Comparison of DNATrax and Bacillus anthracis Surrogate Resuspension from Subway Surfaces


EPA/600/R-20/113 | May 2020
www.epa.gov/homeland-security-research
United States
Environmental Protection
Agency
oEPA
Comparison of DNATrax and
Bacillus	anthracisSurrogate
Resuspension from Subway
Surfaces
Office of Research and Development
Homeland Security Research Program

-------
EPA/600/R-20/113
Comparison of DNATrax and Bacillus anthracis
Surrogate Resuspension from Subway Surfaces
Assessment and Evaluation Report
Homeland Security and Materials Management Division
Center for Environmental Solutions and Emergency Response
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711

-------
EPA/600/R-20/113
Comparison of DNATrax and Bacillus anthracis
Surrogate Resuspension from Subway Surfaces
Assessment and Evaluation Report
EPA CESER/HSMMD Technical Lead: John Archer
Homeland Security and Materials Management Division
Center for Environmental Solutions and Emergency Response
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
ii

-------
Disclaimer
The U.S. Environmental Protection Agency (EPA), through its Office of Research and
Development's Center for Environmental Solutions and Emergency Response (Homeland
Security and Materials Management Division), directed and managed this research through
Contract No. EP-C-15-008 with Jacobs Technology, Inc. The research described in this study
has been funded wholly or in part by the U.S. Department of Homeland Security (DHS) Science
and Technology Directorate (S&T) under an interagency agreement (EPA No. RW-070-
95935001 and DHS S&T IA# 70RSAT18KPM000176).
This report has been peer and administratively reviewed and approved for publication as an
EPA document. This report does not necessarily reflect the views of the EPA. No official
endorsement should be inferred. This report includes photographs of commercially available
products. The photographs are included for the purpose of illustration only. Any mention of trade
names, manufacturers or products does not imply an endorsement by the United States
Government or the EPA. EPA and its employees do not endorse any commercial products,
services, or enterprises.
Questions concerning this report or its application should be addressed to the following
individual:
John Archer, MS, CIH
Homeland Security and Materials Management Division
Center for Environmental Solutions and Emergency Response
U.S. Environmental Protection Agency (MD-E343-06)
Office of Research and Development
109 T.W. Alexander Drive
Research Triangle Park, NC 27711
Telephone No.: (919) 541-1151
Fax No.: (919) 541-0496
E-mail Address: archer.iohn@epa.gov

-------
Foreword
The Center for Environmental Solutions and Emergency Response (CESER) within the Office of
Research and Development (ORD) conducts applied, stakeholder-driven research and provides
responsive technical support to help solve the Nation's environmental challenges. The Center's
research focuses on innovative approaches to address environmental challenges associated
with the built environment. We develop technologies and decision-support tools to help
safeguard public water systems and groundwater, guide sustainable materials management,
remediate sites from traditional contamination sources and emerging environmental stressors,
and address potential threats from terrorism and natural disasters. CESER collaborates with
both public and private sector partners to foster technologies that improve the effectiveness and
reduce the cost of compliance, while anticipating emerging problems. We provide technical
support to EPA regions and programs, states, tribal nations, and federal partners, and serve as
the interagency liaison for EPA in homeland security research and technology. The Center is a
leader in providing scientific solutions to protect human health and the environment.
iv

-------
Acknowledgments
The principal investigator from the Office of Research and Development's Center for
Environmental Solutions and Emergency Response (CESER), Homeland Security and Materials
Management Division (HSMMD) directed this effort with support of EPA and interagency project
teams. Special thanks to the EPA Microbiology Laboratory (BioLab) staff for all of their support
with microbiological sample analysis. Contributions of the following individuals were a valued
asset throughout this effort:
U.S. EPA Principal Investigator
John Archer, CESER/HSMMD/Disaster Characterization Branch (DCB)
U.S. EPA Technical Reviewers
Sang Don Lee, CESER/HSMMD/Wide Area Infrastructure and Decontamination Branch
(WAIDB)
Shannon Serre, Office of Emergency Management (OEM)/Chemical, Biological,
Radiological and Nuclear (CBRN)/Consequence Management Advisory Division (CMAD)
External Technical Reviewers
Marshall Gray, Exposure Reduction Consultants, LLC
Brian Lee, Massachusetts Institute of Technology Lincoln Laboratory (MIT LL)
U.S. EPA Research Team
M. Worth Calfee, CESER/HSMMD
Katherine Ratliff, CESER/HSMMD
Leroy Mickelsen, OEM/CBRN/CMAD
Katrina McConkey, CESER/HSMMD (ERG contractor)
Jacobs Technology, Inc.
D. Adam Hook
Robert Yaga
Jerome Gilberry
DHS S&T Project Manager
Donald Bansleben, DHS S&T, Office of Mission Capability and Support (MCS)
Interagency Project Team
Jane Tang, DHS S&T (contractor)
Andrea Wiggins, DHS S&T (contractor)
Benjamin Ervin, MIT LL
Meghan Ramsey, MIT LL
Frances Nargi, MIT LL
Mandeep Virdi, MIT LL
Trina Vian, MIT LL
Elizabeth Wheeler, Lawrence Livermore National Laboratory (LLNL)
Ellen Raber, LLNL
Matthias Frank, LLNL
David Brown, ANL
U.S. EPA Quality Assurance Reviewer
Ramona Sherman, CESER
v

-------
Contents
Disclaimer	iii
Acknowledgments	iv
Executive Summary	1
1.0 Introduction	1
1.1	Background	1
1.2	Objectives	1
2.0 Experimental Approach	2
3.0 Experimental Materials and Methods	4
3.1	Test Materials	4
3.2	Environmental Systems	5
3.3	Test Surrogates and Preparation	6
3.3.1	DNATrax	8
3.3.2	Btk	8
3.4	Surrogate Particle Deposition System	9
3.5	Resuspension Wind Tunnel	11
3.6	Test Surrogate Enumeration	14
3.6.1	DNATrax	14
3.6.2	Btk	15
4.0 Dry Deposition Evaluation	17
4.1	Test Matrix	17
4.2	Dry Deposition Testing Approach and Procedure	17
5.0 Resuspension Comparison Tests	21
5.1	Resuspension Text Matrix	21
5.1.1	Filter Sampling of Resuspended Surrogate Materials	21
5.1.2	Particle Size Measurements of Resuspended Material	22
5.2	Resuspension Testing Approach and Procedures	22
5.2.1	Coupon Preparation and Particle Seeding	22
5.2.2	Filter Sampling Resuspension Procedure	23
5.2.3	APS Measurement of Resuspended Material Procedure	25
6.0 Results and Discussion	27
6.1	Dry Deposition System Performance	27
6.1.1	APS Size Distribution and SEM Analysis	27
6.1.2	Deposition Spatial Distribution and Load Variation	32
6.2	Resuspension Testing	33
6.2.1	Resuspended Particle Size Distribution	33
6.2.2	Resuspension Fraction Comparison between Surrogates	36
6.3	Summary and Conclusions	40
7.0 Quality Assurance (QA) and Quality Control (QC)	42
7.1	Equipment Calibration	42
7.2	QA/QC Checks	42
7.3	Data Quality Objectives	43
References	45
Appendices	47
vi

-------
Figures
Figure 3-1. Concrete (left) and Stainless Steel (right) Coupons for Resuspension Testing 4
Figure 3-2. T and RH Controlled Environmental Storage Chamber 5
Figure 3-4. Micro-Venturi Eductorwith Flow Field 10
Figure 3-5. Micro-Venturi Eductor Fitted with Large Particle Separator 10
Figure 3-6. Dry Particle Settling Chamber and Internal Mixing Fans 11
Figure 3-7. RWT Positioned in T and RH Controlled AWT 12
Figure 3-8. Four Filter Holders Mounted in RWT 12
Figure 3-9. RWT Particle Sizing Extension with APS Nozzle 13
Figure 3-10. Air Knife and Actuator 13
Figure 4-1. Distribution Disc Sampling Array 19
Figure 6-1: DNATrax Number/Volume Relative and Cumulative Distributions 27
Figure 6-2. Overview Scan of Deposited DNATrax (A) and 1,8-|jm Particle (B) 29
Figure 6-3. SEM Size Measurement and APS Comparison for DNATrax 30
Figure 6-4: Btk Number/Volume Relative and Cumulative Distributions 31
Figure 6-5. Overview Scan of Btk Deposition and Single Spore 32
Figure 6-6. Coupon Spatial Distribution Heat Maps: A) DNATrax, B) Btk 33
Figure 6-7. Size Distribution - Resuspension versus Deposition (Top) DNATrax (Bottom) Btk 35
Figure 6-8. Cumulative Size Distributions: Resuspension versus Deposition 36
Figure 6-9. Stainless Steel Resuspension Fraction 37
Figure 6-10. Concrete Resuspension Fraction 38
Figure 6-11. Diurnal Cycle Temperature and RH for: (Left) SS and (Right) Concrete 39
Figure 6-12. Resuspension Fraction Per Size Bin SS 30% RH 40
Tables
Table ES-1. Summary of Resuspension Fraction Findings [Mean (SD)] 2
Table 3-1. Ba, Btk, and DNATrax Properties 7
Table 4-1. Deposition Test Matrix 17
Table 5-1. Filter Collection Resuspension Test Matrix 21
Table 6-1. Deposition Variation and Particle Counts per mg of Particles Deposited 33
Table 6-2. Average Surrogate Resuspension Fractions under Varied RH [Mean (SD)] 38
Table 7-1. QA/QC Checks and DQIs 43
Table 7-2. DQOs for Critical Measurements 44
Table A-1. Stainless Steel RF Compendium 48
Table A-2. Concrete RF Compendium 48
vii

-------
Acronyms and Abbreviations
jjm
Micron or micrometer
ANL
Argonne National Laboratory
APS
Aerodynamic Particle Sizer
ATF
Aerosol Test Facility
AWT
Aerosol Wind Tunnel
Ba
Bacillus anthracis
B|_im
Barcodes/jjm3
BioLab
EPA Microbiology Laboratory
Btk
Bacillus thuringiensis var. kurstaki
°C
Degree(s) Celsius
CBRN
Chemical, Biological, Radiological, and Nuclear
CCSEM
Computer-Controlled Scanning Electron Microscope
CCDC CBC
Combat Capability Development Command Chemical Biological

Center
CESER
Center for Environmental Solutions and Emergency Response
CFU
Colony Forming Unit(s)
CMAD
Consequence Management Advisory Division
CV
Coefficient of Variation
DCB
Disaster Characterization Branch
DE
Dextrose equivalent
DHS
US Department of Homeland Security
DNA
Deoxyribonucleic Acid
D NATrax
DNA Tagged Reagents for Aerosol experiments
DQI
Data Quality Indicator
DQO
Data Quality Objective
EPA
US Environmental Protection Agency
°F
Degree(s) Fahrenheit
HEPA
High-efficiency particulate air
HSMMD
Homeland Security and Materials Management Division
LL
Lincoln Laboratory
LLNL
Lawrence Livermore National Laboratory
Lpm
Liter(s) per minute
MIT
Massachusetts Institute of Technology
MMAD
Mass Median Aerodynamic Diameter
OEM
Office of Emergency Management
PBST
Phosphate-buffered saline with 0.05% Tween® 20

-------
Pc	Particles collected
Pd	Particles deposited
psi	Pound(s) per square inch
QA	Quality Assurance
QC	Quality Control
QAPP	Quality Assurance Project Plan
qPCR	Quantitative Polymerase Chain Reaction
ReBoUndS	Resuspension of Bacillus anthracis surrogates on Underground
Subway Surfaces
RF	Resuspension fraction
RH	Relative Humidity
RPM	Revolution(s) Per Minute
RTP	Research Triangle Park
RWT	Resuspension Wind Tunnel
S&T	Science and Technology Directorate
SEM	Scanning electron microscope
SPORE	Scientific Program On Reaerosolization and Exposure
SD	Standard Deviation
SS	Stainless Steel
T	Temperature
Tdna	Total DNA Barcodes
TP	Total Particles
Tv	Total DNATrax volume
Vi%	Particle Volume Percent
V	Particle Volume
WAIDB	Wide Area Infrastructure and Decontamination Branch
ix

-------
Executive Summary
This project supports the interests of the US Environmental Protection Agency (EPA) and the
Department of Homeland Security Science and Technology Directorate (DHS S&T).
Specifically, this research supports their missions to understand the spread of biological
contaminants due to a bioterrorism incident after the contaminants settle onto surfaces in an
urban environment.
The main objective of this study was to compare the resuspension of an inert sugar-based
surrogate, known as DNA Tagged Reagents for Aerosol experiments (DNATrax) developed by
Lawrence Livermore National Laboratory (LLNL), under variable humidity conditions for an
underground subway system to the resuspension of an established biological surrogate of
Bacillus anthracis (Bacillus thuringiensis var. kurstaki [Btk]). The two surrogates were deposited
on representative porous and nonporous subway surfaces, and we determined under what
conditions, if any, DNATrax is an appropriate surrogate for Bacillus anthracis (Ba) when
resuspension is considered. A dry powder micro-eductor deposition system was developed to
accurately and repeatedly deposit the materials at the desired coupon load, spatial distribution,
and with the majority of particles in their singlet and doublet forms with reduced agglomeration.
Resuspensions were conducted in a custom wind tunnel using a wind shear velocity of
approximately 45 mph and particles immediately captured onto filters.
Results from the comparative resuspension testing are shown below in Table ES-1. Quantitative
polymerase chain reaction (qPCR) was used to quantify DNATrax particles and culture counting
methods were used for Btk. Btk resuspension data should be treated as a maximum because
culture counting of colony forming units could potentially lead to overestimation due to
agglomerate disassociation. There was no statistically significant difference in resuspension
fractions (RFs) from deposited Btk and DNATrax for the 30% and 80% relative humidity (RH)
environmental test conditions. In contrast, the variable diurnal cycle (30% —~ 85% —~ 30% for
stainless steel (SS); 30% —~ 80% —~ 30% for concrete) RH conditions resulted in statistically
significant differences in RFs between the two surrogates.
For the surrogate materials on stainless steel following a diurnal cycle, the RF of DNATrax was
reduced significantly below the RF of Btk. However, for concrete, the RF for Btk was reduced to
significantly less than the RF of DNATrax. The lower RF for DNATrax on stainless steel is
hypothesized to be due to softening of the DNATrax material due to the RH extending beyond
its glass transition condition and potentially leading to greater adhesion to stainless steel. Glass
transition is defined as the transition in amorphous materials from a hard and relatively brittle
"glassy" state into a viscous or rubber state. This transition can be abrupt or gradual. For
concrete, the conditions remained below the glass transition, which may explain the higher RF.
The mechanism for the lower Btk RF on concrete is undetermined, though this lower Btk RF on
concrete could potentially be due to a combination of the surface roughness of the concrete
combined with the exosporium (hairy nap) on the surface of the Btk, which is not present on
DNATrax. The variability in results makes it difficult to draw strong conclusions and
recommendations. However, we hypothesize from the results that combinations of temperatures
and relative humidities greater than the glass transition of DNATrax or the presence of standing
water could lead to significant differences in the resuspension of DNATrax and Btk/Ba. For
conditions that remain well below the glass transition (-80% RH), DNATrax (specifically
formulated to be similar in size to Ba) is statistically indistinguishable from Btk (representative
ES-1

-------
Ba surrogate). Therefore, based on current data there is no evidence to suggest that the
resuspension of the DNATrax formulation examined here would be significantly different than
Ba under conditions well below the glass transition. Future work is recommended to examine
the high variation in RFs of DNATrax under all conditions, especially following RH cycles,
saturated porous substrates, and rain events.
Table ES-1. Summary of Resuspension Fraction Findings [Mean (SD)]
Surface, %RH Condition
DNATrax
Btk
SS, 30%
6.32 (±7.42) %
2.26 (±0.91) %
SS, 80%
1.79 (±1.88) %
0.73 (±0.47) %
SS, 30% -> 85% -> 30%
0.31 (±0.21) %
2.77 (±0.84) %
Concrete, 30%
8.36 (±7.92) %
3.75 (±1.00) %
Concrete, 80%
9.49 (±7.10) %
3.25 (±1.10) %
Concrete, 30% —> 80% —> 30%
12.76 (±2.5) %
0.35 (±0.36) %
SD = standard deviation
SS = Stainless Steel
Highlight across row indicates statistically significant difference at 95% Confidence Level
ES-2

-------
1.0 Introduction
1.1 Background
Release of a biological agent aerosol in an urban area has the potential for widespread
contamination and risk to civilian life. Human activity and complex wind dynamics make
determining levels and locations of contamination difficult. To coordinate evacuation and
remediation efforts to protect human health, the US Environmental Protection Agency (EPA)
and Department of Homeland Security Science and Technology Directorate (DHS S&T) are
tasked with determining where these agents may be transported and settle. To that end, these
agencies have employed computational contaminant transport models to examine multitudes of
release scenarios. One such scenario is a biological agent release in an urban subway system.
High wind speeds associated with train traffic in confined spaces and human activities such as
walking may allow biological agents such as Bacillus anthracis (Ba) to travel considerable
distances after an initial release, resuspend after depositing on surfaces, and be transported on
human clothing (fomite transport) far away from the initial release point. To validate these
contaminant transport models, it is necessary to conduct surrogate releases inside real-life
subway environments. However, it is preferable if not mandatory to utilize surrogates for Ba that
are inert and non-pathogenic due to safety (and perceived safety) concerns. One such inert
nonbiological surrogate was developed by Lawrence Livermore National Laboratory (LLNL) to
assist in validating these aerosol transport models, and this surrogate is known as DNA Tagged
Reagents for Aerosol experiments (DNATrax). However, the degree to which DNATrax
interacts with and resuspends from surfaces similarly to Ba is currently unknown. This
information will aid in understanding how DNATrax, an inert nonbiological surrogate, compares
to biological agents and biological spore surrogates in terms of resuspension from surfaces and
will provide a basis for interpretation of future DNATrax test data.
1.2 Objectives
The purpose of the research conducted under the Resuspension of Bacillus anthracis
Surrogates on Underground Subway Surfaces (ReBoUndS) project was to fill some of the key
knowledge gaps whether DNATrax is a representative surrogate for actual biological spores,
e.g., Ba, specifically, for resuspension/reaerosolization. The primary objective of this work was
to evaluate and statistically compare reaerosolization of surrogate spores, Bacillus thuringiensis
var. kurstaki (Btk), and DNATrax from surfaces and materials found in subways under typical
environmental conditions. To meet the primary objective, this study includes a secondary
objective of developing a repeatable small-scale dry deposition method for inoculation of
subway materials with DNATrax and the Ba surrogate Btk.
1

-------
2.0 Experimental Approach
This project involved determining the degree to which the resuspension of dry-deposited
DNATrax under various environmental conditions is similar to a dry-deposited Ba surrogate, Btk.
Previous experiments conducted at the EPA under the Scientific Program on Reaerosolization and
Exposure (SPORE) demonstrated that Btk is a suitable Ba surrogate for reaerosolization studies
(EPA 2014). Thus, the conditions under which DNATrax resuspends similarly to Btk are the same
conditions where DNATrax is an acceptable surrogate for Ba in field studies. In order to make such
a comparison, the parameter Resuspension Fraction is determined experimentally. Resuspension
Fractions (RFs) are calculated by the following equation:
where Pc is the number/mass of particles collected after a resuspension experiment, and Pd is the
number/mass of particles deposited onto the surface used for resuspension testing. Therefore,
two things must be measured: the amount of material deposited onto the test coupon surface and
the amount of material resuspended from the test coupon surface. Due to the size and complexity
of the surfaces examined, non-destructive methods of enumerating deposited materials such as
optical microscopy or fluorescence could not be used. Therefore, prior to the resuspension
studies, characterization studies of the depositions were carried out. These studies were used to
show that the materials were deposited similarly and that small reference discs could be used as
an estimate for the coupon surface coverage. The test surfaces chosen for this study were
concrete and stainless steel, to represent a porous and nonporous surface, respectively. Both
surfaces are found throughout a subway system and are representative of the surfaces most likely
to be exposed to surface stresses sufficient to resuspend particulate matter. Relative humidity
(RH) can be a major factor in the ability of a particle to resuspend. Differences in surface
characteristics of the particles and adhering surfaces can potentially cause different resuspension
characteristics with a change in humidity. Under normal conditions, particles can have a thin liquid
layer on the surface, creating a capillary adhesive force between the particle and the surface. The
curvature and roughness of the particle and adhering surface can affect the magnitude of that
force (Hinds 1999). In addition, reduction in RH from a state in which capillary formation has
occurred can cause the particle to be pulled closer to the surface as the capillary recedes,
resulting in an increase in the adhesion force. Therefore, a comparative study must also include
varying the relative humidity both statically (before deposition) and dynamically (after deposition).
The general experimental approach used to meet the project objectives is described below:
1. Evaluation of eductor-based laboratory-scale dry deposition system. To evaluate
the eductor-based deposition system repeatability, multiple tests were conducted:
a. The repeatability of the size distribution exiting the eductor was measured by
repeated particle size measurements using an Aerodynamic Particle Sizer (APS)
and both DNATrax and Btk.
b. The repeatability and uniformity of the dispersion of test material over the surface
of the test coupons was measured by sampling an array of discs at the bottom of
the deposition chamber for multiple depositions.
2

-------
c.	The repeatability of the total surface load was examined by sampling stainless-
steel surfaces after deposition. A nominal load of 1x107 particles (1.5x10s
particles/in2) for both surrogates was desired and the amount of each surrogate
material required to achieve this goal was established. Variation was desired to
be within an order of magnitude and each coupon's load for resuspension was
measured.
d.	Scanning electron microscopy (SEM) imaging was conducted to show that
particles on surfaces were dominated by singlets and doublets, and
agglomeration of particles was minimized so that particle-to-particle comparisons
could be made.
2. Comparative resuspension of particles on subway materials (coupons) and
conditions. To compare resuspension of test particles, the general testing procedure is
shown below:
a.	Subway material coupons were equilibrated to test conditions for a minimum of
24 hours prior to deposition of DNATrax or Btk.
b.	The subway material coupons for resuspension were inoculated or seeded with
the surrogate particles under controlled conditions at a load of approximately 1
x107 particles (1.5x10s particles/in2).
c.	Subway material coupons were subjected to a wind shear stress in a customized
resuspension wind tunnel normalized to test environment conditions, and
resuspended material was captured onto filters.
d.	Filters were removed from the system and test particles extracted into fluid and
enumerated either via quantitative polymerase chain reaction (qPCR) (for
DNATrax) or culture techniques (for Btk).
e.	Additional resuspension tests were conducted without filter capture to estimate
the size distribution of resuspended test particles by using an aerodynamic
particle sizer (APS) to aid in particle enumeration and characterization.
f.	Resuspension fractions of surrogate materials were calculated by dividing the
number of resuspended particles by the number of deposited particles.
g.	Statistical analysis was conducted to compare resuspension fractions of test
surrogates directly.
3

-------
3.0 Experimental Materials and Methods
This section describes the test materials, environmental chambers, test surrogates, deposition
system, resuspension wind tunnel, and particle counting methods used to achieve the project
objectives.
3.1 Test Materials
The representative subway materials chosen for this test were stainless steel and concrete.
Coupons of 7.75" x 7.75" x 0.75" of these materials were created in bulk at the EPA Research
Triangle Park (RTP) facility to ensure uniformity of material surfaces. Stainless steel coupons
were created by attaching 7.75" x 7.75" sheets of 22-gauge #4 polished stainless steel to 7.75"
x 7.75" x 0.75" thick plywood with spray adhesive (Super 77™, 3M™, Maplewood, MN).
Concrete coupons were created by mixing dry concrete mix (Quikrete® Concrete Mix PN 1101,
The QUIKRETE Companies, Atlanta, GA) to manufacturers' specifications at an approximate
10:1 by weight mix-to-water ratio, pouring into 7.75" x 7.75" x 0.75" molds and curing for five
days at 70 degrees Fahrenheit (°F) and 30% RH. The concrete coupons were removed from the
molds and allowed to continue curing in an environmental chamber at 20 degrees Celsius (°C)
and 30% RH for a minimum of two weeks before use. Figure 3-1 shows completed stainless
steel and concrete test coupons. Loose particles were removed from the concrete surface by
spraying with an air nozzle at 30 pounds per square inch (psi) compressed house air prior to
particle seeding/inoculation. Stainless steel coupons were wiped clean with a lint free cloth and
methanol prior to particle seeding/inoculation.
Figure 3-1. Concrete (left) and Stainless Steel (right) Coupons for Resuspension Testing
4

-------
3.2 Environmental Systems
To achieve specific environmental conditions for material coupon storage/conditioning and
resuspension testing, the EPA RTP facility has multiple environmental chambers to ensure
controlled conditions and containment of particles. These chambers include an environmental
test/conditioning chamber (shown in Figure 3-2) used for storage and conditioning of test
material coupons and the EPA's recirculating aerosol wind tunnel (AWT, shown in Figure 3-3)
used to house a small resuspension wind tunnel (Section 3.5) for reaerosolization of materials
from test coupons and subsequent sampling for resuspended material and test coupon
surfaces. Both containment systems are high-efficiency particulate air (HEPA)-filtered and are at
negative pressure relative to the surrounding laboratory spaces. Both systems are temperature
(T)- and RH-controlled. All set conditions were monitored throughout the experiments via a
calibrated probe (VWR 35519-041, VWR International, Radnor, PA), HOBO Micro Station data
logger (Onset Computer Corp., Bourne, MA), and Humicap HMT330 (Vaisala Inc., Louisville,
CO, USA).
Figure 3-2. T- arid RH-Controlled Environmental Storage Chamber
5

-------
Figure 3-3. Plan view of the EPA Aerosol Wind Tunnel
3.3 Test Surrogates and Preparation
For the resuspension comparison tests, two different Ba surrogates were used: DNA-barcoded
maltodextrin particles known as DNATrax (Section 3.3.1) and barcoded Btk (Section 3.3.2). The
nominal properties of each surrogate from literature are discussed in the following sections and
summarized in Table 3-1 along with the properties of the Ames strain of Ba. Images of DNATrax
and Btk in Table 3-1 are of the materials used and properties measured for the specific batches
of DNATrax and Btk used in this study and are discussed in the results.
6

-------
Table 3-1. Ba, Btk, and DNATrax Properties
Ba Ames f
Physical Shape
Capped Cylinder1*

3k.
m
Capped Cylinder
DNATrax
Spheroid
Median
Dimension
~1.53 jjrn x 0.81 jjrm
-1.61 |jmx0.80 pin
~1.83 |jm*
Cross Sectional
Area
1.10 (jm2 max
0.52 |jm2 min
1.15 |jm2 max
0.5 |jm2 min
2.63 jjm2
Volume
0.58 |jm3
0.608 p3
3.2 |jm3
Aspect Ratio
-1.89
-2.01
-1-1.2
Surface
Roughness
Hairy Nap
Hairy Nap
Smooth
Density
Dry - 1.42 g/cm3
Dry —1.4 g/cnr3
(assumed)
1.54 g/cm3
*Mass Median Aerodynamic Diameter
f Ba Ames Data from Carre ra et a I 2007
ff Ba Ames Image from EPA 2012
References for DNATrax and Btk properties are in Section 3.3.1 and 3.3.2, respectively.
7

-------
3.3.1 DNATrax
DNATrax was developed by Lawrence Livermore National Laboratory (LLNL) as a non-toxic test
particle for aerosol fate and transport experiments in populated areas where the use of other Ba
surrogates is restricted. The DNATrax particles are created by spray drying an aqueous solution
of maltodextrin containing copies of a short DNA barcode (-100 base pairs) to achieve
nominally spherical particles. The size of these particles can be varied depending on specific
application, with small particles (~2 |jm) being made of pure maltodextrin and particles on the
order of 5-10 micrometers (|jm) being achieved by seeding with silica particles. In the current
experiments, the 2-|jm nominal particle size was chosen as particles of this size are more
similar to the size of Btk and have less of a potential for material shedding. The DNATrax
samples were received from LLNL with a projected mass mean aerodynamic diameter of 1.83
|jm. This mass mean aerodynamic diameter gives a cross sectional area of approximately 2.63
|jm2 and a volume of approximately 3.2 |jm3, assuming a perfect sphere. Specifications from
LLNL were that the average 2-|jm particle contains 469 DNA fragments for qPCR counting. The
maltodextrin chosen for the DNATrax has a dextrose equivalent (DE) of 10 and a nominal
density of 1.54 (g/cm3) (Kaeser 2017). The density of Ba, however, is approximately 1.43 g/cm3,
making DNATrax slightly heavier on average than Ba (EPA 2012). A DE of 10 makes DNATrax
moderately resistant to absorption of water vapor onto the surface and into the bulk of the
particle. This resistance to absorption also increases its resistance to softening due to a glass
transition. Glass transition is defined as the transition in amorphous materials from a hard and
relatively brittle "glassy" state into a viscous or rubber state. This glass transition is abrupt as the
water absorption isotherm exhibits a cascade absorption at the critical RH and temperature
(Abramovic 2002). Materials with a DE of approximately 10 have a glass transition at 20 °C at
approximately 80% RH, as the RH is lowered, the glass transition temperature increases to 65%
RH/34 °C, 53% RH/41 °C, etc. (Nurhadi 2016). Once sufficient water is present on the surface,
DNATrax becomes fully soluble, making deoxyribonucleic acid (DNA) extraction for qPCR highly
efficient, though the presence of water on a deposition surface could destroy particle integrity
(Nurhadi 2016). The surface of DNATrax particles is nominally smooth under SEM analysis
(Harding 2016). DNATrax material was stored in an environmental chamber set to 20 °C and
30% RH when not in use. In addition, the tubes containing the material were kept in a Ziploc®
bag filled with desiccant to further lower the RH.
3.3.2 Btk
The barcoded Btk used for this project is a genetically modified strain that was developed by the
U.S. Army Combat Capabilities Development Command Chemical Biological Center (CCDC
CBC, Aberdeen Proving Ground, MD) to allow the spores to be distinguished from naturally
occurring Btkvia qPCR analysis. The barcoded Btk was obtained from Dugway Proving Ground
(Dugway, UT) as a dry powder of lyophilized spores. Btk is a gram-positive, spore-forming, rod-
shaped (hemisphere capped cylinder) bacterium found in soil and has been used extensively in
the past as a surrogate for Ba as many of the physical properties are identical (Carrera 2007;
EPA 2012) . Btk spores have a length of approximately 1.61 |jm and a width of approximately
0.8 |jm (EPA 2012). Assuming a hemispherical capped cylinder, these measurements give an
average cross-sectional area of 1.15 |jm2 at maximum, 0.5 |jm2 at minimum, and a volume of
0.608 |jm3. The dried density of Btk has not been measured directly. However, the wet density
8

-------
is identical to Ba at 1.17 g/cm3. Other bacterial spores of similar size and shape matching the
wet density of Ba match the dry density. Therefore, we assumed that the dry density of Btk is
very close to the dry density of Ba (EPA 2012) . Btk and Ba spores are covered in a hairy nap,
making the surface rough on the nanometer scale (Plomp 2005; Tufts 2014). Btk and Ba have
been shown to be relatively hydrophobic compared to other Bacillus species. However, Btk and
Ba have been shown to change sizes by 4-10% over both semi-major and semi-minor axes as
relative humidity changes from dry (-3% RH) to >95%. This expansion corresponds with a
decrease in particle density from dry to wet spores of approximately 1.42 g/cm3 to 1.17 g/cm3.
The spore core/cortex has been shown to be the main driver of this expansion as the spore-coat
surface area remains relatively constant compared to the volume, i.e., under dry conditions, the
spore coat has nano-sized wrinkles underneath the hairy nap and smooths when the spore is
expanded (Plomp 2005; Carrera 2007; Westphal 2003). The onset RH of this expansion has not
currently been explored. It is possible that this increased smoothness could cause a larger
contact area and decrease resuspension for hydrated spores compared to dry spores.
3.4 Surrogate Particle Deposition System
The surrogate dry particle deposition system developed for these experiments is shown in
Figures 3-4, 3-5, and 3-6. The eductor injection system was designed based on the large-scale
system used for DNATrax dispersion experiments previously conducted, but the system was
reduced in size to laboratory scale with a lower flow rate to deposit smaller masses of particles
(Kaeser 2017). The settling chamber is similar to the chamber used for spray dry depositions in
previous Btk resuspension experiments (EPA 2014). The method of operation uses a
compressed air flow into the eductor, which causes a vacuum at the particle inlet. This vacuum
causes air to rush into the bottom of the eductor, carrying particles through the eductor system
into the settling chamber. Figure 3-4 shows the base microflow eductor system, which consists
of a microflow venturi eductor (Micro-Flo Eductor, Jacobs Process Analytics, Inc., Williamsburg,
VA), a dry compressed air connection, an inlet for particle introduction, and a 90-degree curved
exit tube (elbow) for particle introduction into the settling chamber. The vacuum flow from the
inlet is approximately 1.3 liters per minute (Lpm), and the total flow into the settling chamber is
approximately 2 Lpm. Figure 3-5 shows the system with a more complex particle inlet into the
eductor for elimination of large particles/agglomerates. This system contains two additional 90-
degree bends and a 2" diameter x 1" tall cylindrical settling chamber prior to the eductor particle
inlet. The flows of this system are identical to the flows in the base system. Figure 3-6 shows
the settling chamber with internal mixing fans. The settling chamber is 18" x 8" x 8" to contain
the 7.75" square material coupons and allow thorough mixing of particles prior to settling over
three hours. Scouting experiments were used to determine the proper setup for deposition prior
to validation for each test surrogate. It was shown that DNATrax could not be delivered at high
enough surface concentrations using the complex setup in Figure 3-5. Therefore, the base
system was used (Figure 3-4). Proper material preparation and the violent nature of the eductor
was demonstrated to be sufficient to remove major agglomerates. The system in Figure 3-5 was
used for Btk deposition/inoculation.
9

-------
Figure 3-4. Micro-Venturi Eductorwith Flow Field
Figure 3-5. Micro-Venturi Eductor Fitted with Large Particle Separator
10

-------
Figure 3-6. Dry Particle Settling Chamber and Internal Mixing Fans
3.5 Resuspension Wind Tunnel
The resuspension wind tunnel (RWT) shown in Figure 3-7 is 6 feet long with an approximately 9
in x 9 in cross section. The RWT is open-ended by design and thus can take advantage of the T
and RH environmental settings available inside the environmental chamber. When operating
under total collection of reaerosolized material from the test surface, the tunnel air is continually
pulled through four polyester felt filters (EQXSCIEN-001, Superior Felt and Filtration, Ingleside,
IL, USA) mounted in custom filter holders shown in Figure 3-8 by a blower at a nominal velocity
of 2.5 meters per second (5.5 miles per hour [mph]). With the felt filters removed, the RWT can
be fitted with an aerodynamic particle sizer (APS) (TSI 3321, TSI Inc., Shoreview, MN) to
measure the size distribution of resuspended particles. The RWT is fitted with a HERA filter on
the inlet to provide clean sweeping air into the RWT and another HERA filter just in front of the
outlet to provide clean air and prevent contamination of the outer experimental chamber and the
blower. The RWT is constructed of aluminum and stainless steel to allow for easy
decontamination. The top of the RWT has a door that provides access to load a coupon. The
outlet of the tunnel is hinged to provide access to the filter holders. In addition, the hinged door
can be removed so that the RWT can be fitted with an extended chamber and APS sampling
nozzle for particle sizing experiments as shown in Figure 3-9. Resuspension shear forces are
achieved by a custom air knife mounted to an actuator as shown in Figure 3-10. The air knife is
connected to a clean high-pressure dry air line to prevent particle impingement onto the surface
through the air knife. The actuator traverses the air knife over the coupon at a steady speed.
The height of the jet, angle of the air impinging on the coupon, and the speed of the actuator
can all be adjusted as desired. In these experiments, the shear air speed across the coupon
11

-------
was set to 45 mph to simulate subway conditions when trains are actively moving, and the
knife traversed the coupon in 60 seconds or 0.13 inches/second.
Figure 3-7. RWT Positioned in T- and RH-Controlled AWT
Figure 3-8. Four Filter Holders Mounted in RWT
12

-------
Figure 3-9. RWT Particle Sizing Extension with APS Nozzle
mil \
Actuator

-------
3.6 Test Surrogate Enumeration
3.6.1 DNATrax
As stated previously, the DNATrax particles were tagged with a DNA barcode for enumeration
via quantitative polymerase chain reaction (qPCR). Collected material, either on filters or
aluminum deposition reference discs, was dissolved in phosphate buffered saline containing
0.1% Triton X-100 (Sigma-Aldrich, St. Louis, MO) and processed using a qPCR machine (ABS
7500 Fast, Applied Biosystems, Bedford, MA) using standard practices (Kaeser 2017). A
standard curve was generated daily using DNA sequences and concentrations provided by
LLNL. The resulting output from processed samples was compared with the standard to provide
the total number of DNA barcodes in each sample.
The process for converting the total number of recovered DNA barcodes into a number of
DNATrax particles requires some assumptions and additional measurements. The basis for the
enumeration rests on knowledge of the number of DNA barcodes present in each particle based
on the spray-dry procedure. Information provided by LLNL stated that a 2-|jm diameter particle
contains approximately 496 DNA barcodes. If the particles were monodispersed (single size),
then the process would have been to simply divide the total number of DNA barcodes by the
number of barcodes per particle. However, analysis of the DNATrax material received from
LLNL showed a somewhat broad polydispersed size distribution (via APS and SEM
measurements) ranging from <500 nm to 10 |jm with a mass peak between 1-2 |jm and 90% of
mass below 4 |jm. Therefore, some assumptions, additional measurements, and calculations
were made to avoid overestimating or underestimating particles depending on the measured
size distribution.
First, all particles were assumed to be spheres of uniform maltodextrin density (i.e., solid) and
second, the DNA barcodes were assumed to be evenly dispersed throughout each particle (469
barcodes/2-|jm sphere =112 barcodes/|jm3). Next, the size distribution of material that was to
be enumerated needed to be determined. For depositions, measurement was done during the
deposition process via an APS. For resuspended particles, however, as stated in the previous
section, APS measurements were conducted separately from tests that collected material for
qPCR, as the tunnel wind speeds were too great and material was removed from the chamber
too quickly for appropriate APS measurements to be conducted. In addition, the porous
concrete material shed small particles under wind shear, and those background particles would
likely have dominated the APS signal. Therefore, the size distribution from stainless steel
resuspension was used for all resuspension calculations.
Once the size distributions of the particles were established by APS, the count or number
distributions were converted to percent volume distributions as the amount of DNA was
dependent on the volume of each particle. The total number of DNA barcodes (Tdna) was
divided by the assumed number of DNA barcodes per |jm3 (BMm =112 barcodes/|jm3) to obtain
the total volume of DNATrax present in solution (Tv), Equation 3.1. This value was then
multiplied by the sum of each particle volume percent (Vi%) divided by the volume of the particle
(Vi) to obtain the total number of particles in solution (Tp), Equation 3.2. The APS displays
measured particle sizes sorted into 51 size bins, so the sum is over 51 to reflect all 51 particle
14

-------
volume bins. If the particles are solid, the actual number of barcodes per |jm3 (BMm) is
inconsequential when calculating resuspension fractions, as it is found multiplicatively in the
numerator and denominator of the RF equation after particle enumeration. However, if the
particles are not solid, the volume of DNA per particle would be dependent on the wall thickness
and may not be linearly related to the apparent volume of the particle. In that case, a counting
error would occur.
T T°NA it r o I
W = "s— Equation 3.1
xm
51
TP = Tv x V" — Equation 3.2
4-f Vt
1 = 1
3.6.2 Btk
Initially, we desired to enumerate Btk spores via qPCR, in a fashion similar to DNATrax.
However, we determined that agar plating and colony forming unit (CFU) counting was a more
robust and repeatable measurement of the number of spores than qPCR measurement for a
number of reasons. First, the small number of qPCR targets contained in each spore raises the
minimum countable range of spores to a level where the amount of material collected during
resuspension studies would likely be below the range of qPCR analysis. Second, the efficiency
of extracting DNA from a spore is not known and is likely very low due to the hard spore coat.
Experiments to estimate this efficiency would be prohibitively expensive and outside the realm
of this current work. Third, significant internal work has been conducted at EPA microbiological
laboratories to show that culture spore enumeration is a highly repeatable process with low
variability using triplicate plating. Therefore, the process for enumerating Btk particles in the
current experiments is much more simplified than DNATrax analysis. CFU enumeration does
not account for particles that are non-viable and how they contribute to resuspension. However,
non-viable particles are not counted when determining the number of spores deposited on the
surface. Thus, they do not contribute to either the numerator or denominator of the
resuspension fraction equation and would not bias the calculation. Spores that are viable but do
not germinate would be the same small fraction of deposited particles versus resuspended
particles and cancel out in the RF equation.
Btk spores were extracted from the filter collection media using a Stomacher® system (Seward
Ltd., Worthing, West Sussex, UK), spread onto agar plates, incubated, and the number of
colonies counted either automatically (QCount, Advanced Instruments, Norwood, MA ) or
visually by a laboratory technician, depending on the concentration. These processes are more
fully described in Section 5.2.2.
The major assumption with the Btk analysis is that agglomerates of particles do not break apart
when being extracted from the filters or aluminum deposition discs and form only a single
15

-------
colony. For deposition, nearly 90% of particles are below 1.6 |jm (singlets and doublets). Using
APS data from deposition in Section 6.1.1, treating particles as spheres, and knowing that
particle diameters increase as a cube function of the number of particles in the agglomerate, the
overestimation of particles deposited due to deagglomeration can be calculated as a factor of
-2 (Hinds 1999). For resuspension, however, as large particles are more easily resuspended,
deagglomeration would lead to a much larger overestimation of the number of resuspended
particles compared to deposited particles (EPA 2015; Hinds 1999). Using the same method for
resuspension as for deposition and the size distribution measured as shown in Section 6.2.1,
the maximum overestimation of resuspended particles would be a factor of -13. This method
gives an overall overestimation of the RF as a factor of -6.5, and any RF derived from CFU
enumeration should be treated as a maximum.
16

-------
4.0 Dry Deposition Evaluation
This section discusses the resuspension test matrix and approach for evaluation of the dry
deposition method. The dry deposition system was developed and evaluated for reliable and
repeatable deposition of dry particles to surfaces of interest.
4.1 Test Matrix
There were four main components to evaluation of the dry deposition method as listed in
Section 2. The first is confirmation that the eductor system produces a consistent size
distribution of particles, the second is evaluating the distribution of particles over the surface of
the coupons, and the third is determining the repeatability of surface loading and confirmation
that 1 x 107 particles surface loading can be achieved. Lastly, SEM confirmation of particle size
distribution was conducted. Each of the first three components consisted of five replicates to
obtain an estimate of the variability. The SEM analysis of deposited particles was only
conducted once due to time constraints. Since the APS data showed low variability and the
SEM size distribution was similar for the single analysis, it was determined that the SEM particle
size analysis was representative. Additionally, the SEM particle sizing was conducted to
corroborate the APS data and was not deemed a critical measurement. A summary of the
deposition test matrix can be found in Table 4-1.
Table 4-1. Deposition Test Matrix
Surrogate
Test
Replicates

APS Size Distribution
5
DNATrax
Particle Dispersion
5
Particle Load
5

SEM Analysis
1

APS Size Distribution
5
Btk
Particle Dispersion
5
Particle Load
5

SEM Analysis
1
4.2 Dry Deposition Testing Approach and Procedure
All of the above experiments occurred on the same day to minimize inter-day variability. The
particle dispersion and particle seeding/inoculation experiments were conducted concurrently,
and the APS measurements were conducted between each of the deposition replicate tests. All
experiments were conducted inside an environmental chamber set to 20 °C and 30% RH. For
each deposition dispersion/load test, a new array of nine 1.125" diameter aluminum foil discs
17

-------
was placed on top of a stainless steel pan at locations marked by a grid and labeled A-l as
shown in Figure 4-1. The settling chamber was placed over the top of the grid array, and one of
the two eductor arrangements was connected to the lid, as shown in Figure 3-4 for DNATrax
and Figure 3-5 for Btk. Immediately prior to deposition, the Btk and DNATrax were vigorously
agitated to break up the clumps and provide a more consistent powder of singlets and doublets.
For Btk, 500 mg of the lyophilized powder was placed into a plastic vortex tube with five 3-mm
glass beads and vortexed for two minutes in 30-second bursts. DNATrax was processed
similarly with only 3 beads and 200 mg of powder to reduce stress on the material. After
processing, approximately 2 mg of the processed powder to be tested was weighed, the mixing
fans inside the deposition chamber were turned on, and the material was injected into the
settling chamber. The mixing fan and eductor system were allowed to operate for five minutes
before the compressed air valve was closed, and the mixing fans were turned off. Material
inside the settling chamber was allowed to settle for three hours prior to collection of the
aluminum discs for analysis. Immediately following completion of material injection into the
settling chamber, the eductor system was removed, and the microflow portion was cleaned with
water and ethanol and allowed to dry. After cleaning, the eductor was reconnected to the
compressed air and inlet system. Once the components were reattached, the system was
placed directly over an APS inlet tube, and 2 mg of material was injected into the APS for five
minutes to allow time to obtain a particle size distribution, and the eductor was cleaned again.
This cycle was repeated five times for a total of 10 runs for surrogate material, five depositions
onto aluminum discs and five injections into the APS. For the final DNATrax deposition
experiment, an SEM stub (PN 16111, Ted Pella Inc., Redding, CA) was placed next to the
center aluminum disc (Labeled E) to collect settled material for analysis. For SEM imaging of
Btk, a 25-mm polycarbonate filter with a 0.5 |jm pore size was used to capture Btk material
exiting the eductor, and the filter was mounted to a carbon tape-coated SEM stub.
18

-------
Figure 4-1. Distribution Disc Sampling Array
After the three-hour settling time was complete, each deposition stack was removed, and each
aluminum disc collected into an individually labeled 50-ml_ vortex tube (Falcon 50 ml_, Corning
Inc., Corning, NY) so that disc position and deposition experiment number could be maintained.
Material was extracted from the discs by adding 5 mL of phosphate buffered saline containing
0.1% Triton X-100 (Sigma-Aldrich, St. Louis, MO) to the tube and vortexing for two minutes. For
DNATrax, qPCR analysis was then conducted to determine the number of DNA barcodes on
each disc and for Btk, culture plating was conducted to determine the number of CFU deposited
onto each disc.
The peak concentration of the APS signal during the size distribution measurements was used
as the benchmark for the size distribution for each run. The results were averaged, and
standard deviations were calculated. Once the average size distribution was determined for
DNATrax, it was used along with the qPCR results in the enumeration process for DNATrax
using the method found in Section 3.6.1.
Following determination of the number of particles on each disc, deposition heat maps were
created to show if any deposition contained highly skewed concentrations of surrogate material.
The total number of particles deposited for each run was determined by averaging the surface
concentration of the nine deposition discs and multiplying it by the total area of the stainless
steel or concrete coupon (60 in2).
Both Btk and DNATrax particles were sputter-coated in -30 nanometers of gold prior to SEM
analysis to prevent charging (Harding 2016). The imaging was conducted using a TESCAN
19

-------
Mira3 field emission SEM (Tescan USA, Inc., Warrendale, PA), and the computer controlled
particle analysis was done using IntelliSEM software (RJ Lee Group, Monroeville, PA).
20

-------
5.0 Resuspension Comparison Tests
This section discusses the test matrix and procedures for comparing the resuspension fraction
of DNATrax and Btk from select subway-type surfaces and conditions. All procedures discussed
in this section follow standards for quality required by EPA.
5.1 Resuspension Text Matrix
The resuspension tests consisted of two phases. The first phase was resuspended particle
collection on felt filters for enumeration, and the second phase was APS measurements of
resuspended particle size distributions.
5.1.1 Filter Sampling of Resuspended Surrogate Materials
The surfaces and conditions for resuspension comparison were chosen to reflect common
subway surface materials and the environmental conditions the settled particles may
experience. Three main variables were represented in this study: particles, surface types, and
environmental conditions. The traverse speed of the air knife over the coupon (0.13 in/second)
and the air shear speed (45 mph angled at 15° below horizontal) were kept constant. The
surface types used for resuspension tests with DNATrax and Btk were concrete and stainless
steel. The temperature for each environmental test condition was kept constant at 20 °C, and
only the RH was varied. RH values tested were 30%, 80%, and a cycle from 30% to 80% and
back to 30% - referred to hereafter as a diurnal cycle - prior to resuspension tests. Each day of
testing consisted of a single environmental condition and a single surface with the only variable
being the deposited surrogate for each day. Finally, five replicates in a day were conducted on
each surrogate for a total of 10 resuspension tests per day. A summary of the test matrix can be
found in Table 5-1 with alternating row colors indicating a separate test day.
Table 5-1. Filter Collection Resuspension Test Matrix
Material
Surface
RH/Temp
Wind Speed
Replicates
DNATrax
Steel
30%/20 °C
20 m/s (45 mph*)
5
Btk
Steel
30%/20 °C
20 m/s (45 mph)
5
DNATrax
Concrete
30%/20 °C
20 m/s (45 mph)
5
Btk
Concrete
30%/20°C
20 m/s (45 mph)
5
DNATrax
Steel
80%/20 °C
20 m/s (45 mph)
5
Btk
Steel
80%/20 °C
20 m/s (45 mph)
5
DNATrax
Concrete
80%/20 °C
20 m/s (45 mph)
5
Btk
Concrete
80%/20 °C
20 m/s (45 mph)
5
DNATrax
Steel
30%-80%-30%/20 °C
20 m/s (45 mph)
5
Btk
Steel
30%-80%-30%/20 °C
20 m/s (45 mph)
5
DNATrax
Concrete
30%-80%-30%/20 °C
20 m/s (45 mph)
5
Btk
Concrete
30%-80%-30%/20 °C
20 m/s (45 mph)
5
*miles per hour (mph)
21

-------
5.1.2 Particle Size Measurements of Resuspended Material
In general, particle size measurements of resuspended material can be difficult because low
target particle counts and background particles from porous material coupons make isolation of
the test material particles challenging. Therefore, to maximize the probability of detection, only
ideal conditions (30% RH and 20 °C) and surfaces (clean surface/low background particles)
were considered for these measurements, and assumptions were made that changes in
conditions and surfaces did not affect the size distribution of resuspended material dramatically.
The surface chosen for these tests was stainless steel, and the temperature and humidity were
set to 20 °C and 30%, respectively. Five replicates of each surrogate were tested on the same
day for a total of 10 resuspension tests. Shear wind speed and traverse speed were kept
identical to the filter collection tests. SEM analysis of resuspended material was not conducted
as not enough material could be collected during a test run for analysis.
5.2 Resuspension Testing Approach and Procedures
5.2.1 Coupon Preparation and Particle Seeding
Prior to deposition, all coupons were equilibrated according to experimental conditions to ensure
that coupon surface conditions were static during deposition. There were two
seeding/inoculation/post-inoculation conditions: 20 °C/30% RH for the 30% and diurnal cycle
resuspension tests and 20 °C/80% RH for the 80% RH resuspension tests. After equilibration,
all stainless steel coupons were wiped clean with methanol and allowed to dry fully prior to
inoculation. All concrete coupons were sprayed with compressed air at 30 psi to remove loose
dust prior to seeding/inoculation. All coupon depositions occurred the day before resuspension
tests with a total of five DNATrax coupons and five Btk coupons inoculated per day. After
equilibration, the deposition procedure for each surrogate was identical for all environmental
conditions. However, for the 80% RH deposition condition, care was taken to limit the exposure
of the surrogate materials to the atmosphere to avoid agglomeration. The process for deposition
after coupon preparation was as follows.
1. Five coupons for single surrogate seeding/inoculation were placed on individual
stainless steel bases inside an environmental chamber under resuspension
environmental conditions.
2. A 1.125" aluminum foil disc was placed on the center of each coupon as a measure
of deposition level.
3. All coupons were then covered by a settling chamber and lid.
4. One stack was taken to the deposition table.
5. The fans in the settling chamber were turned on, and an eductor system
corresponding to the surrogate being deposited was connected to the lid.
6. Approximately 2 mg of surrogate material was weighed and delivered through the
eductor system, and air flow was maintained for 5 minutes.
22

-------
7. Airflow through the eductor and the mixing fans was stopped.
8. The eductor system was removed from the settling chamber, was cleaned with water
and ethanol, and allowed to dry.
9. The eductor system was then connected to the APS, and a size distribution was
measured.
10. Steps 3-9 were repeated until all five depositions were completed and five reference
APS measurements taken.
11. Steps 1-10 were then repeated for the next surrogate seeding/inoculation.
12. After three hours of settling, all coupons were removed from their settling chambers,
the aluminum disc was collected and placed into a labeled vortex tube, and the
coupons were placed into labeled individual closed aluminum trays for storage until
resuspension testing.
13. All aluminum foil deposition reference discs were then transported to the EPA BioLab
for analysis.
For the 30% and 80% RH resuspension tests, the environment was maintained at a constant
temperature and humidity. For the diurnal cycle tests, after all coupons were transferred to the
aluminum trays, the humidity set point for the environmental chamber was set from 30% to 80%
RH, and the chamber was maintained at 80% for -14 hours. Then, the tunnel RH was set back
to 30%. Total transition time from 30%—>80% and from 80%—>30% was approximately one hour
each.
5.2.2 Filter Sampling Resuspension Procedure
Resuspension testing was conducted using the RWT described in Section 3.5. The testing
follows procedures similar to the procedures used in previous studies (EPA 2014). All five
coupons for a single surrogate inoculation were tested in succession, followed by five coupons
of the other surrogate. Scouting runs showed that no significant background material was
present inside the RWT after each run, thus no decontamination of the tunnel was conducted
between the five replicate test runs. The tunnel was decontaminated before the second set of
five coupons was tested using DNA Away™ (ThermoFisher Scientific, Waltham, MA) for
DNATrax or pH-adjusted bleach (acetic acid plus bleach) for Btk, depending on which surrogate
had been used in the last set of tests. The procedure for filter collection resuspension tests was
as follows.
1. The back of the RWT was opened, four felt folder holders were inserted into the wind
tunnel, and the section was closed.
2. A coupon in its closed aluminum tray was carefully moved from the deposition zone
to directly next to the RWT.
23

-------
3. The top of the tunnel was opened, the coupon for resuspension testing was slowly
removed from its tray and was gently inserted to minimize airflow over the surface.
The technician changed gloves and sealed the top of the tunnel.
4. The pressure regulator on the air knife was set to emit air at 45 mph.
5. The blower controlling the tunnel sampling flow was started and run for 30 seconds
prior to resuspension testing.
6. The air knife valve was opened, and the traverse electronics were engaged.
7. The air knife and traverse ran for 60 seconds. Then, the valve was closed and the
traverse electronics reversed to return the air knife to its initial condition (the traverse
position was monitored to ensure full extension at 60 seconds.)
8. The blower was allowed to run for 30 seconds to ensure all remaining resuspended
particles were pulled into the felt filters.
9. The blower was disengaged. The filter holders were removed from the tunnel and
placed into a sterile labeled Ziploc® bag.
10. The back of the tunnel was closed. The coupon was removed and placed into a
storage container to be discarded.
11. After removal of the coupon, the technician changed gloves, the tunnel was
resealed, the blower was turned on, the air knife valve was opened, and the traverse
was swept back and forth for 3 cycles to remove any additional particles from the
tunnel prior to the next resuspension test.
12. Steps 1-10 were repeated for the additional four single-surrogate coupons.
13. After all five coupons were tested, the tunnel was decontaminated and prepared for
the next set of five coupons to be tested that day.
After all resuspension testing for that day was completed, the felt filters were removed from their
holders using aseptic techniques, and all four filters from each test were placed into a labeled
Stomacher® bag (Seward Ltd., Worthing, West Sussex, UK). Each test day generated five
DNATrax and five Btk Stomacher® bags containing four felt filters from the resuspension
studies as well as two bags containing field blanks. All samples were delivered to the BioLab for
particle extraction and processing. Material from filters was extracted by adding 120 ml_ of
phosphate buffered saline containing 0.1% Triton X-100 (Sigma-Aldrich, St. Louis, MO) to the
filter bags and stomaching the bags for two minutes at 230 revolutions per minute (RPM). The
eluent was removed without squeezing the filters. Filters were not squeezed due to a slight
potential of dilution during Btk extraction. Btk spores loosened from the filter material but still
entrapped with the filter body may not move through the filter as easily as the extraction liquid
and become entrapped. Internal extraction efficiency measurements have shown -99% particle
removal, so it is assumed that the Btk material is evenly distributed throughout the extraction
24

-------
fluid and that determination of the number of CFU/ml in an aliquot is representative of the total
fluid including the portion remaining in the filter. DNATrax filter extract solutions were analyzed
by qPCR to determine the number of DNA barcodes on each filter. Btk filter extracts were spiral
plated, incubated, and counted by computer to determine the number of CFU in the filter
extraction solution. If the Btk material was not countable by computer methods, the extract
solution was spread plated or filter plated, incubated, and counted manually. The aluminum foil
deposition reference discs were extracted and analyzed as described previously in Section 4.2.
5.2.3 APS Measurement of Resuspended Material Procedure
To measure the size distribution of the resuspended material, the RWT, prior to blower
connection, was fitted with an extended sampling chamber and exit HEPA filter as show in
Figure 3-9. The sampling chamber contained sampling ports for installation of isokinetic nozzles
internally and connection ports for the APS externally. A %" sampling nozzle was installed
parallel to the tunnel flow, and an APS was connected to the external connection port. The APS
samples at 5 liters per minute generated a nozzle sampling velocity of -5.9 mph. The front of
the nozzle was approximately 3 feet (ft) from the leeward edge of the coupon and approximately
4 inches below the top of the coupon. Though the 5.9 mph sampling velocity was significantly
below the 45-mph wind shear speed directly out of the air knife, the significant distance from the
edge of the coupon and the larger space to capture particles over time increased the likelihood
of particle capture.
As stated above, particle sizing of resuspended materials requires ideal conditions and
surfaces. The particle sampling was conducted with stainless steel coupons at 30% RH. All
coupons were cleaned, equilibrated to 30% RH, and tested for background particles by APS in
the tunnel prior to particle seeding/inoculation for resuspension. The tunnel was thoroughly
cleaned and measured for a particle background prior to each resuspension test to show that
only particles coming from the seeded/inoculated surface were measured by the APS.
Resuspension tests were conducted without the tunnel blower on to maximize the time
resuspended particles were inside the chamber. The air knife was set to 45 mph and the
traverse for 60 seconds (same settings as the resuspension tests). Ten coupons were tested for
a particle background. Then, five were inoculated with Btk and five seeded with DNATrax.
Coupon deposition, settling time, and storage prior to resuspension followed the procedure for
the 30% RH resuspension studies described in Section 5.3.2. Identically to Section 5.3.2, all five
of a single surrogate type were tested before the next surrogate type. The process for a single
sampling event was as follows.
1. The seeded/inoculated coupon was placed inside the RWT, and the RWT was
closed.
2. The APS sampling was initiated for 10-second sampling intervals, and six 10-second
samples were taken for a 60-second total background sample.
3. The air knife and traverse were initiated, and the APS sampled during resuspension
for 60 seconds.
25

-------
4.	The air knife valve was then closed, and the traverse was returned to its initial
starting point. The APS continued to sample for five more minutes for a post-
resuspension sample.
5.	The blower was turned on for 30 seconds to clear the tunnel of residual aerosolized
particles and then turned off.
6.	The coupon was then removed from the tunnel.
7.	The tunnel was resealed, the blower was turned on, the air knife valve was opened,
and the traverse was swept back and forth for three cycles to remove any additional
particles from the tunnel prior to the next resuspension test.
8.	Steps 1-7 were repeated for the additional four single-surrogate coupons.
The five minutes of post-resuspension sampling were used to measure the particle size
distribution. The APS sampled in 10-second intervals. Therefore, 30 sample intervals were
summed together to determine the distribution. All five coupon runs for each surrogate were
then averaged to show the final distribution and variance. The DNATrax distribution was then
used for the calculation of the number of resuspended DNATrax particles from the filter
extraction qPCR results.
26

-------
6.0 Results and Discussion
This study sought to examine the resuspension properties and behavior of a suitable Ba
surrogate (Btk) compared to an inert maltodextrin-based surrogate, DNATrax. This comparison
was accomplished by carefully depositing known amounts of each material, using a specially
designed dry deposition eductor system for controlled deposition of DNATrax and Btk onto two
types of subway surfaces: stainless steel (nonporous) and concrete (porous). After deposition,
particles were resuspended from the surfaces under controlled conditions (temperature, relative
humidity, and wind speed), resuspended material was captured onto filters, and resuspension
fractions of both materials were calculated and compared against one another. The results of
the deposition and resuspension experiments are presented in this section.
6.1 Dry Deposition System Performance
In this section, the performance of the dry deposition eductor system is discussed. Since
deposited DNATrax particle counting using qPCR requires knowledge of the size distribution,
the APS results are presented first.
6.1.1 APS Size Distribution and SEM Analysis
Both DNATrax and Btk were tested for repeatable deposition onto coupon surfaces using the
dry deposition eductor system. Initial runs of DNATrax through the eductor system led to
significant clogging of the system and highly variable size distributions from deposition to
deposition. After initiating a cleaning procedure of water and ethanol, the variation of the
distributions was significantly reduced. Figure 6-1 shows the average of five depositions of
DNATrax through the eductor system after the implementation of the cleaning procedure. Figure
6-1A shows the relative number and volume size distributions plotted with standard deviations
as error bars. Figure 6-1B, on the right, shows the cumulative size distributions for DNATrax.
The volume distributions correspond to the distribution of mass in the system. The mass median
aerodynamic diameter (MMAD) through the deposition system was measured to be 2.28 |jm,
which is very similar to the LLNL specifications provided with the DNATrax shipment (1.83 |jm
MMAD), and the median aerodynamic particle size from the number distribution was 0.9 |jm.
These data show that the DNATrax is polydispersed with a wide range of particle sizes with
50% of the particles less than 1 |jm in aerodynamic diameter and 90% of the particles having a
diameter less than 2 |jm.
DNATrax Aerodynamic Distributions
DNATrax Cumulative Distributions
80' ift-

f
6.0 < 4
• DNATrax Nu
• DNATrax Vo
mber
ume
5.0
4.0 1
3.0
2.0
1.0 f*
... ,T\.
-ft—A
frx «— 3
ry \

100.0
90.0
_ 80.0
Z" 70.0
c
0
H 60.0
« 50.0
1 40.0
E 30.0
3
° 20.0
10.0
0.0

-f v
IZ.
/
~ DNATrax Number
DNATrax Volume
Aerodynamic Diameter (nm)
B
Aerodynamic Diameter (nm)
Figure 6-1: DNATrax Number/Volume Relative and Cumulative Distributions
27

-------
Following the APS distribution analysis, the deposited DNATrax captured onto a silicon SEM
stub was analyzed by a computer-controlled scanning electron microscope (CCSEM) capable of
automated particle imaging, sizing, and shape analysis. Figure 6-2 shows an example SEM
image of the deposited DNATrax (A) and a close-up image of a particle with a circle area
equivalent diameter of 1.8 |jm (B). The surface morphology of all observed DNATrax particles
showed dimples or indentations on the surfaces of the particles, possibly indicating that the
DNATrax initially formed hollow spheres during the spray dry process, and these hollow spheres
collapsed as the internal moisture was removed (Vehrinq 2008). The results from the CCSEM
analysis plotted with the average APS measurements can be seen in Figure 6-3. The 9000
particles were measured over an approximately 8-mm2 area. The results of the automatic scan
and visual inspection of the resultant images showed that the DNATrax deposited in mostly
singlet and doublet form with a large spectrum of singlet sizes. The shape of the distributions
from CCSEM match closely to the number size distribution measured through the APS. The
APS measures the aerodynamic diameter whereas the CCSEM gives the circle area equivalent
diameter or the diameter of a circle with the same cross sectional area as the measured
particle. For spherical and near-spherical particles, this value is very close to the actual
geometric diameter. Direct comparison between the aerodynamic diameter and the circle
equivalent diameter requires knowledge of the dynamic shape factor and the particle density
(Hinds 1999). In the case of DNATrax in Figure 6-3, the two graphs align very closely without
correction, suggesting that the density of the DNATrax particles is close to the density of water
(1 g/cm3) instead of the 1.54 g/cm3 reported for the bulk material. This observation suggests that
the particles are partially hollow shells, or that the apparent dimples on the particle surface
change the dynamic shape factor from the dynamic shape factor of a perfect sphere, or a
combination of the two. More experimentation in particle settling and density measurements
must be conducted to determine the source of this discrepancy. In addition, the images show no
sign of loss of particle integrity (shearing/fracturing) as all particles remain spheroidal below 1
micron. The CCSEM serves to show that the APS measurement is not missing larger particles
and that the dry eductor deposition system with prior material vortexing delivers the DNATrax to
the coupon surface in its natural state with few agglomerates.
28

-------
•
«•

• I
1 ir
v -
J
1 1
20 |jm
E |
Figure 6-2. Overview Scan of Deposited DNATrax (A) and 1.8-(jm Particle (B)
29

-------
DNATrax Size Comparison APS v SEM

























-•-Ae
rodynamic Diamet
M (Geometric Diai
ei








—SE
neter)
































































































0.1
100
90
80
70
60
50
40
30
20
10
0
1
Diameter (microns)
DNATrax Cumulative APS v SEM
10








«tf*f
a9

































































































•	APS (Aerodynamic Diam
•	SEM (Geometric Diamete
et
er)








»•)





























0.1
1
Diameter (microns)
10
Figure 6-3. SEM Size Measurement and APS Comparison for DNATrax
Figure 6-4 shows the size distribution results for the Btk deposition system. This figure shows a
strong sharp peak in the distribution for Btk. The mass mean particle size for the Btk was
observed to be 1.7 |jm and the mean aerodynamic size derived from the number distribution
was 1.2 |jm, with 90% of particles being below 1.6 |jm. Particles below 0.8 |jm are not viable Btk
spores and are non-colony forming, thus the particles below 0.8 |jm are not counted during
plating. The mass distribution does demonstrate that a significant portion of the mass is
associated with agglomerates—either doublets or larger—as single spores have a very tight
size distribution. However, this is not a problem if the spore agglomerates do not break apart
significantly when extracted from the filters or aluminum discs, which is the main assumption as
described in Section 3.6.2.
30

-------
Btk Aerodynamic Distributions
16
14
£ 12
c
-------
Figure 6-5. Overview Scan of Btk deposition and Single Spore
6.1.2 Deposition Spatial Distribution and Load Variation
The results from the foil-based deposition spatial distribution and repeatability tests are
presented below. Figure 6-6 shows representative heat maps of the distribution of material on
coupon surfaces for: A) DNATrax and B) Btk. The material is delivered relatively evenly across
the surface measured by the nine-foil array. However, a slight diagonal skew was occasionally
detected due to the orientation of the mixing fans in the deposition stack. In general, this slight
diagonal skew is not an issue for resuspension tests as the air knife in the RWT sweeps the
entire coupon surface. Table 6-1 shows the variation in material sampled on the nine-foil array
for each deposition as well as the variation in number of particles per milligram delivered by the
dry powder system. The variation in the nine foils represented by the coefficient of variation
(CV) was 20% for both DNATrax and Btk with the center foil (E) being a good representation of
the average foil/surface coverage. The number of particles delivered per mg for DNATrax was,
on average, 2 x 107 with 58% variation from run to run. The 58%variation is attributed to the
DNATrax having a tendancy to agglomerate in storage and clog the eductor system. However,
as discussed in Section 6.1.1, the material that exited the eductor was generally free of large
agglomerates. The Btk system delivered particles to the surface at a rate of 1.35 x 106 CFU/ mg
with a variation of 35%. Thus, it requires roughly twice as much Btk to attain the same surface
coverage as DNATrax. However, the deposition of Btk is highly repeatable due to the lack of
large particles clogging the eductor system. These results paired with the size distribution data
show that the dry powder delivery system delivers relatively repeatable size and spatial
distributions of material to the coupon surface and that a single foil In the center of the coupon
during deposition is a good reference for the total amount of material deposited. The number of
particles measured from a foil placed in the center of test coupons during inoculation was
32
-------
therefore used to derive the total number of particles for the denominator (i.e., amount
deposited) of resuspension fraction calculations.
01234567 01234567
Figure 6-6. Coupon Spatial Distribution Heat Maps: A) DNATrax, B) Btk
Table 6-1. Deposition Variation and Particle Counts per mg of Particles Deposited

DNATrax
Btk
Particles per mg
9 Foil Coefficient of
Variation (CV)
CFU per mg
9 Foil Coefficient of
Variation (CV)
Depo 1
1.02 x107
19.2%
2.19 x 10e
22.7%
Depo 2
1.02 x 107
20.2%
1.21 x 106
24.4%
Depo 3
1.31 x 107
17.8%
1.10 x 106
20.6%
Depo 4
3.47 x 107
10.4%
1.08 x 106
27.4%
Depo 5
3.33 x 107
32.9%
1.18 x 106
15.2%
Average
2.09 x 107
20.1%
1.35 x 106
22.1%
Standard
Deviation
1.20 x 107
8.1%
4.70 x 105
4.6%
Overall CV
58%

35%

1.0E-03
30% RH 80% RH Diurnal 30-80-30
Relative Humidity Condition
Figure 6-10. Concrete Resuspension Fraction
Table 6-2 shows the average resuspension fractions and the variation of all environmental
conditions, surrogates, and surfaces. Tables of the RF for each coupon measured can be found
in Appendix A. The magnitude of the difference in each surrogate resuspension fraction per
condition should be viewed as large for all tests. However, those differences for the 30% and
80% RH conditions could be due to outliers and potential variation in resuspended particle
sizes. The differences in the diurnal cycle are considered to be outside the variation due to
shifts in the particle size distribution.
Table 6-2. Average Surrogate Resuspension Fractions under Varied RH Values [Mean
(SD)]
Condition
DNATrax
Btk
SS 30% RH
6.32 (±7.42) %
2.26 (±0.91)%
SS 80% RH
1.79 (±1.88)%
0.73 (±0.47) %
SS 30 —> 80 —> 30 (diurnal cycle)
0.31 (±0.21)%
2.77 (±0.84)%
Concrete 30% RH
8.36 (±7.92) %
3.75 (±1.00)%
Concrete 80% RH
9.49 (±7.10)%
3.25 (±1.10)%
Concrete 30 —> 80 —> 30 (diurnal
cycle)
12.76 (±2.5)%
0.35 (±0.36)%
38
-------
To understand the separation in RFs for the SS diurnal cycle, Figure 6-11 shows the
temperature and RH conditions over the course of the cycle. As the figure shows, the relative
humidity for the SS experiments extended past 80% for approximately one hour before returning
to the 80% set point. This extension may have resulted in softening and an increase in the
adhesive properties of the DNATrax due to a glass transition compared to the Btk, which would
not soften considerably (Nurhadi 2016). Nurhadi et al. established glass transition temperatures
and humidities for a variety of dextrose equivalent maltodextrin particles. They showed that
particles with DEs similar to DNATrax have a glass transition at approximately 20 °C and 80%
RH. This process would be irreversible, and the DNATrax would thus resuspend less than Btk.
This is the only condition under which Btk resuspended at a higher fraction than DNATrax,
which would point to a change in the surface properties of the DNATrax or a capillary
condensation not experienced by Btk.
The diurnal humidity cycle for concrete, however, did not extend above 80% and remained just
at or below the glass transition for DNATrax (Nurhadi 2016). However, the Btk RF (0.35%)
dropped significantly below the RF of DNATrax (12.8%) due to the diurnal cycle in this case. It is
possible that the increased surface roughness of the Btk with the concrete causes a more
complex interaction than that of DNATrax and that surface dynamics of the 30% RH-equilibrated
concrete with the 80% external humidity during the cycle affect the Btk differently than the
DNATrax.

100
Diurnal Cycle Temperature and RH -
SS

Diurnal Cycle Temperature and RH ¦
¦ Concrete