interval)
Without Decontamination With Decontamination
6.1 (5.1-7.1)
5.6(4.5-6.6)
77 (63-91)
57 (43-71)
1.4
0.5
21 (13-30)
5.9(0-15)
102
97
89
91
99
105
4.0(2.9-5.0)
3.4(2.1-4.8)
29 (15-43)
36 (21-50)
3.2
ND, <0.1f
14 (4.7-23)
7.4(0-16)
ND, <0.7
ND, <0.7
ND, <0.7
ND, <0.7
ND, <0.7
ND, <0.7
Decontamination
Efficacy, %a
34%c
39%
62%
37%
No efficacy6
No efficacy6
No efficacy6
No efficacy6
>99%
>99%
>99%
>99%
>99%
>99%
a) Calculated using Eq (6)
b) Reported values are least square means; intervals are joint 95% confidence across time points for the combination
of the CWA and the material
c) Decontamination efficacy shown in bold indicates a statistically significant difference in recovery with and without
decontamination (p < 0.05)
d) Reported values are medians; ND denotes median value is a non-detect; difference interval is Hodges-Lehmann
median shift interval of 95% confidence
e) No decontamination efficacy demonstrated; recovery with decontamination greater than recovery without
decontamination
f) ND = not detected; less than coupon MDL converted to percentage of spike amount
g) No statistically significant decontamination demonstrated
For TGD on laminate, and VX on carpet, laminate, and
metal, the results were obtained with the nonparametric
approach described above.
TGD on laminate exhibited almost complete removal at
both time points for coupons with and without fumigant
C1O2 decontamination. No statistically significant difference
in recovery was observed for the decontaminated coupons
compared to the coupons without C1O2 decontamination at
time 1 h (p=0.15) or at 2 h (p=0.17).
For VX on carpet, metal, and laminate, the coupons
without decontamination exhibited very high recovery at
each time point, while the fumigant C1O2 decontaminated
coupons showed removal to below the MDL in every case.
With p-values of 0.008 for each comparison, the recovery
under the fumigant C1O2 decontamination was statistically
significantly less than the recovery without decontamination.
The estimated median differences in fumigant C1O2
decontamination reduction of VX recovery range from
88 percentage points (VX, laminate, 1 h) to 105 percentage
points (VX, metal, 4 h).
The recovery of GB, TGD, and VX on the different
materials is shown graphically in Figures 4-4, 4-5,
and 4-6, respectively (on pages 28 amd 29).
-------
100
10
D With Decon D Without Decon
o
u
&
S Denotes recovery statistically significantly less with Sabre CI02
Figure 4-4. GB Recovery (%) on Coupons With and Without Sabre CI02
Fumigant Decontamination (Error Bars Show 95% Confidence Interval)
100
HI
a:
• With Decon • Without Decon
0.01
Carpet Larrinate Metal Carpet Larrinate Metal
1 h 2h
Values shown without errors bars are medians rather than means due to nonparametric statistical
model fit to data with non-detects.
S Denotes recovery statistically significantly less with Sabre CI02 fumigant decontamination
Figure 4-5. TGD Recovery (%) on Coupons With and Without Sabre CI02
Fumigant Decontamination (Error Bars Show 95% Confidence Interval)
-------
100
g
o
o
HI
OL
Larrinate
4h
Metal
Values shown are medians rather than means due to nonparametric statistical model fit to data with non-detects.
S Denotes recovery statistically significantly less with Sabre CI02 fumigant decontamination
Figure 4-6. VX Recovery (%) on Coupons With and Without Sabre
CI02 Fumigant Decontamination
-------
4.2.4 CWAs on Laboratory and Procedural Blank
Coupons
There were two different types of blank coupons employed
in the investigation of the Sabre decontamination technology.
The laboratory blank coupon of each material was taken from
storage directly to an extraction vial and was not exposed to
the fume hood where the TICs were spiked. The procedural
blanks were placed in the fume hood during sample spiking.
These blank coupons were then placed into the test chambers
along with the spiked samples, and removed at the designated
times with coupons. These procedural blanks were expected
to show the extent to which volatilization and redeposition
onto other surfaces may account for some of the losses
from coupons. Table 4-14 shows the amounts of the CWAs
measured on these different blanks.
As was observed with the TICs, the data in Table 4-14
indicate that GB and TGD, which are more volatile than
VX, may migrate from the initially spiked coupons to other
surfaces in the chamber, including the procedural blank
coupons. For TGD, the carpet coupons absorbed considerably
more than the laminate or ductwork coupons.
Concentration of CI02 in Test Chambers
4.2.5
The concentrations of C1O2 measured in the test chambers
during decontamination of the three different CWAs are listed
in Table 4-15.
Table 4-14. Comparison of Mean CWA Levels on Laboratory
and Procedural Blank Coupons
Table 4-15. Mean Concentration of CI02 in Test
Chamber for CWA Decontamination Tests
GB
TGD
VX
Carpet
Carpet
Laminate
Ductwork
Carpet
Laminate
Ductwork
Oh
1 h
4h
0 h
1 h
2 h
Oh
1 h
2 h
Oh
1 h
2 h
Oh
1 h
4 h
0 h
1 h
4h
Oh
1 h
4h
Mean Mass on Blank, ug
Without
:ontaminati
Laboratory
Blank
ND,
-------
4.3 Results for Liquid Decontamination of CWAs
4.3.1 Recovery of CWAs from Liquid Decontamination
Solutions
Initial tests were conducted to determine whether CWAs
could be effectively recovered from liquid decontamination
solutions or from neutralized liquid decontamination
solutions. The purpose of the method development was to
determine (1) the maximum percentage of spiked CWA
that could be recovered from neutralized decontamination
solution and (2) the percentage of spiked CWA recovered
after 1 h in the decontamination solution. High recovery from
the neutralized decontamination solution would indicate that
the neutralization was effective and could be used to end the
loss/degradation of the CWA caused by the decontamination
solution. A high recovery of CWA from the decontamination
solution after 1 h of contact would suggest that the
decontamination was ineffective against the CWA.
Recovery of GB, TGD, and VX from 10% bleach (0.6%
hypochlorite) and 3000 ppm C1O2 was tested, as was
recovery from neutralized solutions that were created
via addition of sodium thiosulfate to the bleach or C1O2
solutions. CWAs were held in decontamination solutions for
1 h prior to addition of hexane for extraction and the addition
of the SRS; the CWAs were held in the neutralized solutions
for only 15 sec before addition of the SRS and hexane. The
SRS (TBP used as SRS for all three CWAs) was spiked into
the liquid decontamination or neutralized decontamination
solution just prior to addition of hexane. Contact time for the
SRS with any solution was approximately 15 sec. The results
of these tests are shown in Table 4-16.
This investigation produced unexpected results. Clearly,
the SRS is not completely similar to the CWAs in physico-
chemical properties. Either by virtue of slower aqueous
hydrolysis or greater solubility in hexane, it is almost
fully recovered from the neutralized solutions (neutralized
bleach and neutralized C1O2). In contrast, recovery of GB
and VX from neutralized solutions was less than 10%, and
in some cases these compounds were not recovered at all;
TGD recovery was approximately 50% from the neutralized
solutions. Volatilization from the aqueous solution, rapid
hydrolysis and/or poor extraction efficiency into hexane
might explain the low recoveries of the CWAs from the
neutralized solutions.
When comparing recoveries from neutralized solutions
versus oxidant solutions, it appears that bleach is clearly
effective in decontamination of TGD. The other cases are
somewhat ambiguous because of the potential effect of
aqueous hydrolysis or poor extraction efficiency. Because of
the possibility that the small, polar CWAs were preferentially
retained in the aqueous solution over the hexane, the
decision was made to conduct liquid decontamination tests
with coupons that would be removed from the liquid before
extraction. In addition, the decision was made to collect and
record the SRS recovery values (not reported here), but not to
correct CWA recovery by the SRS recovery, as they appear to
be quite dissimilar in recovery from aqueous media. The data
suggest that SRS correction for recovery would not be useful.
4.3.2 Recovery over Time of CWAs on Building Materials
With and Without Liquid Decontamination
The recovery of CWAs from building materials
immersed in bleach or C1O2 solutions was evaluated
using decontamination times of 10 to 30 min because the
results described in Section 4.3.1 suggested that complete
decontamination could be expected in less than 1 h. In the
preliminary solution tests described in Section 4.3.1, no
efficacy of liquid C1O2 was observed against GB or TGD.
Therefore, the decontamination of CWAs by liquid C1O2 was
only tested against VX. In parallel with the VX tests, spiked
coupons were placed in vials containing slightly acidic
water (pH = 4.5-7). This differed from the liquid bleach
decontamination testing where coupons not undergoing
decontamination were placed in sealed vials without any
liquid solution; initial tests showed that CWAs were largely
destroyed by, or not recovered from, aqueous solution.
Coupons without liquid C1O2 decontamination were placed in
slightly acidified water (pH comparable to the C1O2 solution)
with the hope that VX, which is known to be especially prone
to hydrolysis under basic conditions, might be more stable in
an acidic aqueous condition.
Table 4-16. Mean Recovery of CWAs and SRS from Liquid Decontamination Solutions
% ± SD (n = 3)
Solution - Hold Time Before SR'
Addition and Hexane Extractioi
GB
TGD
VX
Recovered From CWA Solut
TBP (from GB)
TBP (from TGD)
TBP (from VX)
Bleach - 1 h
ND,
ND,
ND, <0.7
96 ±3
96 ±2
97 ±2
Neutralized bleach - 15 sec
1 ±0
41 ± 10
ND, <0.7
87 ± 1
84 ±2
86 ±3
CI02- 1 h
±2
58 ±5
ND, <0.7
91 ±2
93 ± 1
89 ±3
Neutralized CICL - 15 sec
± 1
56 ± 10
ND, <0.7
75 ±8
70 ±23
60 ±3
ND = not detected
-------
Tables 4-17 and 4-18 show the recoveries of CWAfrom
spiked coupons without decontamination (i.e., stored in
sealed vials in air for the duration of the decontamination
period), in acidified water, or in decontamination solutions
(bleach or C1O2 solution), showed significant loss of VX;
and VX was not recovered from any coupon with liquid
C1O2 decontamination. Table 4-17 reports recoveries as a
percentage of the mass of CWA spiked onto the coupons.
Table 4-18 (on page 33) reports recoveries as a percentage
of the Time 0 recoveries from spiked coupons.
In all cases with 30 min of the decontamination regimen, the
CWAs were at non-detectable levels or had recoveries of less
than one percent. These are impressive results, especially for
the highly sorptive and large surface area carpet material.
While these results show that there are non-detectable levels
on the building material, this investigation did not explore
the mechanism of this disappearance. The CWA may have
undergone aqueous hydrolysis, in which case water might
have been an equally effective decontamination agent.
Alternatively, the CWAs may have been soluble in the bleach
but not degraded. To fully understand the ramifications
of this decontamination method, it may be necessary to
measure both the CWAs and their degradation products
directly (i.e., without extraction into organic solvent) in the
decontamination solution.
Because there was no effective control for the CWA solubility
and/or aqueous hydrolysis, the decontamination efficacy of
the bleach (specifically CIO") was not calculated. Similarly.
Table 4-17. Mean Recovery of CWAs from Building Materials after Various Treatments
as Percent of Mass Applied
CWA
GB
TGD
VX
Material
Carpet
Carpet
Laminate
Ductwork
Carpet
Laminate
Ductwork
••
0 min
10 min
20 min
30 min
0 min
10 min
20 min
30 min
0 min
10 min
20 min
30 min
0 min
10 min
20 min
30 min
0 min
10 min
20 min
30 min
0 min
10 min
20 min
30 min
0 min
10 min
20 min
30 min
Mean CWA Recovery, % of Mass Applied6 ± SC
Without A^iHifi^ wuot^K Bleach
Decontamination flciamea water Decontamination
75 ±5
70 ±5
66 ±7
63 ± 1
38 ±4
46 ± 13
53 ± 12
51 ± 18
62 ± 13
56 ±6
48 ±9
47 ± 10
38 ± 15
49 ± 15
41 ± 17
34 ± 11
84 ±5
NTC
NT
90 ±6
100 ±5
NT
NT
102 ±6
101 ± 10
NT
NT
96 ± 11
0)
1
£
c
03
a.
£* 00
oj to
~0 0
0) O3
~O _0)
o
"o
0)
1
"o
77 ±4 a
4.5 ± 1.1
1.7 ±0.4
2. 5 ±0.5
111 ±7
2.1 ±0.2
2.8 ± 1.4
1.9 ±0.3
105 ±8
0.5 ±0.2
ND, <0.07
ND, <0.07
NAa
ND, <0.1b
ND, <0.1
ND, <0.1
NA
ND, <0.1
ND, <0.1
ND, <0.1
NA
2.3 ± 1.2
1.1 ± 1.0
0.1 ±0.03
NA
1.2 ±0.7
0.2 ±0.3
0.3 ±0.5
NA
ND, <0.7
ND, <0.7
ND, <0.7
NA
ND, <0.7
ND, <0.7
ND, <0.7
NA
ND, <0.7
ND, <0.7
ND, <0.7
Decontamination
int efficacy observed in
ing.
0 «
•f: 3
Not tested because no sign
solution
NAb
ND, <0.7d
ND, <0.7
ND, <0.7
NA
ND, <0.7
ND, <0.7
ND, <0.7
NA
ND, <0.7
ND, <0.7
ND, <0.7
a) NA = not applicable; no decontamination done at Time 0
b) ND = not detected; less than MDL converted to equivalent percentage
c) NT = not tested at this time interval
d) n = 4; one outlier
e) Recovery for individual coupons calculated in the same manner as analytical method recovery efficiency, Eq (4) for each
individual coupon. The mean and SD of the recoveries for a given condition are reported here.
-------
there was no statistical modeling of the recovery. The raw
data do, however, suggest that the decontamination with
aqueous bleach was very effective.
Because the initial tests showed that CWAs were largely
destroyed by, or not recovered from, aqueous solution, the
control samples (without decontamination) were not placed
in a liquid solution. Instead, the spiked coupons that did not
receive the decontamination technology were allowed to
remain in the laboratory fume hood until they were retrieved
for extraction. Coupons were withdrawn from the liquid
decontamination bath and the fume hood at nearly the same
times. The wet coupons were shaken lightly to remove
excess liquid, spiked with the SRS and then placed in the
extraction vial with hexane for extraction. The recoveries
of the CWAs mirrored the results obtained for the initial 1 h
decontamination tests in that the CWAs were not recovered.
However, this time, solubility in water cannot be suspected as
the cause.
The recoveries of VX on these coupons, normalized to the
recoveries at 0 h, are presented in Appendix A, Table A-3
of this report. These normalized data were used as the input
to the statistical analyses described below. The statistical
analyses are more informative than the normalized recovery
data, in that they provide insight into whether there is a
statistical difference between recoveries with and without
decontamination; therefore, to avoid somewhat repetitive
data, the normalized recovery data are presented in the
appendix.
Table 4-18. Mean Recovery of CWAs from Building Materials After Various Treatments
as Percent of TO Recovery
CWA
GB
TGD
VX
Material
Carpet
Carpet
Laminate
Ductwork
Carpet
Laminate
Ductwork
10 min
20 min
30 min
10 min
20 min
30 min
10 min
20 min
30 min
10 min
20 min
30 min
10 min
20 min
30 min
10 min
20 min
30 min
10 min
20 min
30 min
Normalized CWA Recovery, % of Mass Recovered at Time Od ± SD (n = 5)
Without Acidified Bleach CIO.,
Decontamination Water Decontamination Decontamination
93 ±7
88 ± 9
84 ±2
121 ±33
139 ±31
135 ±47
90 ± 10
77 ± 14
77 ± 16
130 ±39
109 ±44
91 ±28
NTb
NT
107 ±7
NT
NT
102 ±6
NT
NT
95 ± 11
<
a
W
£J 00
C -J3
Not tested; not included
plan for baseline tes
ND, <0.7
ND, <0.7
ND, <0.7
ND,
1.1
TO -o
0 OJ
CD >
~O (/)
CD _Q
to °
0) >^
-^ o
•4-1 TO
0 0
z it
0)
ND, <0.7C
ND, <0.7
ND, <0.7
ND, <0.7
ND, <0.7
ND, <0.7
ND, <0.7
ND, <0.7
ND, <0.7
a) ND = not detected, less than MDL converted to equivalent percentage
-------
4.3.3 Statistical Analysis of Recovery Trends and Liquid
Decontamination Efficacy
Statistical analysis was performed on the recovery results.
For this investigation, liquid C1O2 decontamination was
compared to coupons in acidified water. The residual VX on
the surfaces was below the MDL for some samples without
decontamination. For coupons with decontamination, VX
was below the detection limit for every material (carpet,
laminate, and ductwork). For this type of data, it is not
appropriate to fit an ANOVA statistical model, since the
data are censored and may violate the assumptions of
normality and constant variance in an ANOVA. Instead, a
less constrained non-parametric analysis was performed.
Consistent with this type of analysis, median (rather than
mean) recovery was reported for each treatment, material,
and time point. In cases where the median value was a non-
detect (i.e., three or more of the five replicate coupons were
non-detects), the results are reported as "ND 86%a
>68%
>79%
>63
>63
>59
lndeterminateb
Indeterminate
Indeterminate
a) Calculated using Eq (8); values given in bold are statistically significant
b) Efficacy cannot be determined due to non-detects both with and without decontamination
-------
Carpet Laminate Metal Carpet Laminate Metal Carpet Laminate Metal
10 min 20 min 30 min
Values shown are medians rather than means due to nonparametric statistical model fit to data with non-detects
S Denotes recovery statistically significantly less with Sabre CI02 fumigant decontamination
*Denotes non-detect, ¥2. MDL used in place of 0 in the analysis
Figure 4-7. Statistical Analysis Results of VX Decontamination with Liquid CIO
Table 4-20. CWA Levels on Laboratory and Procedural Blank Coupons with Bleach Decontamination
r-n
TGD
\/Y
Laminate
ImQu
Laboratory
Procedural
Laboratory
Procedural
Laboratory
Procedural
Procedural
Laboratory
Procedural
0 min
10, 20, 30 min
0 min
10, 20, 30 min
0 min
10, 30 min
20 min
0 min
10, 20, 30 min
ND,
-------
after the liquid bleach decontamination tests. The levels of
CWAs on the procedural blank coupons acquired during the
liquid C1O2 decontamination tests are listed in Table 4-21. VX
was not detected on any of the procedural blanks.
After liquid decontamination of the test coupons, the
decontaminated coupons were visually inspected; and any
obvious changes in the coupon surfaces were recorded.
When testing decontamination with liquid bleach, the control
condition consisted of exposure to room air; when testing
decontamination with liquid C1O2, the control condition
consisted of exposure of coupons to slightly acidic water.
No damage or visible change to any of the carpet, laminate
or ductwork test coupons was observed comparing extracted
laboratory blank coupons (not exposed to decontamination)
to extracted procedural blank coupons (exposed to
decontamination) directly after decontamination treatment.
Coupons were not re-examined after 3 and 6 months for this
part of the effort.
Table 4-21. VX Levels on Procedural Blank Coupons with Liquid CI02 Decontamination
| VX
Carpet, Laminate, Ductwork
10, 20, 30 min
Mass, ug
ND, <7a
a) ND = not detected, less than stated MDL
-------
5.0
Summary
The effectiveness of fumigant C1O2, liquid C1O2, and
liquid bleach decontamination technologies for selected
TICs and/or CWAs was evaluated on building materials.
specifically, carpet, ceiling tile, laminate, and ductwork.
Table 5-1 is a summary of the recovery and decontamination
efficacy results at the longest time interval tested for each
combination of decontamination technology, TIC, or CWA.
and building material. The recovery of each chemical is
listed in this table for samples with decontamination or
under positive control conditions (without decontamination
or in acidified water). For instance, fumigant C1O2
decontamination was operated at about 80% RH;
therefore the control conditions without decontamination
consisted of coupons exposed to 80% RH air. Liquid C1O2
decontamination was in aqueous solution; therefore the
control conditions without decontamination consisted of
coupons in acidified water. With this design, comparing
the recovery of a given chemical under the two conditions
allowed assessment of the effectiveness of the C1O2 itself.
rather than the combined effects of C1O2 and water (liquid or
vapor), as many of these chemicals are susceptible to aqueous
hydrolysis.
For the fumigant C1O2 decontamination, it appears that this
technology was most effective against malathion and VX.
Because these chemicals persisted to a large extent in the
presence of high RH under control conditions, it was clear
Table 5-1. Summarization of Percent Recovery With and Without Decontamination
Technologies, and Decontamination Efficacy for TICs and CWAs
\Barn
Fumigant
CI02
3000 ppm
Liquid
Bleach
Liquid
CI02
RTC
Malathion
DMMP
GB
TGD
VX
GB
TGD
VX
VX
Carpet
Laminate
Carpet
Ceiling tile
Carpet
Carpet
Laminate
Ductwork
Carpet
Laminate
Ductwork
Carpet
Carpet
Laminate
Ductwork
Carpet
Laminate
Ductwork
Carpet
Laminate
Ductwork
Without Decon
7h
7 h
7h
7 h
4h
2h
2h
2 h
4h
4h
4h
30 min
30 min
30 min
30 min
30 min
30 min
30 min
30 min
30 min
30 min
878
948
198
9.48
5.68
578
0.58
5.98
978
918
1058
84h
135h
77h
91h
107h
102h
95h
3.38
1.78
<0.7C
With Decon
248
0.48
228
8.78
3.48
368
<0.1C
7.48
<0.7C
<0.7C
<0.7C
<0.1C
0.1
0.2
0.8
<0.7C
<0.7C
<0.7C
<0.7C
<0.7C
<0.7C
Decontamination
Efficacy %a
72
99.6
No efficacyb
8
39
37
No efficacy*
No efficacyb
>99
>99
>99
Complete or nearly
complete loss
of recoverable
CWA from
treated compared
with untreated
couponsd
>86
>59
NAe
a) Statistically significant decontamination efficacy between recovery with and without decontamination test
conditions
b) No efficacy; recovery with decontamination is greater than recovery without decontamination
c) Not detected; MDL expressed as wpercentage of mass applied to the coupon
d) While decontamination efficwacy may be high, efficacy due to the bleach effect cannot be distinguished from
the effect of the water hydrolysis or dissolution because the control coupons were stored in room air
e) NA= not applicable; cannot calculate decontamination efficacy when the analyte is not detected under either
the control or test conditions
f) No efficacy; no statistically significant differences in recovery with and without decontamination
g) Statistically modeled data using Eq (11)
h) Calculated using Eq (7)
-------
that the C1O2, and not water, was responsible for degradation
of these two compounds. The high volatilities of DMMP.
GB, and TGD make it difficult to assess the effectiveness of
fumigant C1O2 decontamination against these compounds.
However, for DMMP where air levels were measured in both
the test and control chambers, the considerable difference in
these air levels suggested that fumigant C1O2 was responsible
for decontamination of gas-phase DMMP; the loss of DMMP
from surfaces may still be due to some extent to either
volatilization or aqueous hydrolysis. However, the gas-phase
decontamination may be a powerful technique for highly
volatile compounds.
When testing with liquid bleach, the positive controls were
not placed in any liquid solution. Based on findings from
the bleach testing, the approach was revised for testing
with liquid C1O2. For liquid C1O2 the control condition
consisted of exposure of coupons to slightly acidic water.
The test coupons and associated controls of the liquid
decontamination tests were placed in sealed vials. The
temperature and RH of the laboratory hood in which these
vials were placed was typically about 24°C and 40% RH.
Shown in Table 5-2, recovery of GB, GD, and VX from
100% bleach (6% CIO') and 3000 ppm C1O2 was tested, as
was recovery from neutralized solutions that were created
via addition of STS to the bleach or C1O2 solutions. No
CWAs spiked directly into bleach were recovered after 1 h.
Likewise, little GB (1%) and no VX were recovered from
neutralized bleach after 15 sec. This testing was inconclusive
as to whether chemical degradation or inefficient recoveries
account for the loss of recoverable CWA. In contrast, a
relatively high recovery of GD from neutralized bleach
solution with no detectable recovery of GD after 1 h in
bleach indicates that bleach was effective against GD.
GB (8% of spike) and GD (58% of spike) were recovered 1
h after being spiked directly into 3000 ppm C1O2 solution.
Within the margin of error, the same amounts of GB (8%)
and GD (56%) were recovered from neutralized CIO
solution. This testing indicated that the C1O2 solution was not
efficacious against GB or GD. VX was not recovered from
neutralized C1O2 solution at 15 sec or from C1O2 solution
after 1 h. The efficacy of C1O2 solution against VX could not
be determined from these results.
For testing the two liquid decontamination technologies
against CWAs on coupons, a slightly different control
approach was used for each. When testing with liquid bleach.
the positive control coupons (without decontamination) were
not placed in any liquid solution — they were maintained in a
vial exposed to air. When testing with liquid C1O2, the control
condition (without decontamination) consisted of immersion
of coupons in slightly acidic water. The test coupons and
associated controls of the liquid decontamination tests
were placed in sealed vials. The temperature and RH of
the laboratory hood in which these vials were placed was
typically about 24°C and 40% RH.
The results of the testing, shown in Table 5-1, suggest that
liquid bleach may be a very effective decontamination agent.
as none of the CWAs were detected on the coupons after a
20 min soak in the bleach. However, the test was not without
confounding parameters. The solution tests, shown in Table
5-2 in which the CWA was directly spiked into bleach and
neutralized bleach showed that both GB and VX were not
detected in extracts of neutralized bleach, either because of
rapid hydrolysis or poor extraction efficiency from aqueous
solution into hexane. Thus the low measured recovery of
these CWAs from the decontaminated coupons may be due in
part to hydrolysis, volatilization, and/or partition of the CWA
into the aqueous bleach and poor extraction efficiency into
hexane. The low recovery of GD in bleach, combined with
the fact that nearly 50% was extracted from a neutralized
bleach solution, suggested that the bleach itself was effective
in decontaminating GD on the materials tested.
The long-term stability and integrity of building materials
after exposure to fumigant C1O2 indicated that the coupons
were relatively unchanged after three months, and the carpet
showed some minor "bleaching" after six months.
Table 5-2. Mean Recovery of CWAs Directly Spiked into Liquid
Decontamination Solutions
Addition and Hexane Extraction % ± SD (n = 3)
Bleach - 1 h
Neutralized bleach - 15 sec
CI02- 1 h
Neutralized CI02 - 15 sec
GB
ND, <0.1
1 ±0
8±2
8± 1
GD
ND, <0.1
41 ± 10
58 ±5
56 ± 10
VX
ND, <0.7
ND, <0.7
ND, <0.7
ND, <0.7
Recoveries calculated using Eq (1); mean and SD of recoveries shown here
ND = not detected
-------
6.0
References
1. Persistence of Toxic Industrial Chemicals and Chemical Warfare Agents on Building Materials
Under Conventional Environmental Conditions; (EPA/600/R-08/075); July 2008.
http://www.epa.gov/nhsrc/pubs/600r08075.pdf
2. Method 8000 "Determinative Chromatographic Separations" as part ofSW-846 Third edition.
3. Battelle, Quality Management Plan (QMP) for the Technology Testing and Evaluation Program
(TTEP); Version 2. January 2006.
4. Battelle, Test/QA Plan for the Systematic Evaluation of Technologies for Decontaminating
Surfaces Inoculated with Highly Hazardous Chemicals (Chemical Warfare Agents and TICs),
Manipulation of Environmental Conditions to Alter Persistence, Version 1. June 2005.
-------
-------
7.0
Appendix A
The recoveries of DMMP and malathion on carpet, ceiling
tile, and laminate, normalized to recoveries at Time 0 (0 h).
both with and without fumigant C1O2 decontamination, are
listed in Table A-l. Note that the statistical model that uses
these data incorporates all data from an experiment into one
model; in modeling that data, the averages at each time point
change slightly to fit the single model. Thus averages listed
in this table do not match precisely the averages listed in
Table 4-4.
The recoveries of GB, TGD, and VX on carpet, laminate and
ductwork, normalized to recoveries at Time 0 (0 h), both with
and without fumigant C1O2 decontamination, are listed in
Table A-2 (on page 42).
Note that the statistical model that uses these data
incorporates all data from an experiment; in modeling that
data, the averages at each time point change slightly to fit
one model. Thus averages listed in this table do not match
precisely the averages listed in Table 4-12.
The recoveries of VX on carpet, laminate and ductwork.
normalized to recoveries at Time 0 (0 h), with liquid C1O2
decontamination, are listed in Table A-2 (on page 42).
Note that the statistical model that uses these data
incorporates all data from an experiment; in modeling that
data, the averages at each time point change slightly to fit
one model. Thus averages listed in this table do not match
precisely the averages listed in Table 4-19.
Table A-l. Normalized Mean Recovery of TICs on Building Materials
over Time With and Without Sabre CI02 Fumigant Decontamination
1 1
TIC
DMMP
Malathion
Material
Carpet
Ceiling tile
Carpet
Laminate
Time (n)
1 h (n = 5)
3 h (n = 5)
7 h (n = 5)
1 h (n = 5)
3 h (n = 5)
7 h (n = 5)
1 h (n = 5)
3 h (n = 5)
7 h (n = 5)
1 h (n = 5)
3 h (n = 5)
7 h (n = 5)
iNoniidiizeu mean Recovery, /o ± su
Without Decon With Decon
33 ±2
24 ±3
16±3
39 ±4
28 ± 1
12 ± 1
89 ±2
90 ±3
87 ± 1
100 ± 12
91 ±5
94 ± 10
43 ±6
28 ±6
23 ±4
30 ±2
15± 1
8. 7 ±0.3
76 ±3
51 ±5
24 ±2
28 ± 19
3.3 ± 1.6
0.4 ±0.2
a) Calculated using Eq (7)
-------
Table A-2. Normalized Mean Recovery of CWAs on Building
Materials over Time With and Without Sabre CI02 Fumigant
Decontamination
CWA
GB
TGD
VX
Material
Carpet
Carpet
Laminate
Ductwork
Carpet
Laminate
Ductwork
g
1 h
4h
1 h
2h
1 h
2 h
1 h
2 h
1 h
4h
1 h
4h
1 h
4h
Normalized Mean Recovery, % ± SDb
Without Decon With Decon
5.9 ± 1.1
5.7 ±0.2
77 ± 17
57 ±6.1
1.5± 1.6
1.3 ± 1.6
21 ±7.5
5. 9 ±2. 6
105 ±31
97 ± 10
86 ±8.0
94 ±5. 5
101 ±3.1
101 ±9
4.1 ±2.1
3.3 ± 1.0
29 ±5. 2
36 ± 17
6.1 ±6.9
0.10±0.11
14 ± 11
7.4 ±9.0
0.35±0.00a
0.35 ±0.00
0.35 ±0.00
0.35 ±0.00
0.35 ±0.00
0.35 ±0.00
a) VX not detected in sample; value reported here is half the MDL
b) Calculated using Eq (7)
Table A-3. Normalized Mean Recovery of VX on Building
Materials With and Without Liquid CI02 Decontamination
Normalized Mean Recovery, % ± SDa
CWA Material Time
Without Decon With Decon
VX
Carpet
Laminate
Ductwork
10 min
20 min
30 min
10 min
20 min
30 min
10 min
20 min
30 min
5.8± 1.5
2. 2 ±0.5
3. 2 ±0.6
1.9 ±0.2
2.5± 1.2
1.8 ±0.3
0.4 ±0.2
ND, <0.07
ND, <0.07
ND, <0.7
ND, <0.7
ND, <0.7
ND, <0.7
ND, <0.7
ND, <0.7
ND, <0.7
ND, <0.7
ND, <0.7
a) Calculated using Eq (7)
-------
-------
&EPA
United States
Environmental Protection
Agency
PRESORTED STANDARD
POSTAGES FEES PAID
EPA
PERMIT NO. G-35
Office of Research and Development
National Homeland Security Research Center
Cincinnati, OH 45268
Official Business
Penalty for Private Use
$300
Recycled/Recyclable
Printed with vegetable-based ink on
paper that contains a minimum of
50% post-consumer fiber content
processed chlorine free
------- |