EPA/600/R-17/187 August 2017
www.epa.gov/homeland-security-research
United States
Environmental Protection
Agency
oEPA
Decontamination of Subway Infrastructure
Materials Contaminated with Biological
Spores Using Methyl Bromide
Office of Research and Development
National Homeland Security Research Center

-------
This page left intentionally blank

-------
EPA/600/R-17/187
May 25, 2017
Decontamination of Subway
Infrastructure Materials
Contaminated with Biological
Spores Using Methyl Bromide
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
l

-------
Disclaimer
The U.S. Environmental Protection Agency (EPA) directed and managed this work through
Contract Number EP-C-15-002 with Battelle. This study was funded through the Underground
Transport Restoration Program by the U.S. Department of Homeland Security Science and
Technology Directorate under interagency agreement (No. 7095866901). This report has been
peer and administratively reviewed and has been approved for publication as an EPA document.
The views expressed in this report are those of the authors and do not necessarily reflect the
views or policies of the Agency. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use of a specific product.
This work was funded under an Interagency Agreement (HSHQPM-14-X-00178) with the
Homeland Security Advanced Research Projects Agency of the Department of Homeland
Security, Science and Technology Directorate. The contents are the sole responsibility of the
authors and do not necessarily represent the official views of S&T, DHS, or the United States
Government.
Questions concerning this document or its application should be addressed to:
Shannon D. Serre, Ph.D.
Chemical, Biological, Radiological, and Nuclear Consequence Management Advisory Division
Office of Emergency Management
Office of Land and Emergency Management
U.S. Environmental Protection Agency
Mail Code E343-06
Research Triangle Park, NC 27711
919-541-3817
11

-------
Acknowledgments
Contributions of the following individuals and organizations to this report are gratefully
acknowledged:
United States Environmental Protection Agency (EPA)
Worth Calfee (EPA NHSRC)
Leroy Mickelsen (EPA OLEM CBRN CMAD)
Lukas Oudejans (EPA NHSRC)
Shawn Ryan (EPA NHSRC)
Joe Wood (EPA NHSRC)
US EPA Technical Reviewers of Report
Tim Boe (EPA NHSRC)
Elise Jakabhazy (EPA OLEM CBRN CMAD)
US EPA Quality Assurance
Eletha Brady Roberts
Ramona Sherman
Battelle Memorial Institute
111

-------
Executive Summary
The U.S. Environmental Protection Agency (EPA) is striving to protect human health and the
environment from adverse impacts resulting from the intentional release of threat agents. This
report provides the results of an assessment to determine the decontamination efficacy of methyl
bromide (MB) fumigant in inactivating Bacillus cmthracis (B.a.; causative agent for anthrax)
spores on materials typically found in subway system infrastructure. To facilitate future
decontaminations employing MB in a subway environment, this investigation focused on finding
efficacious conditions when using MB at temperatures that may be encountered in an
underground subway system (i.e., temperatures lower than used in previous studies).
This investigation focused on the decontamination of four types of common subway materials
(with and without simulated subway grime application): ceramic tile, painted carbon steel,
weathered concrete, and granite. Decontamination efficacy tests were conducted with spores of
virulent B.a. Ames and avirulent B.a. Sterne. Decontamination efficacy was quantified in terms
of log reduction (LR), based on the difference in the number of bacterial spores recovered from
positive control coupons and test coupons. Ten tests were conducted at a target concentration of
212 milligrams per liter (mg/L) MB, target temperatures of 4.5 or 10 degrees Celsius (°C), target
relative humidity (RH) of 50% or 75%, and contact times (CT) ranging from 2 to 9 days to
assess the effect of these operational parameters on decontamination efficacy.
Summary of Results
As seen in other similar fumigant evaluations1, the temperature, RH, and CT affect the efficacy
of MB against B.a. Ames. Table E-l shows the CT required to achieve >6 LR (a decontaminant
that achieves an LR value >6 is considered effective)2 on all materials tested for a given set of
fumigation conditions (temperature and RH). For example, a CT of 4 days was required to
achieve >6 LR of B.a. Ames on all materials when fumigating at 212 mg/L, 10 °C, and 75% RH.
This study corroborates the importance of RH when fumigating with MB. There were no tests in
which >6 LR of B.a. Ames was achieved on all materials when fumigating at 50% RH. When
fumigating at 50% RH, increasing the MB concentration, temperature, or CT generally did not
improve decontamination efficacy. In contrast, when fumigating at 75% RH, increasing the
temperature and CT improved efficacy. Application of grime to the test materials resulted in
longer required CTs to achieve >6 LR. Efficacy of MB on B.a. Sterne was evaluated against B.a.
Ames to assess the potential use of B.a. Sterne as a suitable surrogate for the virulent strain.
Statistical analysis found no significant difference in efficacy for ceramic tile, weathered
concrete, and an increased efficacy against B.a. Ames on painted carbon steel as compared to
B.a. Sterne.
Table E-l. CT Required to Achieve >6 LR of B.a. Ames on all Materials*
Target MB
Concentration
(mg/L)
* Materials tested were ceramic tile, painted carbon steel, weathered concrete, and granite.
a Detailed data from each test number can be referenced in Table A-l in Appendix A.
iv

-------
Impact of Study
This research provides information on the efficacy of MB fumigation to decontaminate subway
relevant materials that have been contaminated with B.a. spores. Such results may be useful in
the development of guidance to aid in deployment of MB fumigation after a release of B.a.
spores within an underground transportation system. These results will provide decision makers
with information for effective use of MB at temperatures lower than what has been previously
tested, which will facilitate its use in subway systems as well as other applications at cold
temperatures.
v

-------
Contents
Disclaimer	ii
Acknowledgments	iii
Executive Summary	iv
Abbreviations/Acronyms	x
1.0 Introduction	1
2.0 Procedures	2
2.1	Technology Description	2
2.2	Test Matrix	2
2.3	Biological Agents	2
2.4	Test Materials	3
2.5	Preparation of Coupons	4
2.6	Coupon Extraction and Biological Agent Quantification	5
2.7	Decontamination Efficacy	6
2.8	Statisical Analysis	7
2.9	Surface Damage	7
3.0 Fumigation Description and Procedures	8
4.0 Quality Assurance/Quality Control	11
4.1	Equipment Calibration	11
4.2	QC Results	11
4.2.1. Operational Parameters	11
4.3	Audits	12
4.3.1	Performance Evaluation Audit	12
4.3.2	Technical Systems Audit	12
4.3.3	Data Quality Audit	13
4.4	Quality Assurance Project Plan Deviations	13
4.5	QA/QC Reporting	13
4.6	Data Review	13
5.0 Summary of Results and Discussion	14
5.1	Effects of Test Materials on MB efficacy for B.a. Ames	14
5.2	Effect of Temperature on Efficacy of MB against B. anthracis Ames	16
5.3	Effect of Relative Humidity on Efficacy of MB against B. anthracis Ames	17
5.4	Effect of CT on Efficacy of MB against B. anthracis Ames	18
5.5	Effect of Grime	19
5.6	Comparison of B.a. Ames vs Sterne	19
5.7	Surface Damage to Materials	20
5.8	Summary and Conclusion	20
6.0 References	22
Appendices
Appendix A Detailed Test Results	24
Appendix B Detailed Statistical Analysis	1
vi

-------
Figures
Figure 2-1. Coupon not applied in simulated subway grime. From left to right: ceramic tile,
painted carbon steel, weathered concrete, and granite	4
Figure 2-2. Coupons applied in simulated subway grime. From left to right: ceramic tile, painted
carbon steel, weathered concrete, and granite	4
Figure 2-3. Liquid inoculation of coupon using a micropipette	5
Figure 3-1. Schematic of MB decontamination test chamber housed inside custom compact
glove box	9
Figure 5-1. Summary of MB efficacy results, by material, for B. anthracis Ames and B.
anthracis Sternea. Results shown are average LR ± 95% CI	15
Figure 5-2. Effect of relative humidity on MB decontamination efficacy against B. anthracis
Ames. Results shown as average log reduction ± CI	17
Figure 5-3. Effect of relative humidity on MB decontamination efficacy against B. anthracis
Ames. Results shown as average log reduction ± CI	17
Figure 5-4. Summary of the effect of CT on average MB decontamination efficacy against B.
anthracis Ames (Non Grimed Coupons)	18
Figure 5-5. Summary of the effect of CT on average MB decontamination efficacy against B.
anthracis Ames (Grimed Coupons)	18
Figure 5-6. Summary of the effect of grime on average MB decontamination efficacy against B.
anthracis Ames	19
Figure 5-7. MB decontamination efficacy against B. anthracis Ames and Sterne (Grimed
Coupons)	20
Figure B-l. Plot of Control Coupon Percent Recovery of Inoculum by Material and Simulated
Subway Grime for B. anthracis Ames. Note That Percent Recovery Values Greater
than 200% Are Not Included in the Plot	11
Figure B-2. Plot of Control Coupon Percent Recovery of Inoculum by Material and Strain. Note
That Percent Recovery Values Greater than 200% Are Not Included in the Plot	12
Figure B-3. Plot of Decontaminated Coupon Log (Base 10) Reduction of Inoculum by Material,
Simulated Subway Grime, Temperature, Relative Humidity, and CT	12
Figure B-4. Plot of Decontaminated Coupon Log (Base 10) Reduction of Inoculum by Material
and Strain	13
Tables
Table E-l. CT Required to Achieve >6 LR of B.a. Ames on all Materials	iv
Table 2-1. Text Matrix	3
Table 2-2. Test Materials	4
Table 4-1. Actual Fumigation Conditions for MB Tests	12
Table 4-2. Performance Evaluation Audits	12
Table 5-1. CT Required to Achieve >6 LR of B. anthracis on all Materials	14
Table 5-2. Summary of B.a Ames Average Log Reductions by Material Type Error! Bookmark
not defined.
Table 5-3. Summary of B.a Sterne Log Reductions by Material Type	Error! Bookmark not
defined.
Table 5-4. Average Difference in Efficacy between Test 10 (4.5) and Test 8 (10°C)	16
Table A-l. Inactivation of B. anthracis Ames Spores using Methyl Bromide	24
Table A-2. Inactivation of B. anthracis Sterne Spores using Methyl Bromide	25
vii

-------
Table A-5. Effect of Increasing Relative Humidity at High Temperatures on B. anthracis Ames
	26
Table A-6. Effect of Increasing Relative Humidity at Low Temperatures on B. anthracis Ames
	26
Table A-7. Effect of Increasing CT at High Relative Humidity on B. anthracis Ames with no
Grime	27
Table A-8. Effect of Increasing CT at High Relative Humidity on B. anthracis Ames with
Grime	28
Table B-l. Mean Percent Recovery for Control Coupons for Each Strain, Material, and
Simulated Subway Grime Coating Combination with 95 Percent Confidence
Intervals	3
Table B-2. Percent Recovery ANOVA Summary Table Testing the Effect for Simulated Subway
Grime on Ceramic Tile Control Coupons	3
Table B-3. Percent Recovery ANOVA Summary Table Testing the Effect for Simulated Subway
Grime on Painted Carbon Steel Control Coupons	3
Table B-4. Percent Recovery ANOVA Summary Table Testing the Effect for Simulated Subway
Grime on Weathered Concrete Control Coupons	3
Table B-5. Percent Recovery Least Squares Means for Simulated Subway Grime Conditions on
Weathered Concrete Control Coupons	4
Table B-6. Percent Recovery ANOVA Summary Table Testing the Effect for Simulated Subway
Grime on Granite Control Coupons	4
Table B-7. Percent Recovery Least Squares Means for Simulated Subway Grime Conditions on
Granite Control Coupons	4
Table B-8. Percent Recovery ANOVA Summary Table Testing the Effect for Material Without
Simulated Subway Grime Applied Control Coupons	4
Table B-9. Percent Recovery Tukey Comparisons for Materials Without Simulated Subway
Grime Applied Control Coupons	4
Table B-10. Percent Recovery ANOVA Summary Table Testing the Effect for Material with
Simulated Subway Grime Applied Control Coupons	5
Table B-l 1. Percent Recovery Tukey Comparisons for Materials with Simulated Subway Grime
Applied Control Coupons	5
Table B-12. Percent Recovery ANOVA Summary Table Testing the Effect for Strain and
Material for Control Coupons with Simulated Subway Grime	5
Table B-13. Percent Recovery ANOVA Summary Table Testing the Main Effects for Strain and
Material for Control Coupons with Simulated Subway Grime	5
Table B-14. Percent Recovery Tukey Comparisons for Material for Control Coupons	6
Table B-15. Mean Log (Base 10) Reduction for Decontaminated Coupons for Each
Decontamination Scenario with 95 Percent Confidence Intervals	6
Table B-16. Log-reduction ANOVA Summary Table for Testing Main Effects of Material,
Simulated Subway Grime, Temperature, Relative Humidity, and CT for
Decontaminated Coupons	8
Table B-17. Log-reduction Tukey Comparisons for Materials for Decontaminated Coupons	9
Table B-l8. Log-reduction Least Squares Means for Simulated Subway Grime for
Decontaminated Coupons	9
Table B-19. Log-reduction Least Squares Means for Temperature for Decontaminated Coupons.
	9
Table B-20. Log-reduction Least Squares Means for Relative Humidity for Decontaminated
Coupons	9
Table B-21. Log-reduction Tukey Comparisons for CT for Decontaminated Coupons	9
viii

-------
Table B-22. Log-reduction ANOVA Summary Table Testing the Effect for B. anthracis Strain
and Material for Decontaminated Coupons	10
Table B-23. Log-reduction Tukey Comparisons for Material for B. anthracis Ames
Decontaminated Coupons	10
Table B-24. Log-reduction Tukey Comparisons for Material for B. anthracis Sterne
Decontaminated Coupons	10
Table B-25. Log-reduction Tukey Comparisons for B. anthracis Strain by Material for
Decontaminated Coupons	10
IX

-------
Abbreviations/Acronyms
ANOVA
analysis of variance
B.a.
Bacillus anthracis
BBRC
Battelle Biomedical Research Center
BSC
biological safety cabinet
CBRN
Chemical, Biological, Radiological, and Nuclear
CMAD
Consequence Management Advisory Division
CFU
colony-forming unit(s)
CI
confidence interval
Cm
centimeter(s)
°C
degrees Celsius
CT
contact time
DNA
deoxyribonucleic acid
EPA
U.S. Environmental Protection Agency
ft3
cubic feet
HC1
hydrochloric acid
HS
homeland security
HSRP
Homeland Security Research Program
L
liter(s)
LAL
Limulus Amebocyte Lysate (assay)
LED
light emitting diode
LR
log reduction
MB
methyl bromide
Min
minute(s)
Mg
milligram(s)
mL
milliliter(s)
jiL
microliter(s)
NHSRC
National Homeland Security Research Center
OLEM
Office of Land and Emergency Management
Oz
ounce(s)
PBST
phosphate-buffered saline + 0.1% Triton X-100
PCR
polymerase chain reaction
ppm
part(s) per million
QA
quality assurance
QC
quality control
QMP
Quality Management Plan
RH
relative humidity
rpm
revolution(s) per minute
SD
standard deviation
SE
standard error
SFW
sterile filtered water
T&EII
Testing and Evaluation II Program
TSA
technical systems audit(s)
X

-------
1.0 Introduction
The U.S. Environmental Protection Agency (EPA) is helping to protect human health and the
environment from adverse impacts resulting from the release of chemical, biological, or
radiological agents. With emphasis on decontamination and consequence management, water
infrastructure protection, and threat and consequence assessment, the EPA is working to develop
tools and information that: help detect the intentional introduction of chemical or biological
contaminants into buildings or water systems; contain these contaminants; decontaminate
buildings, outdoor environments, or water systems; and facilitate the disposal of material resulting
from restoration activities.
In this investigation, the efficacy of methyl bromide (MB) against Bacillus anthracis (B.a.)
Ames spores applied to subway materials (ceramic tile, painted carbon steel, weathered concrete,
and granite) was tested. Simulated subway grime was applied to the surface of the materials
during Tests 5 through 10 prior to testing to determine impact on the efficacy of MB.
Decontamination efficacy was determined based on the log reduction (LR) in viable spores
recovered from the inoculated samples (with and without exposure to MB). A decontaminant or
fumigant technology is considered effective if a 6 LR or greater is achieved on the materials
tested for a given set of fumigation conditions (sporicidal liquid volume, temperature, and
relative humidity [RH]).(1) This study builds on previous laboratory research conducted by EPA
to assess decontamination efficacy of MB for inactivating B.a. spores on various materials and
adds data for low temperature as well as grimed material exposures.
Lastly, another objective of this work was to obtain efficacy data for B.a. Ames and B.a. Sterne,
which could be used to assess its suitability as a potential surrogate for B.a. Ames when
decontaminating with MB. Previous tests'2"4^ with B. atrophaeus or B. subtilis have shown these
species to be more resistant to MB compared with B.a. Ames. The Ames strain of B.a. was
chosen for use as a standard since it was the strain identified in the Amerithrax incident in
2001(8).
1

-------
2.0 Procedures
This section provides an overview of the procedures used for the evaluation of MB fumigant to
inactivate B.a. Ames on up to four material types. Testing was performed in accordance with the
EPA and Battelle Quality Assurance Programs.
2.1	Technology Description
MB (Chemtura, Philadelphia, PA) has been registered by the EPA for soil fumigation (injected
into the soil before a crop is planted to effectively sterilize the soil), commodity treatment (used
for post-harvest pest control), structural pest control (used to fumigate buildings for termites, and
warehouses and food processing facilities for insects and rodents), and quarantine uses (used to
treat imported commodities). Although MB has also been demonstrated to be an effective
biocide against B.a. Ames on building materials and soil(6), the focus of this study was to
determine effective conditions at lower temperatures, to generate evidence that MB fumigation
for B.a. Ames may be implemented in the event of a potential release in a subway system.
Furthermore, although MB use is being phased out under the Montreal Protocol, MB is still
currently and widely used via critical use exemptions as a soil and commodity (quarantine)
fumigant(7).
2.2	Test Matrix
The test matrix for the MB fumigation is shown in Table 2-1. Each test was performed using four
material types inoculated with B.a. Ames and B.a. Sterne. A subset of tests was conducted using
those same four materials, but with simulated subway grime added prior to inoculation of B.a.
Ames (Tests 5-10). An adaptive management approach was used such that adjustments were
made to the fumigation parameters (CT, organism, material grimed or clean, temperature, or
relative humidity [RH]) to assess the impact of that parameter and to find efficacious conditions.
Tests 1 through 8 and 10 were conducted with all four materials, using B.a. Ames. Test 9 was
conducted with all four materials, using B.a. Sterne to preliminarily assess its use as a
comparable surrogate for B.a. Ames for future testing.
2.3	Biological Agents
The virulent B.a. spores used for this testing were prepared from a qualified stock of the Ames
strain at the Battelle Biomedical Research Center (BBRC, Lot B21, West Jefferson, OH). The
spore lot was subjected to a stringent characterization and qualification process required by
Battelle's standard operating procedure for spore production. Specifically, the spore lot was
characterized prior to use by observation of colony morphology, direct microscopic observation
of spore morphology, and size and determination of percent refractivity and percent
encapsulation. In addition, the number of viable spores was determined by colony count and
expressed as colony forming units per milliliter (CFU/mL). Theoretically, once plated onto
bacterial growth media, each viable spore germinates and yields one CFU. Variations in the
expected colony phenotypes were recorded. Endotoxin concentration of each spore preparation
was determined by the Limulus Amebocyte Lysate (LAL) assay to assess whether contamination
from gram-negative bacteria occurred during the propagation and purification process of the
spores. Genomic deoxyribonucleic acid (DNA) was extracted from the spores and DNA
fingerprinting by polymerase chain reaction (PCR) was done to confirm the genotype. This work
was performed by Dr. Paul Keim at Northern Arizona University. The virulence of the spore lot
was measured at Battelle by challenging guinea pigs intradermally with a dilution series of spore
suspensions, and virulence was expressed as the intradermal median lethal dose. In addition,
testing was conducted for robustness of the spores via hydrochloric acid (HC1) resistance.
2

-------
Table 2-1. MB Test Matrix
Test
Number


Target Fumigation Parameters

Contact
Materials
Organisms
MeBr
Concentration
(mg/L)
Temperature
(°C)
RH
(%)
Time
(days)
1


212
10
75
7
2


212
10
75
2
3


212
10
75
4
4


212
10
75
3


Bacillus anthracis Ames




5*
Ceramic Tile
Painted Carbon Steel
212
10
75
4
6*
Weathered Concrete
Granite

212
10
75
5
'J*


212
4.5
75
7
8*


212
10
50
7
9*

Bacillus anthracis Sterne
212
10
75
5
10s

Bacillus anthracis Ames
212
4.5
50
9
*Tests used grimed materials
2.4 Test Materials
Decontamination testing was conducted on ceramic tile, painted carbon steel, weathered
concrete, and granite. Information on these materials is presented in Table 2-2, and a picture of
each is presented in Figure 2-1. Material coupons were cut to uniform length and width from a
larger piece of stock material. Materials were prepared for testing by sterilization via autoclaving
at 121 °C for 15 minutes. Autoclaved coupons were sealed in sterilization pouches (Fisher,
Pittsburgh, PA) to preserve sterility until the coupons were ready for use. Additionally, when
required, simulated grime was prepared by combining 94% Arizona fine dust (Powder
Technology Inc., PP2G4A2 find), 2.50% Carbon black (Powder Technology Inc, Raven 410),
0.25% Diesel particulate (NIST, SRM 1650b), 0.13% Motor oil, 0.13% alpha-Pinene 97%
(Fisher Scientific, AC16436-0050), 1.00% Lycopodium (Fisher Scientific, S755301), 1.00%
Ragweed pollen (Polysciences Inc., 7673), and 1.00% Paper Mulberry pollen (Polysciences Inc.,
7670). Application of the prepared grime was achieved by combining 14.0 g of the sterile grime
and 300 mL of 95% ethanol into a Binks SV100 spray can. The HPLV sprayer was operated
over the test materials in a sweeping motion to achieve the targeted 0.02 g per test material (14.5
cm2). The simulated grime was applied to ceramic tile, painted carbon steel, weathered concrete,
and granite coupons (Figure 2-2).
3

-------
Table 2-2. Test Materials
Material
Lot, Batch, or ASTM No., or
Observation
Manufacturer/
Supplier Name
Approximate Coupon
Size,
width x length x thickness
Material
Preparation
Ceramic
Tile
Style Selections White Matte
Ceramic Floor Tile Item #: 437485
Lowes
ITilliard, OH
1.9 centimeter (cm) x 5.0 cm x
0.8 cm
Autoclave
Painted
Carbon
Steel
ASTM A1008 Grade CS, Type B
Paint: Bond Plex Water based Acrylic
Adept Products,
West Jefferson, OH
Sherwin Williams,
Columbus, OH
1.9 cmx 5.0 cm x 0.1 cm
Autoclave
Weathered
Concrete
Military-grade runway concrete;
aged 11 years
U.S. Government
1.9 cmx 5.0 cm x 1.0 cm
Autoclave
Granite
Color: Luna Pearl
Konkus Marble,
Columbus, OH
1.9 cmx 5.0 cmx 1.0 cm
Autoclave
Figure 2-1. Coupons not covered with simulated subway grime. From left to right: ceramic
tile, painted carbon steel, weathered concrete, and granite.
Figure 2-2. Coupons coated with simulated subway grime. From left to right: ceramic tile,
painted carbon steel, weathered concrete, and granite.
2.5 Preparation of Coupons
Test and positive control coupons were placed on a flat surface within a Class II biological safety
cabinet (BSC) and inoculated with approximately 1 • 10s CFU of viable B.a. Ames or B.a.
4

-------
Sterne spores per coupon. A 100 microliter (|iL) aliquot of a stock suspension of approximately
1 x lo9 CFU/mL was dispensed using a micropipette applied as 10 droplets across the
coupon surface (see Figure 2-3). This approach provided a more uniform distribution of spores
across the coupon surface than would be obtained through a single drop of the suspension. After
inoculation, the coupons were transferred to a Class III BSC and left undisturbed overnight to
dry under ambient conditions, approximately 22°C and 40% RH.
Figure 2-3. Liquid inoculation of coupon using a micropipette.
The number and type of replicate coupons used for each combination of material, decontaminant,
concentration, and environmental conditions were:
•	five test coupons (inoculated with B.a. and exposed to decontaminant)
•	five positive controls (inoculated with B.a. but not exposed to decontaminant)
•	one laboratory blank (not inoculated and not exposed to the decontaminant)
•	one procedural blank (not inoculated and exposed to the decontaminant)
On the day following spore inoculation, coupons intended for decontamination (including
blanks) were transferred into the test chamber and exposed to the MB fumigant using the
apparatus and application conditions specified in Section 3.0 of this report. Control coupons
were added to the control chamber as described in Section 3.0.
2.6 Coupon Extraction and Biological Agent Quantification
For sample extraction, test coupons, positive controls, and blanks were placed in 50 mL
polypropylene conical tubes containing 10 mL of sterile phosphate buffered saline + 0.1% Triton
X-100 (PBST). The vials were capped, placed on their sides and agitated on an orbital shaker for
15 minutes (min) at approximately 200 revolutions per minute (rpm) at room temperature.
Residual viable spores were determined using a dilution plating approach. Following extraction,
the extract was removed and a series of 10-fold dilutions was prepared in sterile filtered water
(SFW). An aliquot (0.1 mL) of either the undiluted extract and/or each serial dilution were plated
onto tryptic soy agar in triplicate and were incubated for 18-24 hours at 37 ± 2 °C. Colonies were
counted manually and CFU/mL were determined by multiplying the average number of colonies
per plate by the reciprocal of the dilution. Dilution data representing the greatest number of
individually definable colonies were expressed as arithmetic mean ± standard deviation (SD) of
the numbers of CFLJ observed. Laboratory blanks controlled for sterility, and procedural blanks
5

-------
controlled for viable spores inadvertently introduced to test coupons. The target acceptance
criterion was that extracts of laboratory or procedural blanks were to contain no CFU.
After each decontamination test, the BSC III and the MB test and control chambers were
thoroughly cleaned (using separate steps involving bleach, ethanol, water, then drying).
2.7 Decontamination Efficacy
The mean percent spore recovery from each coupon was calculated using results from positive
control coupons (inoculated, not decontaminated), by means of the following equation:
where Mean CFUpc is the mean number of CFU recovered from five replicate positive control
coupons of a single material, and CFUspike is the number of CFU spiked onto each of those
coupons. The value of CFUspike is known from enumeration of the stock spore suspension. One
aliquot of the stock suspension is plated and enumerated on each day of testing to confirm
CFUspike concentration. Spore recovery was calculated for B.a. Ames or Sterne on each coupon,
and the results are included in Section 5 and Appendix A.
The efficacy of MB was assessed by determining the number of viable organisms remaining on
each test coupon after decontamination. Those numbers were compared to the number of viable
organisms extracted from the positive control coupons.
The number of viable spores of B.a. Ames or Sterne in extracts of test and positive control
coupons was determined to calculate efficacy of the decontaminant. Efficacy is defined as the
extent (as logio reduction or LR) to which viable spores extracted from test coupons after
decontamination were less numerous than the viable spores extracted from positive control
coupons. The logarithm of the CFU abundance from each coupon extract was determined, and
the mean of those logarithm values was then determined for each set of control and associated
test coupons, respectively. Efficacy of a decontaminant for a test organism/test condition on the
z'th coupon material was calculated as the difference between those mean log values, i.e.:
where logio CFUcy refers to the j individual logarithm values obtained from the positive control
coupons, logio CFUtij refers to the j individual logarithm values obtained from the
corresponding test coupons, and the overbar designates a mean value. In tests conducted under
this plan, there were five positive controls and five corresponding test coupons (i.e. J = 5) for
each test. A decontaminant that achieves a 6 LR or greater is considered effective.(2)
In the case where no viable spores were found in any of the five test coupon extracts after
decontamination, a CFU abundance of 1 was assigned, resulting in a logio CFU of zero for that
material. This situation occurred when the decontaminant was highly effective, and no viable
spores were found on the decontaminated test coupons. In such cases, the final efficacy on that
material was reported as greater than or equal to (>) the value calculated by Equation 2.
The variances (i.e., the square of the SD) of the logio CFUctj and logio CFUtij values were also
calculated for both the control and test coupons (i.e., S2Cij and S2ty), and were used to calculate
the pooled standard error (SE) for the efficacy value calculated in Equation 2, as follows:
Mean % Recovery = [Mean CFUpc/CFUspike] x 100
(1)
Efficacy (LR) = (log10 CFUcy) - (log10 CFUty)
(2)
(3)
6

-------
where the number 5 again represents the number j of coupons in both the control and test data
sets. Each efficacy result is reported as a LR value with an associated 95% confidence interval
(CI), calculated as follows:
95% CI = Efficacy (LR) ± (1.96 x SE)	(4)
The significance of differences in efficacy across different test conditions was assessed based on
the 95%) CI of each efficacy result.
2.8	Statistical Analysis
The mean and 95% CIs on the percent recovery for the control coupons were calculated by
strain, material, and simulated subway grime coating. For B.a. Ames and each material, an
analysis of variance (ANOVA) model with main effect for simulated subway grime was fit to
percent recovery. An ANOVA model with main effect for material was fit to the percent
recovery data for control coupons with and without the simulated subway grime for B.a Ames.
Finally, an ANOVA model with main effects for B.a. strain and material and the two-factor
interaction effect was fit to the percent recovery data, comparing Test 9 to the combined data
from Tests 5, 6, 7, 8, and 10. For any effects of factors with more than two levels found to be
statistically significant, Tukey comparisons were used to identify which levels of the effect are
different; for effects of factors with two levels found to be statistically significant, least squares
means were calculated.
The mean and 95% CIs on the reduction of the logarithm (base 10) B.a for the decontaminated
coupons were calculated by strain, material, simulated subway grim coating, temperature,
relative humidity and CT. For B.a Ames, an ANOVA model with main effects for material,
simulated subway grime coating, temperature, RH, and CT was fit to the reduction of logarithm
(base 10) spores. For any effects of factors with more than two levels found to be statistically
significant, Tukey comparisons were used to identify which levels of the effect are different; for
effects of factors with two levels found to be statistically significant, least squares means were
calculated.
An ANOVA model with main effects for B.a strain and material and the two-factor interaction
was fit to the reduction of logarithm (base 10) spores for Tests 6 and 9. For any effects of factors
with more than two levels found to be statistically significant, Tukey comparisons were used to
identify which levels of the effect are different; for effects of factors with two levels found to be
statistically significant, least squares means were calculated.
All statistical analyses were performed using Statistical Analysis Software (SAS; version 9.4,
Cary, NC). All results are reported at the 0.05 level of significance.
2.9	Surface Damage
The physical effect of MB on the materials was also qualitatively monitored during the
evaluation. This approach provided a gross visual assessment of whether the decontaminants
altered the appearance of the test materials. The procedural blank (coupon that is
decontaminated, but has no spores applied) was visually compared to a laboratory blank coupon
(a coupon not exposed to the decontaminant and that has no spores applied). Obvious visible
damage might include structural damage, surface degradation, discoloration, or other aesthetic
impacts.
7

-------
3.0 Fumigation Description and Procedures
Methyl bromide is a colorless and odorless volatile gas. Chloropicrin (Sigma Aldrich, St. Louis,
MO) was added to the MB source gas (0.5% chloropicrin, 99.5% MB) as a warning irritant
(lacrimator) for the safety of laboratory staff. The gas mixture was used at full strength and
injected into the test chamber at the indicated target concentrations.
Figure 3-1 shows a schematic drawing of the MB test chamber and containment system.
Decontamination testing was conducted inside an approximately 38 liter (L) stainless steel
chamber. The chamber was insulated to prevent condensation on the inside chamber walls. As a
means of secondary containment and laboratory personnel safety, this test chamber was housed
inside a custom acrylic compact glove box (Plas Labs, Inc., Lansing, MI) that was hard-ducted to
the facility exhaust system.
Temperature was controlled using a heated/cooled water bath, and RH was elevated using a
Nafion tube pervaporation system (controlled using a water bath). Temperature and RH in the
test chamber were measured using an HMT368 temperature and humidity probe (Vaisala, Inc.,
Woburn, MA). Temperature, RH, and MB concentration were controlled with a CNI-822
controller (Omega Engineering, Stamford, CT) and the data were recorded every minute during
the contact time (CT) using the associated iLOG software.
The MB concentration in the test chamber was measured continuously during the contact period
using a Fumiscope™ Version 5.0 (Key Chemical and Equipment Company, Clearwater, FL).
MB was added to the chamber, as necessary, to maintain the 212 mg/L within ±10% as was
show to be efficacious previously(2~4'8). The Fumiscope meter was calibrated by the
manufacturer for MB, displaying the concentration on a digital light-emitting diode (LED)
display in ounces (oz) of MB per 1000 cubic feet (ft3). 1 oz per 1000 ft3 is approximately 257
parts per million (ppm) at 25 °C and is approximately 1 mg/L (independent of temperature). The
Fumiscope meter included an air pump that pulled a gas sample from the test chamber through
the thermal conductivity meter at a controlled rate and exhausted the gas back into the test
chamber. Moisture was removed from the gas sample using a small paper filter before it was
measured in the Fumiscope to eliminate interference from water.
8

-------
% RH
MeBr
Omega Controllers
n
Fumiscope

T™H





Indicates RH lines
Indicates MB Fumigant Lines
Indicates Temperature Control Lines
Indicates Temperature RH Cable
Vaisala
Pressure
Regulator
Inlet Valve
Shell 2
Radiator
Temp
Water Bath
Water Bath
Outlet Valve
Plaslabs Glove Box
Temp/RH Probe
Shelf 1
Transfer Chamber
Figure 3-1. Schematic of MB decontamination test chamber housed inside custom compact glove box.

-------
A 9-L Lock & Lock® airtight container (Lock & Lock, Farmers Branch, TX) served as the
positive control chamber. Fixed humidity point salts(9) were added as a slurry to a separate
container placed in the bottom of the positive control chamber. Sodium chloride was used to
control the RH at 75% and sodium bromide was used to the control the RH at 50%. The control
chamber was placed in an incubator (Thermo Scientific, Waltham, MA) for all tests and set to
the appropriate temperature (i.e., 10 °C). The temperature and RH of the positive control
chamber were measured and the data logged once every minute using a HOBO® data logger
model U12-11 (Onset Computer Corporation, Cape Cod, MA).
As in previous studies with MB(1), multiple coupons of each material were inoculated with the
biological agent and placed on a wire rack inside the test chamber. Blank (i.e., uninoculated) and
positive control (i.e., inoculated but not decontaminated) samples were also prepared for each
material and were utilized with data from the test samples (inoculated and decontaminated) to
determine decontamination efficacy.
Ten MB tests were conducted at a concentration of 212 mg/L (Table 2-1). Target CTs ranged
from 2 to 9 days, target temperature from 4.5 or 10 °C and RH from 50 or 75%. During each test
run, inoculated test samples were placed inside the MB test chamber, and the chamber was
sealed. The chamber was allowed sufficient time to equilibrate to the target temperature and RH
prior to start of the run. Once the temperature and RH were stable, MB was slowly injected into
the chamber until the target concentration was reached. The test chamber remained sealed until
the end of the required CT. At this time, the MB was turned off and the seal of the test chamber
broken by removing the lid. The test chamber and BSC III were allowed to off-gas until the MB
levels in the chamber reach 0.00 mg/L, which occurred within minutes of lid removal. At this
time, the samples were removed and processed as stated in Section 2.6.
10

-------
4.0 Quality Assurance/Quality Control
Quality Assurance (QA) and quality control (QC) procedures were performed in accordance with
the Testing and Evaluation (T&E) II program Quality Management Plan (QMP).. The QA/QC
procedures and results are summarized below.
4.1	Equipment Calibration
All equipment (e.g., pipettes, incubators, biological safety cabinets) and monitoring devices (e.g.,
thermometer, hygrometer) used at the time of evaluation were verified as being certified,
calibrated, or validated.
4.2	QC Results
QC efforts conducted during decontaminant testing included positive control samples
(inoculated, not decontaminated), procedural blanks (not inoculated, decontaminated), laboratory
blank (not inoculated, not decontaminated), and inoculation control samples (analysis of the
stock spore suspension).
All positive control results were within the target recovery range of >5 to <120% of the
inoculated spores, and all procedural and laboratory blanks met the criterion of no observed CFU
for both organisms.
Inoculation control samples were taken from the spore suspension on the day of testing and
serially diluted, nutrient plated, and counted to establish the spore density used to inoculate the
samples. The spore density levels met the QA target criterion of 1 x io9 CFU/mL (±1 log) for all
tests.
4.2.1. Operational Parameters
The temperature, RH, and MB concentration during each test were controlled using Omega
controllers, as described in Section 3.0. These controllers were set to the target conditions and
allowed temperatures, RH, or MB to be adjusted or injected as needed to stay within target
ranges of ±2 °C, ±20% RH and ±10% MB. Readings were taken once every minute for the
duration of the CT. The actual operational parameters for each test are shown in Table 4-1 and
reported as the average value ± SD.
11

-------
Table 4-1. Actual Fumigation Conditions for MB Tests.
Test
Number
MB Concentration
(mg/L)
Target Actual*
Temperature (°C)
Target Actual*
Target
RH (%)
Actual*
CT
(days)
It
212
214.1 ± 14.3
10
10.3 ± 1.6
75
72.3 ±4.1
7
2
212
216.0 ±3.21
10
9.5 ±0.6
75
77.9 ±2.4
2
3
212
220.8 ±8.6
10
9.5 ± 1.4
75
80.9 ±5.1
4
4
212
230.5 ± 15.7
10
10.4 ±0.8
75
78.7 ±4.2
3
5
212
212.8 ±3.4
10
10.0 ±0.3
75
82.3 ±5.5
4
6
212
215.7 ±4.1
10
10.2 ±0.9
75
81.6 ±4.0
5
7
212
221.8± 11.7
4.5
3.1 ±0.8
75
79.9 ±4.5
7
8
212
214.4 ±3.2
10
9.6 ±0.1
50
52.4 ±6.4
7
9
212
215.4 ±4.3
10
9.6 ±0.0
75
77.1 ±0.5
5
10
212
217.1 ±6.9
4.5
4.5 ±0.1
50
48.9 ±4.8
9
* Data reported as average ± SD.
t Parameters deviated from target during Test 1 which is outlined in Section 4.4
4.3
4.3.1
Audits
Performance Evaluation Audit
Performance evaluation audits were conducted to assess the quality of the results obtained during
these experiments. Table 4-2 summarizes the performance evaluation audits that were
performed.
No performance evaluation audits were performed to confirm the concentration and purity of
B.a. Ames or B.a. Sterne spores because quantitative standards do not exist for these organisms.
The control coupons and blanks support the spore measurements.
Table 4-2. Performance Evaluation Audits.
Measurement
Audit
Procedure
Allowable
Tolerance
Actual
Tolerance
Volume of liquid from
micropipettes
Time
Temperature
Relative Humidity
Gravimetric evaluation
Compared to independent clock
Compared to independent calibrated
thermometer
Compare to independent calibrated
hygrometer
± 10%
±0.15% to 2.5%
± 2 seconds/hour 0 seconds/hour
All differences
± 2 °C
± 10%
were <0.3 °C
All differences
were <1.8%
4.3.2 Technical Systems Audit
Observations and findings from the technical systems audit (TSA) were documented and
submitted to the laboratory staff lead for response. TSAs were conducted on August 8, August 9,
and August 11, 2016 to ensure that the tests were being conducted in accordance with the EPA
and Battelle quality assurance programs. As part of the audit, test procedures and data acquisition
and handling procedures were reviewed. None of the findings of the TSA required corrective
action.
12

-------
4.3.3 Data Quality Audit
At least 10% of the data acquired during the evaluation were audited. A QA auditor traced the
data from the initial acquisition, through reduction and statistical analysis, to final reporting to
ensure the integrity of the reported results. All calculations performed on the data undergoing the
audit were checked.
4.4	Quality Assurance Project Plan Deviations
Section 4.2.1 Operation Parameters states "MB concentration has an allowable test measurement
tolerance of ±10%." Test #1 started on July 21, 2016 with target parameters of 212 mg/L, 10 °C,
75%) RH, and a 7-day CT. Due to the target temperature of 10°C, the water circulating through
the radiator was set to 4 °C. This low temperature combined with the high level of desired RH
(75%>) resulted in large amounts of condensation within the radiator inside the test chamber. This
buildup of moisture in turn caused the temperature to rise in the test chamber. To mitigate this,
the test chamber had to be opened approximately every 8 to 10 hours to remove the condensation
from the radiator. As the door to the testing chamber was opened, the MB concentration (mg/L)
in the chamber would briefly drop below the allowable test measurement tolerance. To mitigate
any issues in further testing, the circulating water bath was adjusted to stay above the dew point
of the chamber which minimized condensation inside the test chamber, which reduced the need
to periodically open the door to the chamber.
4.5	QA/QC Reporting
Each assessment and audit were documented in accordance with EPA and Battelle quality
assurance programs. For these tests, findings were noted (none significant) in the data quality
audit, but no follow up corrective action was necessary. The findings were mostly minor data
transcription errors requiring some recalculation of efficacy results, but none were gross errors in
recording. Copies of the assessment reports were distributed to the EPA and laboratory staff.
QA/QC procedures were performed in accordance with the EPA and Battelle quality assurance
programs.
4.6	Data Review
Records and data generated in the evaluation received a QC/technical review before they were
utilized in calculating or evaluating results and prior to incorporation in reports.
13

-------
5.0 Summary of Results and Discussion
The decontamination efficacy of MB against virulent B.a. Ames was evaluated at a target
concentration of 212 mg/L, target temperatures of 4.5 or 10 °C, target RH of 50 or 75%, and CTs
ranging from 2 to 9 days for a total of nine tests. Table 5-1 shows the CT required to achieve >6
LR on all material types tested (ceramic tile, painted carbon steel, weathered concrete, and
granite) with or without grime added, and at all target operational parameters. Actual operational
parameters, as measured, were well within acceptable ranges for all tests except for Test 1 which
is outlined in Section 5.4. The detailed decontamination efficacy results are found in Appendix A.
Table 5-1. CT Required to Achieve >6 LR of B. anthracis on all Materials*
Target MB

Target
Target
RH (%)
Time (days) Required to Achieve >6 LR on All
Materials
Test
Concentration
(mg/L)
Grimed
Temperature
(°C)
B.a. Ames
Number
Reference3
212
No
10
75
4
3
212
Yes
10
75
5
6
212
Yes
4.5
75
7
7
* Materials tested were ceramic tile, painted carbon steel, weathered concrete and granite.
'Detailed data from each test number can be referenced in Table A-l in Appendix A.
5.1 Effects of Test Materials on MB efficacy for B.a. Ames
The LR results by material, for each test, are shown in the bar graphs in Figure 5-1. Differences
in efficacy between two materials are significant if the 95% CIs of the two efficacy results do not
overlap. In general, ceramic tile and weathered concrete were most difficult to decontaminate
(exhibited lower efficacy than painted carbon steel or granite) when testing with Ames. Further
details and statistical analysis on the decontamination efficacy results are found in Appendices A
and B.
14

-------
Sfi 1
Test #1: 212 mg/LMB, 10 °C, 75% RH, 7 days
llll
Test #2: 212 mg/L MB, 10 °C, 75%RH, 2 days
5
$ 4
Sfi 1
Ceramic Tile Painted Carbon Steel Weathered Concrete	Granite
Test #5 (grimed): 212 mg/LMB, 10 °C, 75%RH, 4
days
(2 5
Ceramic Tile Painted Carbon Steel Weathered Concrete	Granite
Test #7(grimed): 212 mg/LMB, 4.5 °C, 75%RH, 7
days
5
$ 4
n i
Ceramic Tile Painted Carbon Steel Weathered Concrete	Granite
Test #9 (grimed): 212 mg/L MB, 10 °C, 75% RH, 5
days
Ceramic Tile Painted Carbon Steel Weathered Concrete	Granite
n	7
3	6
(2	5 -
oJ) .
o	4 -
J
3 "
§	2 -
s	11
Ceramic Tile Painted Carbon Steel Weathered Concrete	Granite
Test #3: 212 mg/L MB, 10 °C, 75% RH, 4 days
I* *
III
p6 5
sfi i
Ceramic Tile Painted Carbon Steel Weathered Concrete	Granite
Test #4: 212 mg/L MB, 10 °C, 75% RH, 3 days
i 11
i
Ceramic Tile Painted Carbon Steel Weathered Concrete	Granite
Test #6 (grimed): 212 mg/L MB, 10 °C, 75% RH, 5
days
uap	y 1	A	uays
II	H	_
i I IIII
Ceramic Tile Painted Carbon Steel Weathered Concrete	Granite
Test #8 (grimed): 212 mg/LMB, 10 °C, 50%RH, 7
days
I* «ays	9	days
t	*	8	*
l ill
Ceramic Tile Painted Carbon Steel Weathered Concrete	Granite
Test #10 (grimed): 212 mg/LMB, 4.5 °C, 50%RH, 9
days
J	y -	ufljs
*	8	T	«¦
III llll
Ceramic Tile Painted Carbon Steel Weathered Concrete	Granite
Figure 5-1. Summary of MB efficacy results, by material, for B. anthracis Ames and B.
anthracis Sterne3. Results shown are average LR± 95% CI.
* Complete inactivation achieved
a Test 9 was tested with Bacillus anthracis Sterne, Test 6 is the comparative test with B.a. Ames
15

-------
5.2 Effect of Temperature on Efficacy of MB against B. anthracis Ames
The decontamination efficacy of MB against virulent B.a. Ames was evaluated at target
temperatures of 4.5 or 10 °C. These temperatures were tested at various combinations of RH and
CTs with and without grime added to the test materials. Due to complexity of test matrix, a direct
comparison of temperature was not achievable.
In general, increasing temperature either increased decontamination efficacy or had no
significant impact on efficacy. Test 10 (212 mg/L, 4.5 °C, 50% RH, 9 Day CT), achieved
complete inactivation on painted carbon steel and granite but only 5.89 LR was achieved on
ceramic tile and 7.30 LR on weathered concrete. For Test 8, all parameters but temperature (10
°C) were the same and the CT decreased to 7 days. Complete inactivation was achieved on
painted carbon steel, weathered concrete, and granite but only 4.95 LR was achieved on ceramic
tile (Table 5-2). Thus, it can be concluded that increasing temperature at 50% RH will have
minimal effect on efficacy. Additional analyses of the effect of temperature, including LR data for
each specific material, are included in Appendix A.
Table 5-2. Average Difference in Efficacy between Test 10 (4.5°C) and Test 8 (10°C)
Material Type
Test 10
Test 8
Average
Difference
in
Efficacy
212 4.5 9
mg/L °C Days
212 10 7
,T or 50% «
mg/L °C Days
Ceramic Tile
5.89
4.95

Painted Carbon Steel
>7.73
>7.73




-0.24
Weathered Concrete
7.30
>7.34

Granite
>7.58
>7.52





16

-------
5.3 Effect of Relative Humidity on Efficacy of MB against B. anthracis Ames
The decontamination efficacy of MB against B.a. Ames was evaluated at a target RH of 50 and
75%. The actual %RH conditions for each test are shown in Table 4-1. These RH levels were
tested at various temperatures and CTs with and without grime added to the test materials. The
comparisons are shown in Figures 5-2 and 5-3. Detailed tabulated results to assess the effect of
RH are summarized Appendix A.
RH Difference at 10°C
Ceramic Tile
Painted Carbon Steel Weathered Concrete
Granite
Figure 5-2. Effect of relative humidity at 10°C on MB decontamination efficacy against B.
anthracis Ames. Results shown as average log reduction ± CI.
RH Difference at 4.5°C
Ceramic Tile
Painted Carbon Steel Weathered Concrete
Granite
Figure 5-3. Effect of relative humidity at 4.5°C on MB decontamination efficacy against B.
anthracis Ames. Results shown as average log reduction ± CI.
Complete inactivation was achieved in Test 1 on all materials tested. For Test 8, the RH was
lowered to 50% and simulated subway grime was added to each coupon materials prior to
inoculation. In this test, complete inactivation was achieved on all coupons except ceramic tile
(4.95 LR), while this may suggest that 75 %RH promoted greater decontamination efficacy for
17

-------
ceramic tile compare to 50 %RH, this cannot be confirmed due to the application of simulated
subway grime for test 8.
5.4 Effect of CT on Efficacy of MB against B. anthracis Ames
The effect of increasing the CT on the efficacy against B.a. Ames was also assessed. The CTs
tested ranged from 2 to 9 days; four non-grimed test conditions (Tests 1-4) and two grimed test
conditions (Tests 5 and 6) could be compared to assess the effect of increasing CT. These
comparisons are summarized in Figures 5-4 and 5-5 and presented in full detail in Appendix A.
Effect of Contact Time (Non Grimed)
~ 7
S?
$ 6
+i
c 5
o
y 4
12	Days
13	Days
14	Days
7 Days
Ceramic Tile	Painted Carbon Steel Weathered Concrete	Granite
Figure 5-4. Summary of the effect of CT on average MB decontamination efficacy against
B. anthracis Ames (Non Grimed Coupons).
Effect of Contact Time (Grimed)
~ 7
S?
$ 6
+1
c 5
U 4
~a
QJ 3
Ctf)
5 2
14	Days
15	Days
Ceramic Tile	Painted Carbon Steel Weathered Concrete
Granite
Figure 5-5. Summary of the effect of CT on average MB decontamination efficacy against
B. anthracis Ames (Grimed Coupons).
18

-------
A significant increase in LR was observed with extended CT for both grimed and non-grimed
test materials. In Test 2 (212 mg/L, 10 °C, 75% RH, 2 Day CT) complete inactivation was not
achieved on any materials tested and achieved an average of 1.59 LR across materials. Using
those same test parameters but extending CT to 4 days resulted in complete inactivation on all
non-grimed test materials. Similarly, using grimed test materials, increasing CT from 4 days to 5
resulted in complete inactivation on all test materials.
5.5 Effect of Grime
The addition of a simulated subway grime to each of the four test materials was evaluated in
Tests 5-10. Comparing Tests 3 and 5 (212 mg/L, 10 °C, 75% RH, 4 Day CT), test materials that
did not have grime applied resulted in complete inactivation for all test materials, but when
grime was applied significant reductions in efficacy for all materials but granite were observed.
These comparisons are summarized in Figures 5-6 and presented in full detail in Appendix B.
While the primary focus of this study was to determine log reduction values, it is worth noting
that statistical analysis showed the addition of grime increased the recovery of spores from
weathered concrete and granite as shown in Appendix B, page B-3 conclusions.
Effect of Grime
Ceramic Tile Painted Carbon Steel Weathered Concrete
Granite
Figure 5-6. Summary of the effect of grime on average MB decontamination efficacy
against B. anthracis Ames.
5.6 Comparison of B.a. Ames vs Sterne
Efficacy of MB on B.a. Sterne was evaluated against B.a. Ames to assess the potential use of
B.a. Sterne as a suitable surrogate for the virulent strain. Comparing Test 6, which used B.a.
Ames (212 mg/L, 10 °C, 75% RH, 5 Day CT) and Test 9 which had identical testing parameters
but used B.a. Sterne, statistical analysis found no significant difference in efficacy for ceramic
tile, weathered concrete, and granite. MB was more efficacious against B.a. Ames on Painted
Carbon Steel than B.a. Sterne, resulting in >7.85 and 5.25 LR, respectively. These comparisons
are summarized in Figure 5-7 and presented in full detail in Appendix B.
19

-------
B.a. Ames vs B.a. Sterne
¦	Sterne (Test9)
¦	Ames (Test 6)
Ceramic Tile
Painted Carbon Steel Weathered Concrete
Granite
Figure 5-7. MB decontamination efficacy against B. anthracis Ames and Sterne (Grimed
Coupons).
5.7	Surface Damage to Materials
At the end of each decontamination test, the procedural blanks were visually compared to the
laboratory blanks, and test coupons were visually compared to positive controls, to assess any
impact MB may have had on each material type. Based on the visual appearance of the
decontaminated coupons, there were no apparent changes in the color, reflectivity, or roughness
of the six material surfaces after being exposed to MB. Note that chloropicrin is often added to
MB as a warning agent and chloropicrin has the potential to cause oxidation to some surfaces(9).
5.8	Summary and Conclusion
This investigation focused on decontamination efficacy when fumigating with MB at
temperatures (low temperatures) and examining the effect of RH. Eliminating or reducing the
need to humidify and/or heat an area of interest would facilitate MB fumigation when used to
decontaminate a subway contaminated with B.a. spores.
This study highlights the roles of CT, RH, temperature, and application of simulated subway
grime when fumigating with MB. There was a clear impact of time as evidenced by the increase
in LR of B.a. Ames with the increase of time (2, 3, 4 & 7 days) resulting in >6 LR at 4 and 7
days. There were no tests (only two tests conducted at 50% RH) in which >6 LR of B.a. Ames
was achieved on all materials when fumigating at 50% RH. Application of grime to the test
materials in the case of weathered concrete, painted steel, and granite decrease decontamination
efficacy. It also increased the time to achieve complete inactivation on all test materials from 4
to 5 days. Lastly when comparing B.a. Ames to B.a. Sterne, three of the test materials resulted in
similar reductions while B.a. Sterne was more resistant on Painted Carbon Steel. This was only
conducted for a single test but shows that B.a. Sterne may be a suitable surrogate for the virulent
Ames strain. Additional testing would need to be performed to confirm this result.
20

-------
Impact of Study
This work provides information on the operational parameters of MB fumigation that are
required to achieve efficacy when tested at temperatures and on materials that would be typical
of a subway underground transit system that has been contaminated with B.a Ames and Sterne
spores. Such results may be useful in the development of guidance to aid in deployment of MB
fumigation after a wide-area release of B.a Ames spores in a subway environment.
21

-------
6.0 References
1.	Determining the Efficacy of Liquids and Fumigants in Systematic Decontamination
Studies for Bacillus anthracis Using Multiple Test Methods. US EPA Report 600/R-
10/088, December 2010.
(http://cfpub.epa.gov/si/si public record report.cfm?dirEntrvld=227175&fed org id=12
53&address=nhsrc/si/&view=desc&sortBv=pubDateYear&showCriteria=l&count=25&s
earchal 1='i ndoor%20outdoor%20decontami nati on').
2.	Decontamination of Soil Contaminated with Bacillus anthracis Spores. US EPA Report
600/R-13/110, July 2013
(http://cfpub.epa.gov/si/si public record report.cfm?dirEntryId=258007&fed org id= 12
53&address=nhsrc/&view=desc&sortBv=pubDateYear&showCriteria=l&count=25&sea
rchall=).
3.	Evaluation of Fumigant Decontamination Technologies for Surfaces Contaminated with
Bacillus anthracis Spores. US EPA Report 600/S-l 1/010, November 2011.
(https://cfpub.epa.gov/si/si public record report.cfm?dirEntrvld=239584&fed org id=l
253&address=nhsrc/&view=desc&sortBv=pubDateYear&showCriteria=l&count=25&se
archall='indoor%2520outdoor%2520decontamination'%2520AND%2520'biologican.
4.	Systematic Investigation of Liquid and Fumigant Decontamination Efficacy against
Biological Agents Deposited on Test Coupons of Common Indoor Materials. US EPA
Report 600/R-11/076, August 2011.
(http://cfpub.epa.gov/si/si public record report.cfm?dirEntryId=235044).
5.	Evaluation of Ethylene Oxide for the Inactivation o/Bacillus anthracis. EPA Technology
Evaluation Report. December 2013. EPA/13/R-13-220.
(http://cfpub.epa.gov/si/si public record report.cfm?dirEntrvld=262949&fed org id= 12
53&address=nhsrc%2Fsi%2F&view=desc&sortBv=pubDateYear&count=25&showCrite
ria= 1 &searchall=EPA° o2F600° o2FR+13%2F220+&submit=Search).
6.	Decontamination of Soil Contaminated with Bacillus anthracis Spores. EPA Technology
Evaluation Report. August 2013. EPA/600/R13/110.
(https://cfpub.epa.gov/si/si public record report.cfm?dirEntrvld=258007&fed org id=l
253&subiect=Homeland%20Securitv%20Research&view=desc&sortBy=pubDateYear&
showCriteria=l&count=25&searchall='indoor%20outdoor%20decontamination'%20AN
D%20'biologicaD
7.	The Phaseout of Methyl Bromide, (http://www, epa. uo v/ozone/mbr/i n dex. ht m 1) Accessed
February 24, 2017
8.	Wood, J.P., Wendling, M., Richter, W., Lastivka, A., Mickelsen, L. (2016) Evaluation of
the Efficacy of Methyl Bromide in the Decontamination of Building and Interior
Materials Contaminated with B. anthracis spores. Applied and Environmental
Microbiology. April 2016; 82:7 2003-2011
9.	ASTM International. E104-02, Reapproved 2012. Standard Practice for Maintaining
Constant Relative Humidity by Means of Aqueous Solutions.
22

-------
10. Compatibility of Material and Electronic Equipment with Methyl Bromide and Chlorine
Dioxide Fumigation, EPA Technology Evaluation Report. October 2012. EPA/600/R-
12/664. (https://cfpub.epa.gov/si/si public file download.cfm?p download id=508138)
23

-------
Appendix A
Detailed Test Results
Efficacy Results
The detailed decontamination efficacy results for methyl bromide against B. a. Ames on four material
types (ceramic tile, painted carbon steel, weathered concrete and granite) are shown in Table A-l.
Zero CFU were observed on all laboratory and procedural blanks.
Table A-l. Inactivation of B. anthracis Ames Spores using Methyl Bromide3
Test
Number





Concentration (mg/L) /
CT (days)
Temp (°C) /
RH (%)
Positive Controlb
Test Couponc
1
212/7
10/75
Ceramic Tile
7.97 xlO7
4.50 ± 1.47x107
0.00 ± 0.00
>7.64 ± 0.11
Painted Carbon Steel
7.55 ± 0.54 xlO7
0.00 ± 0.00
>7.88 ± 0.03
Weathered Concrete
1.48 ± 0.90 xlO7
0.00 ± 0.00
>7.12 ±0.21
Granite
1.50 ± 2.46 xlO6
0.00 ± 0.00
>7.17 ±0.06
2
212/2
10/75
Ceramic Tile
8.23 x 107
3.60 ± 0.78 xlO7
7.56 ± 3.49 xlO5
1.72 ±0.22
Painted Carbon Steel
5.63 ± 0.80 xlO7
8.22 ± 3.32 xlO5
1.87 ±0.20
Weathered Concrete
1.24 ± 0.39 xlO7
5.29 ± 3.23 xlO5
1.44 ±0.30
Granite
1.20 ± 0.26 xlO7
6.01 ±2.44 xlO5
1.32 ±0.17
3
212/4
10/75
Ceramic Tile
8.93 x 107
4.32 ± 1.30x107
0.00 ± 0.00
>7.62 ±0.11
Painted Carbon Steel
8.08 ± 0.34 xlO7
0.00 ± 0.00
>7.91 ±0.02
Weathered Concrete
2.17 ± 1.70 xlO7
0.00 ± 0.00
>7.26 ± 0.23
Granite
1.79 ± 0.67 xlO7
0.00 ± 0.00
>7.23 ±0.13
4
212/3
10/75
Ceramic Tile
8.13 xlO7
2.65 ± 0.58 xlO7
8.99 ± 11.3 xlO4
2.66 ±0.37
Painted Carbon Steel
6.00 ± 0.89 xlO7
9.26 ± 5.72 xlO2
5.03 ±0.61
Weathered Concrete
1.35 ± 0.76 xlO7
1.86 ± 2.40 xlO4
3.46 ± 0.90
Granite
8.05 ± 2.23 xlO6
7.75 ± 1.31 xlO3
3.55 ± 0.76
5 e
212/4
10/75
Ceramic Tile
7.77 xlO7
5.87 ± 0.58 xlO7
3.20 ± 1.46 xlO2
5.34 ±0.30
Painted Carbon Steel
5.19 ± 0.84 xlO7
6.69 ± 6.60 xlO1
6.50 ±0.98
Weathered Concrete
3.08 ± 2.00 xlO7
1.05 ± 1.41 xlO3
5.38 ± 1.28
Granite
3.14 ± 0.47 xlO7
4.73 ± 7.27 xlO1
6.68 ±0.98
6 e
212/5
10/75
Ceramic Tile
7.70 X 107
1.70 ± 0.96 xlO7
0.00 ± 0.00
>7.18 ±0.21
Painted Carbon Steel
7.07 ± 0.53 xlO7
0.00 ± 0.00
>7.85 ± 0.03
Weathered Concrete
1.01 ±0.43 xlO7
0.00 ± 0.00
>6.97 ±0.16
Granite
1.33 ± 0.14 x 107
0.00 ± 0.00
>7.12 ±0.04
7 e
212/7
4.5/75
Ceramic Tile
7.63 x 107
5.57 ± 1.26 xlO7
0.00 ± 0.00
>7.74 ±0.08
Painted Carbon Steel
7.34 ± 0.43 xlO7
0.00 ± 0.00
>7.87 ±0.02
Weathered Concrete
2.99 ± 1.62 xlO7
0.00 ± 0.00
>7.41 ± 0.25
Granite
1.80 ± 0.60 xlO7
0.00 ± 0.00
>7.24 ±0.11
8 e
212/7
10/50
Ceramic Tile
9.13 xlO7
3.95 ± 0.55 xlO7
2.67 ± 3.03 xlO3
4.95 ± 1.36
Painted Carbon Steel
5.43 ± 0.41 xlO7
0.00 ± 0.00
>7.73 ± 0.03
Weathered Concrete
2.52 ± 1.27 xlO7
0.00 ± 0.00
>7.34 ±0.23
Granite
3.33 ± 0.48 xlO7
0.00 ± 0.00
>7.52 ± 0.06
10 e
212/9
4.5/50
Ceramic Tile
9.60 x 107
5.93 ± 1.17 x 107
1.20 ± 1.86 xlO3
5.89 ± 1.53
Painted Carbon Steel
5.38 ± 0.34 xlO7
0.00 ± 0.00
>7.73 ± 0.02
Weathered Concrete
4.12 ± 0.98 xlO7
7.46 ± 14.4 xlO1
7.30 ± 0.60
Granite
3.87 ± 0.74 xlO7
0.00 ± 0.00
>7.58 ± 0.07
a Data are expressed as the mean (± SD) of the logs of the number of spores (CFU) observed on five individual samples, and decontamination
efficacy (log reduction).
b Positive Controls = samples inoculated, not decontaminated.
c Test Coupons = samples inoculated, decontaminated.
11 CI = confidence interval (± 1.96 SE).
e Test 5-10 had materials applied with simulated subway grime
A-24

-------
Table A-2. Inactivation of B. anthracis Sterne Spores using Methyl Bromide3

Target Parameters





Test
Number
Concentration (mg/L) /
CT (days)
Temp (°C) /
RH (%)
Material
Inoculum
(CFU/coupon)
Mean Recovered B. a. Sterne
(CFU/coupon)
Decontamination
Efficacy ± CId





Positive Control1*
Test Couponc




Ceramic Tile

3.25 ± 0.80 xlO7
0.00 ±0.00
>7.50 ±0.11
ge
212/5
10/75
Painted Carbon Steel
6.10 xlO7
5.71 ± 1.00 xlO7
5.99 ± 5.25 xlO2
5.25 ±0.59
Weathered Concrete
2.10 ± 0.94 xlO7
0.00 ±0.00
>7.29 ±0.18



Granite

2.07 ± 0.76 xlO7
0.00 ±0.00
>7.29 ±0.13
a Data are expressed as the mean (± SD) of the logs of the number of spores (CFU) observed on five individual samples, and decontamination
efficacy (log reduction).
b Positive Controls = samples inoculated, not decontaminated.
c Test Samples = samples inoculated, decontaminated.
11 CI = confidence interval (± 1.96 SE).
e Test 9 had materials applied with simulated subway grime
A-2 5

-------
Effect of Relative Humidity on Efficacy of MB against B. anthracis Ames
The decontamination efficacy of MB against B. a. Ames was evaluated at target relative humidity of
50 or 75%. The actual %RH conditions for each test are shown in Section 4.2. These RH levels were
tested at various temperatures, and CTs and results are summarized in Table A-5 below and
discussed in Section 6.4. The comparisons are made for two test conditions which share the same
fumigation parameters except RH and grime application.
Table A-5. Effect of Increasing Relative Humidity at High Temperatures on
B. anthracis Ames*
Material Type
Test la
Test 8ab
Average
Increase
in
Efficacy
10 °C 75% 7 days
10 °C 50% 7 days
Ceramic Tile
>7.64
4.95

Painted Carbon Steel
>7.88
>7.73




-0.57
Weathered Concrete
>7.12
>7.34

Granite
>7.17
>7.52





* Data are expressed as decontamination efficacy (log reduction).
" Parameters of each test listed in order of MB concentration (mg/L), temperature (°C), %RH, and CT (days).
b Grime applied to test materials
Table A-6. Effect of Increasing Relative Humidity at Low Temperatures on
B. anthracis Ames*
Material Type
Test 7ab
Test 10ab
Average
Increase
in
Efficacy
212 4.5 __n. _ .
mg/L °C 75% 7dayS
212 4.5 cao/ ft i
mg/L °C 50% 9dayS
Ceramic Tile
>7.74
5.89

Painted Carbon Steel
>7.87
>7.73




-0.70
Weathered Concrete
>7.41
7.30

Granite
>7.24
>7.58





* Data are expressed as decontamination efficacy (log reduction).
a Parameters of each test listed in order of MB concentration (mg/L), temperature (°C), %RH, and CT (days).
b Grime applied to test materials
A-26

-------
Effects of CT on Efficacy of MB against B. anthracis Ames
The effect of increasing the CTs to MB at low and high %RH on the efficacy against B.a. Ames
was assessed by comparing Tests 1-4 for non-grimed test materials and Tests 5 and 6 for grimed
materials. The CTs tested ranged from 2 to 7 days and actual CTs did not deviate from these
targets except for Test 1 which is described in Section 5.4. The results are summarized in Table
A-6. The comparisons are made for two test conditions that share the same fumigation
parameters except CT.
Table A-7. Effect of Increasing CT at High Relative Humidity on B.
anthracis Ames* with no Grime.

Test 2a
Test 4a
Average
Material Type
212
mg/L
10 nsn/
OQ 75/0
2
days
212
mg/L
10
OQ 75/0
3
days
Increase in
Efficacy
Ceramic Tile

1.72


2.66


Painted Carbon

1.87


5.03


Steel




2.09
Weathered

1.44


3.46

Concrete





Granite

1.32


3.55



Test 4a
Test 3a
Average
Material Type
212
mg/L
10 nsn/
OQ 75/0
3
days
212
mg/L
10
OQ 75/0
4
days
Increase in
Efficacy
Ceramic Tile

2.66


>7.62


Painted Carbon

5.03


>7.91


Steel




6.62
Weathered

3.46


>7.26

Concrete





Granite

3.55


>7.23



Test 3a
Test la
Average
Material Type
212
mg/L
o£ 750/0
4
days
212
mg/L
!c 75%
7
days
Increase in
Efficacy
Ceramic Tile

>7.62


>7.64


Painted Carbon

>7.91


>7.88


Steel




-0.05
Weathered
Concrete

>7.26


>7.12

Granite

>7.23


>7.17


* Data are expressed as decontamination efficacy (log reduction).
a Parameters of each test listed in order of MB concentration (mg/L), temperature (°C), %RH, and CT (days).
A-27

-------
Table A-8. Effect of Increasing CT at High Relative Humidity on B.
anthracis Ames* with Grime.

Test 5a
Test 6a
Average
Material Type
212
10 nsn/
OQ 75/0
4
212
10 nsn/
OQ 75/0
5
Increase in

mg/L
days
mg/L
days
Efficacy
Ceramic Tile

5.34


>7.18


Painted Carbon

6.50


>7.85


Steel




1.31
Weathered

5.38


>6.97

Concrete





Granite

6.68


>7.12


* Data are expressed as decontamination efficacy (log reduction).
" Parameters of each test listed in order of MB concentration (mg/L), temperature (°C), %RH, and CT (days).
A-28

-------
Appendix B
Detailed Statistical Analysis
Results
Table B-l contains the mean percent recoveries for each strain, material, and simulated subway
grime coating, with 95 percent confidence intervals. Percent recoveries for each B. anthracis
Ames control coupon are plotted in Figure B-l. Percent recoveries comparingB. anthracis strain
for tests 5 through 10 are plotted in Figure B-2.
Table B-2, B-Table B-3, B-Table B-4, and Table B-6 present the ANOVA summary tables for
testing the effect of simulated subway grime for each material. Percent recovery was
significantly different with and without simulated subway grime for weathered concrete (p =
0.0071) and granite (p < 0.0001), with simulated subway grime resulting in greater percent
recovery for both materials (Table B-5 and Table B-7).
ANOVA summary tables for testing the effect of material without and with simulated subway
grime are presented in Table B-8 and Table B-10. Percent recovery was significantly different
among the materials both without and with simulated subway grime. Table B-9 and Table B-l 1
present the Tukey comparisons among the differences, indicating only granite and weathered
concrete were not significantly different with respect to percent recovery on the control coupons.
Table B-12 presents the ANOVA summary table for testing whether there is a difference in
control coupon recovery between the two strains and materials. There was no statistically
significant difference in percent recovery between the two strains (p = 0.2108), but there was a
statistically significant difference among the materials (p < 0.0001). The Tukey comparisons for
materials are presented in Table B-14. All materials were significantly different from one another
except for granite and weathered concrete. Painted carbon steel had a significantly greater
percent recovery than all other materials. Ceramic tile had a significantly greater percent
recovery than granite and weathered concrete.
Estimates with exact 95 percent confidence intervals for the reduction in log (base-10)
B. anthracis spores are presented in Table B-l5.
The main effects ANOVA model fitted to the B. anthracis Ames strain log-reduction data
summary table is presented in Table B-16. All main effects were statistically significant. The
Tukey comparisons for material are presented in Table B-17 and for CT are presented in Table
B-21; least squares means for simulated subway grime, temperature, and relative humidity are
presented in Table B-18, Table B-19, and Table B-20, respectively. From Table B-17, the log-
reduction for ceramic tile was significantly less than the other materials, and no other materials
were significantly different from each other. Log-reduction was statistically greater when there
was no simulated subway grime (Table B-18); log-reduction was statistically greater at 4.5
degrees (Table B-19) and 50% relative humidity (Table B-20). From Table B-21, a CT of 4 days
was not statistically different from CTs of 7 and 9 days, while all other pairs of CTs were
statistically different with respect to reduction in log (base 10) B. anthracis Ames spores.
Generally, longer times had greater log-reduction except for 7 and 9 days which had significantly
less log-reduction than 5 days and 9 days had significantly less log-reduction than 7 days.
Table B-22 presents the ANOVA summary table for testing whether there was a difference in
decontaminated coupon reduction between the two B. anthracis strains and materials. This model
only included data from Tests 6 and 9 because they had the same combination of subway grime,
temperature, relative humidity and CT. There was a statistically significant interaction between
B-l

-------
strain and material (p < 0.001), indicating that the difference in reduction between the two strains
of log (base 10) B. anthracis depends on the material. Table B-25 presents the Tukey
comparisons among the material combinations for Ames strain, and shows the model did not
estimate any difference in log-reduction between the materials for Ames strain. Table B- presents
the Tukey comparisons among the material combinations for Sterne strain, and shows the that
reduction of log (base 10) B. anthracis Sterne spores on painted carbon steel was statistically less
than that for all other materials. Table B-16 presents the Tukey comparisons between the strains
for each material, and shows that there was a significantly greater reduction for Ames strain than
Sterne on painted carbon steel. Results are consistent with the data that show all combinations of
material and strain had a complete kill except for Sterne strain on painted carbon steel.
Conclusions
Analysis of the percent recovery showed statistically significant differences in percent recovery
with and without simulated subway grime for weathered concrete and granite. For both
materials, the percent recovery was greater for coupons with simulated subway grime. In
addition, painted carbon steel had a significantly greater percent recovery than all other materials
while ceramic tile had a significantly greater percent recovery than granite and weathered
concrete. It can also be concluded that B. anthracis Ames and B. anthracis Sterne were not
significantly different with respect to the percent recovery.
Reduction in log (base-10) B. anthracis Ames spores was statistically different among the
different materials, simulated subway grime, temperatures, relative humidity, and CTs. The least
percent reduction in spores was seen for ceramic tile. Simulated subway grime resulted in lower
reduction in spores. Greater reduction in spores was observed for temperature equal to 4.5
degrees and relative humidity equal to 50%. Finally, CT of 5 days resulted in the greatest
reduction in B. anthracis Ames spores.
B. anthracis strains were differentially reduced depending on the material. A lower log (base 10)
reduction in B. anthracis Sterne on painted carbon steel was observed compared to B. anthracis
Ames on painted carbon steel and compared to Sterne on other materials.
B-2

-------
Table B-l. Mean Percent Recovery for Control Coupons for Each Strain, Material, and
Simulated Subway Grime Combination with 95 Percent Confidence Intervals.
Strain
Material
Simulated
Subway Grime
N
Mean Percent Recovery
(95% Confidence Interval)
B. anthracis Ames
Ceramic Tile
No
20
45.28 (38.26,52.29)
B. anthracis Ames
Ceramic Tile
Yes
25
55.12 (45.60,64.63)
B. anthracis Ames
Granite
No
20
15.82 (13.09,18.56)
B. anthracis Ames
Granite
Yes
25
31.60 (27.01,36.19)
B. anthracis Ames
Painted Carbon Steel
No
20
81.87 (75.51,88.24)
B. anthracis Ames
Painted Carbon Steel
Yes
25
74.08 (66.59,81.57)
B. anthracis Ames
Weathered Concrete
No
20
18.65 (13.13,24.17)
B. anthracis Ames
Weathered Concrete
Yes
25
32.47 (24.57,40.38)
B. anthracis Sterne
Ceramic Tile
Yes
5
53.25 (36.92,69.57)
B. anthracis Sterne
Granite
Yes
5
33.93 (18.48,49.39)
B. anthracis Sterne
Painted Carbon Steel
Yes
5
93.64 (73.32,100.0)a
B. anthracis Sterne
Weathered Concrete
Yes
5
34.48 (15.36,53.61)
a Confidence limits less than 0 or greater than 100 truncated to 0 or 100 to reflect valid range of percent recovery
values.
Table B-2. Percent Recovery ANOVA Summary Table Testing the Effect for Simulated
Subway Grime on Ceramic Tile Control Coupons.
Source
Degrees of
Freedom
Sum of
Squares
F Statistic
P-value
Grime
1
1075.88
2.72
0.1064
Residual Error
43
17007.56


Table B-3. Percent Recovery ANOVA Summary Table Testing the Effect for Simulated
Subway Grime on Painted Carbon Steel Control Coupons.
Source
Degrees of
Freedom
Sum of
Squares
F Statistic
P-value
Grime
1
674.69
2.54
0.1182
Residual Error
43
11412.63


Table B-4. Percent Recovery ANOVA Summary Table Testing the Effect for Simulated
Subway Grime on Weathered Concrete Control Coupons.
Source
Degrees of
Freedom
Sum of
Squares
F Statistic
P-value
Grime
1
2123.58
7.98
0.0071*
Residual Error
43
11437.93


Effect is statistically significant at a = 0.05 level.
B-3

-------
Table B-5. Percent Recovery Least Squares Means for Simulated Subway Grime
Conditions on Weathered Concrete Control Coupons.
Simulated
Subway Grime
Mean
Standard
Deviation
No
18.65
3.65
Yes
32.47
3.26
Table B-6. Percent Recovery ANOVA Summary Table Testing the Effect for Simulated
Subway Grime on Granite Control Coupons.
Source
Degrees of
Freedom
Sum of
Squares
F Statistic
P-value
Grime
1
2766.32
32.87
<0.001*
Residual Error
43
3618.36


* Effect is statistically significant at a = 0.05 level.
Table B-7. Percent Recovery Least Squares Means for Simulated Subway Grime
Conditions on Granite Control Coupons.
Simulated Subway
Grime
Mean
Standard
Deviation
No
15.82
2.05
Yes
31.60
1.83
Table B-8. Percent Recovery ANOVA Summary Table Testing the Effect for Material
Without Simulated Subway Grime Applied Control Coupons.
Source
Degrees of
Freedom
Sum of
Squares
F Statistic
P-value
Material
3
56415.41
129.09
<0.001*
Residual Error
76
11070.90


* Effect is statistically significant at a = 0.05 level.
Table B-9. Percent Recovery Tukey Comparisons# for Materials Without Simulated
Subway Grime Applied Control Coupons.

Granite
Painted Carbon Steel
Weathered Concrete
Ceramic Tile
29.45*
-36.6*
26.63*
Granite

-66.05*
-2.82
Painted Carbon Steel


63.22*
*	Differences are statistically significantly at a = 0.05 level.
#	Positive values indicate the row level results in greater percent recovery than the column level; negative values indicate the
column level results in greater percent recovery.
B-4

-------
Table B-10. Percent Recovery ANOVA Summary Table Testing the Effect for Material
with Simulated Subway Grime Applied Control Coupons.
Source
Degrees of
Freedom
Sum of
Squares
F Statistic
P-value
Material
3
31007.48
30.62
<0.001*
Residual Error
96
32405.58


* Effect is statistically significant at a = 0.05 level.
Table B-ll. Percent Recovery Tukey Comparisons'* for Materials with Simulated Subway
Grime Applied Control Coupons.

Granite
Painted Carbon Steel
Weathered Concrete
Ceramic Tile
23.51*
-18.96*
22.64*
Granite

-42.48*
-0.87
Painted Carbon Steel


41.61*
*	Differences are statistically significantly at a = 0.05 level.
#	Positive values indicate the row level results in greater percent recovery than the column level; negative values indicate the
column level results in greater percent recovery.
Table B-12. Percent Recovery ANOVA Summary Table Testing the Effect for Strain and
Material for Control Coupons with Simulated Subway Grime.
Source
Degrees of
Freedom
Sum of
Squares
F Statistic
P-value
B. anthracis Strain
1
505.49
1.58
0.2108
Material
3
28820.79
30.11
<0.001*
B. anthracis Strain*Material
Interaction
3
1142.58
1.19
0.3155
Residual Error
112
35736.33


* Effect is statistically significant at a = 0.05 level.
Table B-13. Percent Recovery ANOVA Summary Table Testing the Main Effects for
Strain and Material for Control Coupons with Simulated Subway Grime.
Source
Degrees of
Freedom
Sum of
Squares
F Statistic
P-value
B. anthracis Strain
1
505.49
1.58
0.2118
Material
3
41641.43
43.28
<0.001*
Residual Error
115
36878.90


* Effect is statistically significant at a = 0.05 level.
B-5

-------
Table B-14. Percent Recovery Tukey Comparisons'* for Material for Control Coupons.

Granite
Painted Carbon Steel
Weathered Concrete
Ceramic Tile
22.81*
-22.54*
22*
Granite

-45.35*
-0.82
Painted Carbon Steel


44.53*
*	Differences are statistically significantly at a = 0.05 level.
#	Positive values indicate the row level results in greater percent recovery than the column level; negative values indicate the
column level results in greater percent recovery.
Table B-15. Mean Log (Base 10) Reduction" for Decontaminated Coupons for Each
Decontamination Scenario with 95 Percent Confidence Intervals.
Strain
Material
Simulated
Subway
Grime
Temperature
Relative
Humidity
CT
N
Number of
Samples
that were
Complete
Kill
Mean LoglO
Reduction
(95% Confidence
Interval
B. anthracis Ames
Ceramic Tile
No
10
75
2
5
0
2.08 (1.79,2.38)
B. anthracis Ames
Ceramic Tile
No
10
75
3
5
0
3.15 (2.64,3.67)
B. anthracis Ames
Ceramic Tile
No
10
75
4
5
5
7.95 (-. -)
B. anthracis Ames
Ceramic Tile
No
10
75
7
5
5
7.90 (--, ~)
B. anthracis Ames
Ceramic Tile
Yes
4.5
50
9
5
2
6.11 (3.95,8.28)
B. anthracis Ames
Ceramic Tile
Yes
4.5
75
7
5
5
7.88 (-, ~)
B. anthracis Ames
Ceramic Tile
Yes
10
50
7
5
1
5.32 (3.39,7.25)
B. anthracis Ames
Ceramic Tile
Yes
10
75
4
5
1
5.46 (5.04, 5.88)
B. anthracis Ames
Ceramic Tile
Yes
10
75
5
5
5
7.89 (-, ~)
B. anthracis Ames
Granite
No
10
75
2
5
0
2.16 (1.96,2.37)
B. anthracis Ames
Granite
No
10
75
3
5
0
4.57 (3.51,5.63)
B. anthracis Ames
Granite
No
10
75
4
5
5
7.95 (-. -)
B. anthracis Ames
Granite
No
10
75
7
5
5
7.90 (-, ~)
B. anthracis Ames
Granite
Yes
4.5
50
9
5
5
7.98 (-, ~)
B. anthracis Ames
Granite
Yes
4.5
75
7
5
5
7.88 (-, ~)
B. anthracis Ames
Granite
Yes
10
50
7
5
5
7.96 (-. -)
B. anthracis Ames
Granite
Yes
10
75
4
5
4
7.08 (5.69, 8.47)
B. anthracis Ames
Granite
Yes
10
75
5
5
5
7.89 (-, ~)
B. anthracis Ames
Painted Carbon
Steel
No
10
75
2
5
0
2.04 (1.77,2.31)
B. anthracis Ames
Painted Carbon
Steel
No
10
75
3
5
1
5.17(4.31,6.03)
B. anthracis Ames
Painted Carbon
Steel
No
10
75
4
5
5
7.95 (-. -)
B-6

-------
Table B-15. Mean Log (Base 10) Reduction for Decontaminated Coupons for Each
Decontamination Scenario with 95 Percent Confidence Intervals. (Continued)
Strain
Material
Simulated
Subway
Grime
Temperature
Relative
Humidity
CT
N
Number of
Samples
that were
Complete
Kill
Mean LoglO
Reduction
(95% Confidence
Interval
B. anthracis Ames
Painted Carbon
Steel
No
10
75
7
5
5
7.90 (--, ~)
B. anthracis Ames
Painted Carbon
Steel
Yes
4.5
50
9
5
5
7.98 (--, ~)
B. anthracis Ames
Painted Carbon
Steel
Yes
4.5
75
7
5
5
7.88 (--, ~)
B. anthracis Ames
Painted Carbon
Steel
Yes
10
50
7
5
5
7.96 (-. -)
B. anthracis Ames
Painted Carbon
Steel
Yes
10
75
4
5
3
6.68 (5.29, 8.06)
B. anthracis Ames
Painted Carbon
Steel
Yes
10
75
5
5
5
7.89 (--, ~)
— Confidence interval could not be calculated because all sample were a complete kill.
B-7

-------
Table B-15. Mean Log (Base 10) Reduction for Decontaminated Coupons for Each
Decontamination Scenario with 95 Percent Confidence Intervals. (Continued)
Strain
Material
Simulated
Subway
Grime
Temperature
Relative
Humidity
CT
N
Number of
Samples
that were
Complete
Kill
Mean LoglO
Reduction
(95% Confidence
Interval
B. anthracis Ames
Weathered
Concrete
No
10
75
2
5
0
2.27(1.88,2.66)
B. anthracis Ames
Weathered
Concrete
No
10
75
3
5
0
4.30 (3.06,5.54)
B. anthracis Ames
Weathered
Concrete
No
10
75
4
5
5
7.95 (-. -)
B. anthracis Ames
Weathered
Concrete
No
10
75
7
5
5
7.90 (--, ~)
B. anthracis Ames
Weathered
Concrete
Yes
4.5
50
9
5
5
7.68 (6.83, 8.52)
B. anthracis Ames
Weathered
Concrete
Yes
4.5
75
7
5
5
7.88 (-, ~)
B. anthracis Ames
Weathered
Concrete
Yes
10
50
7
5
5
7.96 (-. -)
B. anthracis Ames
Weathered
Concrete
Yes
10
75
4
5
3
5.86 (4.08, 7.64)
B. anthracis Ames
Weathered
Concrete
Yes
10
75
5
5
5
7.89 (-, ~)
B. anthracis Sterne
Ceramic Tile
Yes
10
75
5
5
5
7.79 (-, ~)
B. anthracis Sterne
Granite
Yes
10
75
5
5
5
7.79 (-, ~)
B. anthracis Sterne
Painted Carbon
Steel
Yes
10
75
5
5
1
5.28 (4.46,6.11)
B. anthracis Sterne
Weathered
Concrete
Yes
10
75
5
5
5
7.79 (-, ~)
— Confidence interval couk
not be calculated because all samp
e were a complete kil

Table B-16. Log-reduction ANOVA Summary Table for Testing Main Effects of Material,
Simulated Subway Grime, Temperature, Relative Humidity, and CT for Decontaminated
Coupons.
Source
Degrees of
Freedom
Sum of
Squares
F Statistic
P-value
Material
3
22.05
14.02
<0.001*
Coat
1
28.30
53.97
<0.001*
Temperature
1
13.83
26.38
<0.001*
RH
1
5.84
11.14
0.001*
CT
5
518.62
197.85
<0.001*
Residual Error
168
88.08


* Effect is statistically significant at a = 0.05 level.
B-8

-------
Table B-17. Log-reduction Tukey Comparisons'* for Materials for Decontaminated
Coupons.

Granite
Painted Carbon Steel
Weathered Concrete
Ceramic Tile
-0.85*
-0.86*
-0.66*
Granite

-0.01
0.19
Painted Carbon Steel


0.2
*	Differences are statistically significantly at a = 0.05 level.
#	Positive values indicate the row level results in greater log-reduction than the column level; negative values indicate the
column level results in greater log-reduction.
Table B-18. Log-reduction Least Squares Means for Simulated Subway Grime for
Decontaminated Coupons.
Simulated Subway
Grime
Mean
Standard
Deviation
No
7.74
0.28
Yes
6.06
0.13
Table B-19. Log-reduction Least Squares Means for Temperature for Decontaminated
Coupons.
Temperature
Mean
Standard
Deviation
10
6.07
0.09
4.5
7.73
0.33
Table B-20. Log-reduction Least Squares Means for Relative Humidity for
Decontaminated Coupons.
Relative
Humidity
Mean
Standard
Deviation
50
7.44
0.33
75
6.36
0.09
Table B-21. Log-reduction Tukey Comparisons'* for CT for Decontaminated Coupons.
Time (days)
3
4
5
7
9
2
-2.16*
-5.81*
-7.43*
-5.76*
-4.24*
3

-3.65*
-5.27*
-3.6*
-2.08*
4


-1.62*
0.05
1.57
5



1.67*
3.19*
7




1.52*
*	Differences are statistically significantly at a = 0.05 level.
#	Positive values indicate the row level results in greater log-reduction than the column level; negative values
indicate the column level results in greater log-reduction.
B-9

-------
Table B-22. Log-reduction ANOVA Summary Table Testing the Effect for B. anthracis
Strain and Material for Decontaminated Coupons.
Source
Degrees of
Freedom
Sum of
Squares
F Statistic
P-value
B. anthracis Strain
1
5.28
95.24
<0.001*
Material
3
11.73
70.56
<0.001*
B. anthracis Strain*Material
Interaction
3
11.73
70.56
<0.001*
Residual Error
32
1.77


Effect is statistically significant at a = 0.05 level.
Table B-23. Log-reduction Tukey Comparisons** for Material for B. anthracis Ames
Decontaminated Coupons.
Material
Granite
Painted Carbon Steel
Weathered Concrete
Ceramic Tile
0.00
0.00
0.00
Granite

0.00
0.00
Painted Carbon Steel


0.00
*	Differences are statistically significantly at a = 0.05 level.
#	Positive values indicate the row level results in greater log-reduction than the column level; negative values indicate the
column level results in greater log-reduction.
Table B-24. Log-reduction Tukey Comparisons** for Material for B. anthracis Sterne
Decontaminated Coupons.
Material
Granite
Painted Carbon Steel
Weathered Concrete
Ceramic Tile
0.00
2.50*
0.00
Granite

2.50*
0
Painted Carbon Steel


-2.50*
# Positive values indicate the row level results in greater log-reduction than the column level; negative values indicate the
column level results in greater log-reduction.
*Differences are statistically significantly at a = 0.05 level.
Table B-25. Log-reduction Tukey Comparisons for B. anthracis Strain by Material for
Decontaminated Coupons.
Material
Difference*
Ceramic Tile
0.10
Granite
0.10
Painted Carbon Steel
2.60*
Weathered Concrete
0.10
#	Positive values indicate B. anthracis Ames results in greater log-reduction than B. anthracis Sterne;
negative values indicate B. anthracis Sterne results in greater log-reduction.
*	Difference is statistically significantly at a = 0.05 level.
B-10

-------
110

100
~
90
•

~
80
~~

~~
70


*
60-
•

*
50
*
40
S
30
F
20

10

0

•	No Coat
~	Yes Coat
Ceramic Tile
%
*
i
<
~5
I
%
~
/~
~
*•
Granite
Painted Carbon Steel
Weathered Concrete
Material
Figure B-l. Plot of Control Coupon Percent Recovery of Inoculum by Material and
Simulated Subway Grime for B. anthracis Ames. Note That Percent Recovery Values
Greater than 200% Are Not Included in the Plot.
B-l 1

-------
d)
>
o
o
(U
ai
c
0
o
<5
0.
120
110
100
90
80
70
60
50
40
30
20
10
On
•	B. anthracis Ames
~	8. anthracis Sterne
}
%
4
~
t
~
*.
I
e
Painted Carbon S
%
\
%
**
i
Weathered Concrete
Material
Figure B-2. Plot of Control Coupon Percent Recovery of Inoculum by Material and Strain.
Note That Percent Recovery Values Greater than 200% Are Not Included in the Plot.
o
"O
0
EC
CT>
O
10
9
® I 0
7
6
5-
4
a. I
2§
1
O CT 2, Tomp 10, RH 75 O CT 3, Tcnrp 10, RH 75 O CT 4. Temp 10, RH 75 O CT 5. Temp 10, RH 75
CT7, Temp 4.5, RH75 • CT 7 Tetrp 10, RH 50 CT 7. Temp 10. RH 75 ~ CT 9. Temp 4.5, RH 50
i
~
o
o
o
o
o
o
Material
Coat
Figure B-3. Plot of Decontaminated Coupon Log (Base 10) Reduction of Inoculum by
Material, Simulated Subway Grime, Temperature, Relative Humidity, and CT.
B-12

-------
•	B. anthracis Ames
~	B. anthracis Sterne
~
~ ~
Ceramic Tile
Granite
Painted Carbon Steel
Weathered Concrete
Material
Figure B-4. Plot of Decontaminated Coupon Log (Base 10) Reduction of Inoculum by
Material and Strain.
B-13

-------
vvEPA
United States
Environmental Protection
Agency
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT NO. G-35
Office of Research and Development (8101R)
Washington, DC 20460
Official Business
Penalty for Private Use
$300

-------