EPA/600/R-14/259 | September 2014 | www.epa.gov/ord
United States
Environmental Protection
Agency
Determination of the Difference
in Reaerosolization of Spores off
Outdoor Materials
Office of Research and Development
National Homeland Security Research Center
-------
Determination of the Difference in
Reaerosolization of Spores off Outdoor Materials
National Homeland Security Research Center
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
-------
Disclaimer
The United States Environmental Protection Agency, through its Office of Research and Development's
National Homeland Security Research Center, funded and managed this investigation through contract
EP-D-10-070, WA 3-19, with Alion Science and Technology. Funding was received from the Interagency
Agreement HSHQPM-12-X-00118 P00001 between the Environmental Protection Agency and the
Department of Homeland Security. This report has been peer and administratively reviewed and has
been approved for publication as an Environmental Protection Agency document. It does not necessarily
reflect the views of the Environmental Protection Agency. No official endorsement should be inferred.
This report includes photographs of commercially available products. The photographs are included for
purposes of illustration only and are not intended to imply that the Environmental Protection Agency
approves or endorses the product or its manufacturer. The Environmental Protection Agency does not
endorse the purchase or sale of any commercial products or services.
Questions concerning this document or its application should be addressed to:
Russell W. Wiener, Ph.D.
Decontamination and Consequence Management Division
National Homeland Security Research Center
U.S. Environmental Protection Agency (MD-D205-03)
Office of Research and Development
109 T.W. Alexander Drive
Research Triangle Park, NC 27709
Phone:919-541-1910
Fax:919-541-0496
E-mail: wiener.russell@epa.gov
-------
Acknowledgments
The authors would like to acknowledge the funding support of the Department of Homeland Security
Science and Technology Directorate through Interagency Agreement HSHQPM-12-X-00118 P00001.
Additionally, we would like to acknowledge the support of Alion Science and Technology, funded under
EPA contract EP-D-10-070. In addition the authors would like to acknowledge the contributions of
ARCADIS U.S., Inc. for microbiological support and Booz Allen Hamilton Inc. for program support. The
quality assurance review by Joan Bursey is also appreciated.
The contributions of the interagency Scientific Program on Reaerosolization and Exposure (SPORE)
team members listed below are also acknowledged:
DHS:
Donald Bansleben
Matthew Moe
Department of Defense:
K. Wing Tsang
Jeffrey Hogan
Angelo Madonna
Nicholas Hogan
HHS:
John Koerner
Angela Weber
EPA:
Shawn Ryan
Russell Wiener
Marshall Gray
Worth Calfee
Sangdon Lee
Sara Taft
-------
Table of Contents
Disclaimer ii
Acknowledgments iii
List of Figures v
List of Tables vi
List of Acronyms and Abbreviations vii
Executive Summary ix
1 Introduction 1
2 Experimental Materials and Methods 3
2.1 Reaerosolization Wind Tunnel 3
2.2 Air Jet 6
2.3 Deposition Chambers 9
2.3.1 Wet Deposition Chamber 9
2.3.2 Dry Deposition Chamber 10
2.4 Test Surrogates 12
2.4.1 Bacillus thuringiensis var. kurstaki (Btk) 12
2.4.2 Bacillus atrophaeus subspecies globigii (Bg) 13
2.5 Test Surrogate Preparations 13
2.5.1 Wet Preparation 13
2.5.2 Dry Preparation 14
2.6 Test Surface Characterization 14
2.6.1 Hygroscopicity 15
2.6.2 Surface Tension 15
2.6.3 Surface Roughness 16
2.6.4 Surface Area 17
2.7 Test Protocol 18
2.8 Nondimensional Analysis 20
2.9 Modeling 20
3 Results and Discussion 22
3.1 Materials Characterization 22
3.1.1 Hygroscopicity 22
3.1.2 Surface Tension 23
3.1.3 Surface Roughness 24
3.1.4 Surface Area 25
3.2 Scouting Tests 26
3.3 Reaerosolization Experiments 30
3.4 Propagation of Error Analysis 32
3.5 Modeling 34
4 Quality Assurance 38
4.1 Equipment Calibration 38
4.2 Data Quality Objectives 38
IV
-------
4.3 QA/QC Checks 39
4.4 Acceptance Criteria 40
5 Summary 43
6 References 44
List of Figures
Figure 2-1: T/RH controlled environmental test chamber 3
Figure 2-2: EPA RWT positioned inside the T/RH chamber 4
Figure 2-3: Open front section of EPA RWT showing a coupon, lower part of the internal
flow transition unit, and the slotted nozzle air jet attached to the linear traversing device 4
Figure 2-4: Slotted nozzle air jet driven by the linear traverse actuator 4
Figure 2-5: Coupon and flow transition unit 4
Figure 2-6: Slotted nozzle air jet above the coupon at the end of its forward traverse and
the airspeed monitoring device (anemometer) 5
Figure 2-7: Last open door showing filter holders 5
Figure 2-8: Four filter holders designed to be serviced using heavy-duty gloves 5
Figure 2-9: Vertical velocity profiles in RWT #1 (EPA) and RWT #2 (DPG) 6
Figure 2-10: Average air velocity from the slotted nozzle as a function of nozzle pressure 7
Figure 2-11: Wet deposition chamber 9
Figure 2-12: Top view of chamber 9
Figure 2-13: Deposition chamber coupon platform 9
Figure 2-14: Dry deposition system showing two chambers 11
Figure 2-15: Inside dry deposition chamber, looking down from top 11
Figure 2-16: SEM image of bar-coded Btk spores deposited with wet deposition chamber 12
Figure 2-17: SEM image of Bg spores deposited with wet deposition chamber 13
Figure 2-18: Flow chart describing the test protocol for deposition experiments in the RWT 19
Figure 3-1: Scouting test average F for wet-deposited Btk by surface type and jet velocity 28
Figure 3-2: F for wet-deposited Btk calculated from scouting tests with varying jet velocity for
smooth surfaces 29
Figure 3-3: F for wet-deposited Btk calculated from scouting tests with varying jet velocity for
rough surfaces 29
Figure 3-4: F for wet-deposited Btk calculated from scouting tests with varying jet velocity for
very rough surfaces 30
Figure 3-5: Fit of a model to predict reaerosolization of wet particles from various urban surfaces
categorized by laboratory and test particle 35
Figure 3-6: Fit of a model to predict reaerosolization of wet particles from various urban surfaces
categorized by surface material 35
Figure 3-7: Fit of a model to predict reaerosolization of dry particles from various urban surfaces
categorized by laboratory and test particle 37
Figure 3-8: Fit of a model to predict reaerosolization of dry particles from various urban surfaces
categorized by surface material 37
-------
List of Tables
Table 2-1. Air Jet Exit Velocity for Each Coupon Material, Deposition Type, and Velocity Category 8
Table 2-2. Surface Roughness and Roughness Length Scale, z0, Values 9
Table 2-3. Materials/Surfaces Used for Testing 14
Table 2-4. Hamaker Constants 16
Table 3-1. Summary of Surface Measurements Used for Particle Reaerosolization Experiments and
Modeling 22
Table 3-2. Hygroscopicity Measurements (% Mass Loss) 23
Table 3-3. Goniometry Measurements Given as the Cosine of the Contact Angle (radians) 23
Table 3-4. Microscopic Surface Roughness (urn) Determined from SEM Images 24
Table 3-5. Microscopic Horizontal Surface Roughness (Peak-to-Peak Space, urn) Determined from
SEM Images 24
Table 3-6. Macroscopic Surface Roughness (unitless) Determined from Optical Images 25
Table 3-7. Microscale Surface Area Measurements (urn2) from SEM Images 26
Table 3-8. Macroscale Surface Area Measurements (urn2) from Optical Images 26
Table 3-9. Number of Scouting Tests Completed by Coupon Type 27
Table 3-10. Summary of Findings from Statistical Analysis of Reaerosolization Data 31
Table 3-11. Overall Mean F Values Calculated from Reaerosolization Tests 32
Table 4-1. Data Quality Objectives for Experimental Conditions 38
Table 4-2. QA/QC Checks 40
Table 4-3. Spore Acceptance Criteria 40
Table 4-4. Deposition Acceptance Criteria 41
Table 4-5. Reaerosolization Acceptance Criteria 42
Table 4-6. Acceptable Variation among Replicate Experiments 42
VI
-------
List of Acronyms and Abbreviations
A Hamaker constant
A1 microscale surface area
A2 macroscale surface area
ATF Aerosol Test Facility
atm atmosphere(s)
Ba Bacillus anthracis
Ba-Ames Bacillus anthracis (Ames strain)
BET Brunauer-Emmett-Teller algorithm
Bg Bacillus atrophaeus subspecies globigii
BSL biosafety level
Btk Bacillus thuringiensis var. kurstaki
CPU colony-forming unit(s)
cm centimeter(s)
CV coefficient of variation
cfp particle diameter
DHHS United States Department of Health and Human Services
DHS United States Department of Homeland Security
Dl deionized
DOD United States Department of Defense
DPG Dugway Proving Ground
DQO data quality objective
£ surface energy
ECBC Edgewood Chemical Biological Center
EPA United States Environmental Protection Agency
F fraction reaerosolized
ft foot/feet
g gram(s)
HEPA high-efficiency particulate air
in inch(es)
J Joule
L liter(s)
m meter(s)
mg milligram(s)
min minute(s)
ml milliliter(s)
mm millimeter(s)
MOP miscellaneous operating procedure
mph mile(s) per hour
VII
-------
|j|_ microliter(s)
|jm micrometer(s)
NR jet traverse speed
NHSRC National Homeland Security Research Center
NIH National Institutes of Health
NIST National Institute of Standards and Technology
N Newton
nm nanometer(s)
PBS phosphate buffered saline
PBST phosphate buffered saline with 0.05% Tween® 20
PCR polymerase chain reaction
QA/QC quality assurance/quality control
QAPP quality assurance project plan
R-, microscale roughness height
R2 macroscale roughness height
Ra measured surface roughness
RH relative humidity
RTP Research Triangle Park
RWT reaerosolization wind tunnel
s second(s)
S1 microscale roughness spacing between asperities
SD standard deviation
SEM scanning electron microscope/microscopy
SOP standard operating procedure
SPORE Scientific Program on Reaerosolization and Exposure
SS stainless steel
T temperature
TGA thermal gravimetric analyzer
U0 exit velocity of jet
T shear stress
z0 roughness length scale
pp particle density
VIM
-------
Executive Summary
This report presents the initial study conducted as part of the Scientific Program on Reaerosolization and
Exposure (SPORE), which was established by the United States Department of Defense, Department of
Health and Human Services, Department of Homeland Security, and Environmental Protection Agency
(EPA) to coordinate related missions in biodefense response and recovery regarding the possibility of a
Bacillus anthracis (Ba) spore release in the United States. Two questions of significant public health
concern are addressed by the research in this report. First, can Ba spores be reaerosolized in the
ambient environment due to natural or anthropogenic causes? The research conclusively found the
answer to be yes, Ba spores can be reaerosolized. The second question is, can a benign spore be used
to represent Ba when conducting laboratory and field experimentation relevant to the deposition,
reaerosolization, and transport of Ba? Again, the experiments indicated yes as one of the tested
surrogate spores yielded equivalent results to Ba.
A large portion of the research effort was in the design, construction, and validation of equipment and the
development and validation of methods. In order to determine the efficacy of surrogate particles in a
timely and cost-effective manner, tests were performed at both the EPA facilities in Durham, NC, and at
the Department of Defense (DOD) facilities at Dugway Proving Ground in Dugway (DPG), UT. Testing
was conducted at both laboratories with two nearly identical small wind tunnels placed in environmental
enclosures. The tests at DPG included live Ba (Ames strain) and surrogate spores, while the tests at EPA
were limited to surrogate spores. The validation and performance tests conducted with surrogate spores
demonstrated that the resuspension test results using the same test conditions (spore, temperature,
relative humidity, etc.) showed no evidence of statistical differences between the two laboratories. The
performance equivalency of the two laboratories allowed for direct comparison of the DPG Ba-Ames tests
and the EPA surrogate tests.
The test surrogates studied included Bacillus thuringiensis var. kurstaki (Btk) spores and Bacillus
atrophaeus subspecies globigii (Bg) spores. The Btk used in the experiments was DMA-labeled
(bar coded) and provided by DOD. The experimental data showed that there were no statistically
significant differences in spore reaerosolization between Btk and Ba-Ames for either wet or dry
deposition. Results from resuspension tests using Bg were 79 % lower than the average for Btk and
Ba-Ames. Wet-deposited Btk and Ba-Ames spore reaerosolization results were 81 % lower than dry-
deposited Btk and Ba-Ames spores. Other noteworthy findings of the study were concrete consistently
yielded greater reaerosolization than asphalt and glass, the roughness category within each material type
did not have a significant effect on reaerosolization, and the high jet velocity consistently reaerosolized
significantly more particles than the low and medium jet velocities.
IX
-------
1 Introduction
This project report describes the research effort to characterize the reaerosolization of Bacillus spores, in
particular biological anthrax simulants Bacillus thuringiensis var. kurstaki (Btk) and atrophaeus
subspecies globigii (formerly Bacillus globigii [Bg]), from outdoor surfaces under different environmental
conditions and compare the data with data from identical experiments carried out using B. anthracis
(Ames strain), hereafter referred to as Ba-Ames.
The purpose of this research effort was twofold: Our first and primary focus was to determine the
suitability of the biological simulants Btk and Bg for Ba-Ames. Our second aim was to provide a
quantitative understanding of the reaerosolization of anthrax spores from environmentally available
surfaces.
Over the past decade, the intentional release of the "Category A" biological agent Bacillus anthracis (Ba)
has been deemed a serious public health risk. Such a release has the potential to produce mass
casualties as Ba spores are highly persistent in the environment, thus creating challenges for
decontamination and restoration of infrastructure. Over the past 60 years, numerous studies have
identified the potential for Ba to reaerosolize (become a particulate suspended in air after initial settling).
Unfortunately, there has been a dearth of quantitative information related to the prediction of public health
impacts from reaerosolized Ba spores. Without such information, prediction of the reaerosolization hazard
accompanying an anthrax attack in an outdoor urban environment is problematic. The mechanisms
governing reaerosolization are complex, and developing a quantitative database is scientifically
challenging, time consuming, and expensive.
The United States Department of Defense (DOD), Department of Health and Human Services (DHHS),
Department of Homeland Security (DHS), and Environmental Protection Agency (EPA) execute related
missions in biodefense response and recovery. There are mutual interests in understanding more about
Ba reaerosolization and any associated public health risks. In turn, these organizations are partnering
through the Scientific Program on Reaerosolization and Exposure (SPORE). The primary objective of
SPORE is to develop an understanding of reaerosolization to make informed decisions regarding
response and mitigation activities that will reduce the risks associated with an outdoor Ba release.
The overall SPORE effort will provide guidance to senior leadership, public health officials, and response
personnel so that they can be prepared to respond to and mitigate economic, environmental, and human
health impacts resulting from reaerosolization after an anthrax incident. In such an event, B. anthracis
spores released into an outdoor urban area would disperse and settle on exposed surfaces. The physical
and chemical requirements for decontamination would depend on the initial dispersion and the later
reaerosolization of the biological material as well as the types of surfaces affected, such as concrete,
brick, asphalt, soil, metal piping, plants, plastics, and cloth—the typical materials that make up the
infrastructure of a complex urban environment.
This study investigated the reaerosolization of Btk and Bg surrogate spores in EPA's National Homeland
Security Research Center (NHSRC) Aerosol Test Facility (ATF) located in Research Triangle Park (RTP),
NC. Results from tests conducted at EPA were compared with results from tests conducted at Dugway
Proving Ground (DPG) in Utah. The testing was conducted by a project team consisting of Alion Science
and Technology (Durham, NC), RTI International (RTP, NC), ARCADIS U.S. (Durham, NC), and DPG,
with project oversight from EPA, DHS, and DPG principal investigators and the Interagency SPORE
Workgroup. The work at DPG was carried out in a manner identical to the work conducted in EPA's ATF
but using Ba-Ames.
1
-------
Two small reaerosolization wind tunnels (RWTs) were designed and constructed specifically for
conducting spore reaerosolization studies. The first RWT was used at EPA to conduct reaerosolization
experiments using two surrogate bacterial spore species as the test subjects. The second RWT was
installed in the biocontainment chamber at DPG and used to conduct reaerosolization experiments using
Ba-Ames. The end point for the experimentation was to compare two sets of data, one acquired from
EPA using surrogates and the second from DPG using Ba-Ames. Validation tests were completed at
DPG using the same study protocols and one surrogate (Btk) used in the EPA facility. Reaerosolization
results for Btk from EPA and DPG were compared statistically to verify that the data produced by both
reaerosolization wind tunnels describe the same population. Once the data were demonstrated to be
equivalent, DPG conducted experiments using Ba-Ames. The two data sets were then evaluated
statistically and the behaviors of the Ba-Ames and surrogates compared quantitatively to determine the
efficacy of the surrogate in representing Ba-Ames in terms of environmental reaerosolization. The
reaerosolization data were also used to develop predictive models for reaerosolization of wet- and dry-
deposited spores.
The wind tunnel experiments were concerned primarily with the combination of two processes, namely,
detachment of spores from a surface followed by their reaerosolization in the tunnel airflow. Numerous
experimental variables affect reaerosolization of particles. Particle variables include size, morphology,
chemistry, hygroscopicity, electrostatic properties, and others. Important surface characteristics also
include morphology, chemistry, hygroscopicity, and electrostatic properties, as well as roughness and
asperity. In addition, fluid variables impact reaerosolization; these key parameters include velocity,
turbulence, boundary layer effects, and humidity.
The experimental variables that were selected for this study are particulate type (Btk, Bg, and Ba-Ames),
particle deposition type (wet and dry), jet velocity, surface type (asphalt, glass, and concrete), and
roughness level of each surface type (smooth, rough, and very rough). Although innumerable potential
surrogates exist, the three chosen for these experiments were selected based on programmatic interests,
as well as safety, size similarity, and other key particle properties. At this point, the relative importance of
various spore properties, such as the type of endospore and degree of hygroscopicity, has not been
established. The properties of the test surfaces were analyzed quantitatively. Experimental variables were
prioritized per input from the SPORE Workgroup. It was important that both EPA and DPG operate using
identical protocols, so any parameter that could not be controlled adequately at both facilities was
deemed unsuitable for evaluation in this set of experiments.
-------
2 Experimental Materials and Methods
The laboratory reaerosolization studies using surrogates were conducted at EPA's ATF in RTP, NC.
Processing and analysis of microbiological samples was conducted in EPA's microbiology laboratory in
RTP. Reaerosolization studies using Ba-Ames were conducted in DPG's biosafety level (BSL) 3
laboratories at Dugway Proving Ground, UT. The key materials and pieces of equipment used for the
project are described here. These materials and equipment include a small-scale reaerosolization wind
tunnel and deposition chambers for wet and dry deposition of particulate materials. Two identical versions
of each piece of equipment were designed and constructed for the project. One set was used for tests
with surrogates in EPA's ATF, and the other set was installed at DPG in the BSL 3 facility and used for
tests with Ba-Ames. Material characterization testing, dimensional analysis, and modeling were
conducted by RTI International in RTP.
2.1 Reaerosolization Wind Tunnel
The RWT at EPA is contained in a temperature (T) and relative humidity (RH) controlled environmental
test chamber (Figure 2-1) within a BSL 1 laboratory. The BSL 1 facility is a standard laboratory module in
EPA's RTP facility.
Figure 2-1: T/RH controlled environmental test chamber.
The RWT (Figures 2-2 through 2-8) is 6 feet (ft) long and has a 9-inch (in) x 9-in cross section. The RWT
is open-ended by design and can thus take advantage of the T/RH environmental settings available
inside the chamber. It operates on the principle of total collection of reaerosolized material from the test
surface on four polyester felt filters. The RWT is fitted with two high-efficiency particulate air (HEPA) filters
-------
to provide clean air and prevent contamination of the chamber and the blower. Its walls are laminated
with aluminum sheeting used in building construction. The top of the RWT is divided into independent
doors that provide access to various areas of the tunnel. One door allows for installing and retrieving the
test material (coupon), and another door provides access for installing and retrieving test filters. The RWT
that was constructed for use in the BSL 3 facility at DPG is identical to the EPA version except that the
doors are located on the side of the RWT to provide easier access through glove ports.
Figure 2-2: EPA RWT positioned inside the T/RH
chamber. The HEPA filter can be seen at the entry
end and the blower at the far end.
Figure 2-3: Open front section of EPA RWT
showing a coupon, lower part of the internal
flow transition unit, and the slotted nozzle air
jet attached to the linear traversing device.
Figure 2-4: Slotted nozzle air jet driven by the
linear traverse actuator.
Figure 2-5: Coupon and flow transition unit.
-------
Figure 2-6: Slotted nozzle air jet above the
coupon at the end of its forward traverse and the
airspeed monitoring device (anemometer).
Figure 2-7: Last open door showing filter
holders.
Figure 2-8: Four filter holders designed to be serviced using heavy-duty gloves.
The reaerosolization force applied to the deposited spores was generated by an air jet that traversed over
the surface of the test coupon. The air jet was produced by a slotted nozzle, shown in Figure 2-4, that
was connected to a compressed air supply. The air jet is described in section 2.2.
Downwind of the air jet and test coupon in the RWT is the flow transition unit, shown in Figure 2-5. This
unit was designed to minimize loss of particles and at the same time accommodate four filter holders
(Figure 2-8) with round polyester felt filters. The biological extraction requirements imposed a need to
have filters no larger than 10 centimeters (cm) in diameter. Therefore, multiple filters were installed to
obtain appropriate air velocity and directionality in the tunnel.
Vertical profiles of velocity were measured in both RWTs to verify that they had the same overall flow
characteristics. Results are presented in Figure 2-9. For these measurements, a clean, smooth concrete
coupon was placed in the RWT, the air jet was positioned in the center of the coupon, and a hot-wire
anemometer (series 471 digital thermo-anemometer, Dwyer Instruments, Inc., Michigan City, IN, USA)
was used to measure the velocity 2 cm downwind of the downwind edge of the coupon. Velocity was
measured at eight heights using five different nozzle pressures ranging from 0 to 2 atmospheres (atm).
Statistical analyses were conducted to compare the velocity measurements of the two different RWTs. In
-------
addition to examining whether the measured velocities were significantly different statistically, the
magnitude of the differences was also estimated. The analyses performed here answered the question of
whether velocities in the two RWTs differed from an overall perspective. The EPA RWT (#1) generally
yielded lower measured velocities than the DPG RWT (#2). This difference was established with a
Wilcoxon signed rank test (Hollander and Wolfe, 1973), which gave a p-value of 0.0021. Of the 45 height-
pressure combinations, RWT#1 was lower on 34 and higher on 10, and one combination had equal
velocities. The magnitude of the RWT #1 minus RWT #2 difference was estimated to be -0.17 meters
(m)/second (s), determined from the Hodges-Lehmann estimator based on Walsh averages (Hollander
and Wolfe, 1973). To summarize, RWT #1 consistently yielded lower velocity measurements with the
overall difference estimated as -0.17 m/s. This difference is acceptable at 7 % of the average measured
velocity (2.3 m/s).
-RWTttl P^O.Oatm —I
RWT#2 P-O.Oatm -I
10
•RWTHI P=0.25atm —
RWT#2P-0.25atm -
k—RWT#lP^0.5atm
— RWT#2P-0.5atm
•RWTS1 P-l.Oatm
RWT #2 P=1.0atm
•RWTW1 P=2.0atm
RWT #2 P=2.0atm
& 7
E
o
•t 4
.2 3
10 15
Wind velocity (mph)
Figure 2-9: Vertical velocity profiles in RWT #1 (EPA) and RWT #2 (DPG).
2.2 Air Jet
The air jet applied the reaerosolization force evenly across the width of the test coupon (Figure 2-5). The
actuator was designed to move the air jet 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 changed as desired, but were
held constant for all experiments.
-------
Scouting trials were conducted to determine appropriate jet velocities to reaerosolize wet-deposited
spores. The trials illustrated that wet deposition of spores results in strong surface attachment, requiring
high jet velocities to dislodge significant quantities of spores. A small number of additional scouting trials
were conducted for dry-deposited spores to determine the range of appropriate jet velocities. As
expected, these jet velocities were significantly lower than those used for the wet deposition experiments.
The flow from the air jet was quantified in terms of the average air velocity exiting the slotted nozzle,
calculated by dividing the total volumetric flow rate of air through the nozzle at each pressure setting by
the cross-sectional area of the nozzle opening (Figure 2-10).
300
250
-C
Q.
.§ 200
u
_O
OJ
> 150
o
c
& 100
50
0.5
1.5 2 2.5
Nozzle pressure (atm)
3.5
Figure 2-10: Average air velocity from the slotted nozzle as a function of nozzle pressure.
A literature review was conducted to determine a suitable expression of the shear stress exerted by the
air jet at the coupon surface. Scientific studies have established that three flow zones exist when a jet
obliquely impinges onto a surface: a free jet zone, an impingement zone, and a wall jet zone (Chin and
Agarwal, 1991; Akansu et al., 2008; Beitelmal and Saad, 2000; Abramovich, 1963; Glauert, 1956; Myers
et al., 1961; Rajaratnam, 1967; Schwartz and Cosart, 1961). For this experiment, we are primarily
concerned with the wall jet zone, as it describes the flow of air from the jet over the surface of the coupon.
-------
Loureiro and Freire (2009) described turbulent jet impingement onto a rough surface by
U, K ^Z0
and
C = 0.8432-^2>L-5.9615,
u.
where u is the velocity at height z above the surface, u, is the friction velocity, K is the von Karman
constant (taken to be 0.4), z0 is the roughness length scale, and Umax is the maximum velocity (Loureiro
and Freire, 2009).
Hogg et al. (1997) derived a relationship to model the boundary shear stress, T, at a rough surface as
where C5 is a constant, p is the density of air, U0 is the exit velocity of the jet, x is the horizontal distance,
b0 is the width of the nozzle opening, and m and n are constants. This equation was used to estimate the
shear stress for each combination of surface material, roughness level, and jet velocity that was tested.
This shear stress can be translated to activities in the environment that could cause reaerosolization of
spores (wind, foot traffic, vehicle traffic, etc.).
The parameter values used to calculate rwere C5 = 1, x= 10 millimeters (mm), b0 = 0.5 mm, m = 0.37,
and n = 0.84. U0 values are given in Table 2-1, calculated as the volumetric flow rate through the air jet
divided by the cross-sectional area of the nozzle opening. Values for z0 are given in Table 2-2, calculated
as the measured surface roughness for the appropriate material multiplied by 0.01. Table 2-2 contains the
unitless, relative surface roughness measurements, which were taken from optical, millimeter-scale
measurements (described in section 2.6.3) and converted to roughness values in centimeters. The
conversion was accomplished by multiplying each relative roughness value by the ratio of the physical
surface roughness measurement for very rough glass (0.06 cm) to the relative roughness of very rough
glass (4.13 x 104). To establish the physical surface roughness of the very rough glass, a micrometer
distance gauge was used to record the highest and lowest points on the surface, and then the average
distance between the highest and lowest points was calculated and halved.
Table 2-1. Air Jet Exit Velocity for Each Coupon Material, Deposition Type, and Velocity Category
Deposition
Type
Wet
Dry
Air Jet
Velocity
Low
Medium
High
Low
Medium
High
Air Jet Exit Velocity, U0 (mis)
Asphalt
2.59
46.0
101
4.87
14.4
23.5
Concrete
30.0
56.8
101
4.87
14.4
23.5
Glass
46.0
72.9
113
23.5
44.4
79.1
-------
Table 2-2. Surface Roughness and Roughness Length Scale, z0, Values
Surface Material
Asphalt, smooth
Asphalt, rough
Asphalt, very rough
Concrete, smooth
Concrete, rough
Concrete, very rough
Glass, smooth
Glass, rough
Glass, very rough
Relative
Roughness
3.28E+4
7.93E+4
9.62E+4
9.51 E+4
1.03E+5
1.17E+5
2.99E+2
4.12E+3
4.13E+4
Roughness
(cm)
4.77E-2
1.15E-1
1.40E-1
1.38E-1
1.49E-1
1.71E-1
4.35E-4
5.98E-3
6.00E-2
Zo
(cm)
4.77E-4
1.15E-3
1.40E-3
1.38E-3
1.49E-3
1.71E-3
4.35E-6
5.98E-5
6.00E-4
2.3 Deposition Chambers
Two chambers were constructed and used for deposition of wet and dry spore preparations. Chambers
identical to the wet and dry chambers used in EPA's ATF were constructed and used at the DPG facility.
2.3.1 Wet Deposition Chamber
The wet deposition chamber consists of several key components: a vessel, an ultrasonic nozzle, and two
small mixing fans (Figures 2-11 and 2-12). It was designed to fit over the coupon for aerosol generation
and deposition (Figure 2-13). Two identical chambers were fabricated. The first chamber was used at
EPA, and the second was used at DPG.
\
Figure 2-11: Wet deposition
chamber.
Figure 2-12: Top view of
chamber.
Figure 2-13: Deposition
chamber coupon platform.
-------
Several design features were included in the wet deposition chamber to achieve consistency in surface
loading. An ultrasonic nozzle (model Q060-2-26-17-303-030, Sono-tek Corp., Milton, NY, USA) was used
to atomize the spore solution in the deposition chamber. A microbore was fitted inside the ultrasonic
nozzle to minimize the amount of "dead" liquid trapped inside the nozzle that could drip onto the coupon
at the end of the atomization. The microbore inner diameter was sized to produce a median droplet size
of 31 micrometers (urn). The needle was connected directly to the dispensing syringe pump via
hypodermic (20 gauge) tubing, providing a well-controlled and steady dispensing rate. As shown in Figure
2-12, the two small mixing fans (model DF124010BM-3G, Dynatron Corp., Fremont, CA, USA) were
positioned close to the coupon surface at the edges of the coupon at opposite walls of the chamber. This
arrangement produced swirling air that rose along the walls, descended in the center, and then dispersed
along the coupon, forming a droplet-laden vortex that swirled downward and ultimately produced a
uniform coating on the coupon.
Deposition efficiency was measured for the wet deposition chamber by depositing 0.2 ml of Btk spore
solution onto an array of six stainless steel (SS) coupons that completely covered the test coupon area.
For each deposition test, a reference sample was collected by dispensing 0.2 ml of Btk spore solution
directly into a conical tube containing 10 ml of phosphate buffered saline with 0.05 % Tween® 20 (PBST).
Spores were extracted from the SS coupons according to miscellaneous operating procedure (MOP)
6601 (ARCADIS U.S., 2013a). The reference samples and samples extracted from the SS coupons were
plated and incubated, and colony-forming units (CFU) were enumerated according to MOP 6535a
(ARCADIS U.S., 2009). The average measured deposition efficiency for the wet deposition chamber from
eight separate tests with Btk was 48 %, calculated as the ratio of CFU recovered from the SS coupons to
CFU contained in the reference sample. The repeatability of deposition is estimated by the coefficient of
variation (CV) of reference samples dispensed from the same spore solution. This analysis of the
reference sample data from the deposition efficiency tests showed that the CV was 10 %.
2.3.2 Dry Deposition Chamber
Dry deposition of test particles was accomplished by suspending them in a 90 % ethanol solution and
spraying the suspension in a spray-dry deposition chamber. This procedure was very similar to the
procedure used for the wet deposition chamber, but the dry chamber was taller to allow for the particles to
dry before settling on the coupon surface. Multiple dry chambers were constructed so that dry deposition
could be completed simultaneously on multiple coupons. Two of the chambers are shown in Figure 2-14.
A total of nine chambers were constructed: seven were used at EPA and two at DPG. The settling time
required for dry deposition was a minimum of three hours. The ultrasonic nozzle from the wet deposition
chamber was used for spray-drying. Each chamber was a rectangular aluminum duct 8 in x 8 in x 18 in
with two 1.5-in square mixing fans mounted inside, 4.25 in from the chamber top to the center of the fan
(Figure 2-15). A lid was constructed for each chamber with a recessed hole in the center to accommodate
the ultrasonic nozzle and a cover to slide over the hole when the nozzle was not on the chamber. The lids
were cut from 0.5-in acrylic and lined with aluminum foil. Grounding wires were attached to the foil lining
of the lid and to a post on the side of the chamber so that the entire chamber could be grounded. Mixing
fans were turned on only while the ultrasonic nozzle was dispensing spore solution. After dispensing the
spore solution, the mixing fans were turned off. After two minutes (min), the ultrasonic nozzle was moved
to another identical chamber with the next test coupon, and the hole in the chamber lid was covered. This
procedure was repeated for seven coupons at a time, and the settling chambers were left undisturbed for
a minimum of three hours to allow gravitational settling of spores onto the coupon.
10
-------
Figure 2-14: Dry deposition system showing two chambers.
Figure 2-15: Inside dry deposition chamber, looking down from top.
Deposition efficiency was evaluated in the same manner as described above for the wet deposition
chamber. The only differences in the process were that the SS coupons were placed in closure bags for
the Stomacher 400 Circulator (Seward Laboratory Systems Inc., Davie, FL, USA) with 80 ml of PBST to
extract spores, and the corners of the SS coupons were rounded to avoid sharp corners puncturing the
Stomacher bags. The average measured deposition efficiency for the dry deposition chamber from 12
tests with Bf/
-------
2.4 Test Surrogates
Several species of Bacillus spores are commonly used as surrogates for B. anthracis in laboratories and
field studies, including B. atrophaeus, B. thuringiensis, B. cereus, and 8. subtilis, among others
(Greenberg et al., 2010). This project aims to compare reaerosolization of two of these types of bacterial
spores to reaerosolization of Ba-Ames.
2.4.1 Bacillus thuringiensis var. kurstaki (Btk)
The Ba-Ames genotype and phenotype are similar to B. thuringiensis (EPA, 2012). Btk is a gram-positive
bacterium commonly found in soil. The endotoxin protein produced during sporulation is commonly used
as a pesticide. The average hydrated spore size is 0.8 x 1.4 urn, which is the same as the average
hydrated spore size for B. anthracis (Carrera et al., 2007). Bar-coded Btk is a genetically modified strain
developed by Edgewood Chemical Biological Center (ECBC). The genetic modification alters the DMA
sequence so that polymerase chain reaction (PCR) analysis clearly distinguishes these spores from the
common naturally occurring strain. The applicability of bar-coded Btk as a surrogate for Ba-Ames is under
evaluation in this research.
The bar-coded Btk cells used in this study were cultured by 10-liter (L) batch fermentation. Following
sporulation, the spores were concentrated into a wet pellet, washed three times, and lyophilized. The
lyophilized spores were a dry but very clumpy cake rather than a loose dry powder. The lyophilized
spores were delivered to EPA and mixed with sterile deionized (Dl) water to make a stock solution that
was diluted with PBST to the desired titer of 108 spores/milliliter (ml). Figure 2-16 shows a scanning
electron microscopy (SEM) image of bar-coded Btk spores deposited onto aluminum foil using the wet
deposition chamber. The Btk spores were deposited as 74 % singlets and 21 % doublets with very few
large agglomerates.
Figure 2-16: SEM image of bar-coded Btk spores deposited with wet deposition chamber.
12
-------
2.4.2 Bacillus atrophaeus subspecies globigii (Bg)
Bacillus atrophaeus subspecies globigii is a gram-positive rod capable of producing endospores. The
endospores are elliptically shaped with dimensions of 0.7-0.8 x 1-1.5 urn and a mean aerodynamic
diameter of approximately 1.1 urn (Carrera et al., 2007). Bg has been used as a surrogate for B. anthracis
in more than 40 studies (Greenberg et al., 2010). Bg cells were fermented, harvested, and lyophilized by
DPG in the same manner as the bar-coded Btk. Figure 2-17 shows a SEM image of Bg spores deposited
using the wet deposition chamber onto aluminum foil.
Figure 2-17: SEM image of Bg spores deposited with wet deposition chamber.
2.5 Test Surrogate Preparations
2.5.1 Wet Preparation
The same spore preparation method was used for Btk, Bg, and Ba-Ames to be deposited in the wet
deposition chamber. Spores, received from DPG as a lyophilized paste, were prepared for wet deposition
by suspending the paste in sterile Dl water to achieve a titer on the order of 108 spores/mL. The spore
suspension was sonicated for one minute to break up agglomerates and agitated using a vortex mixer
(BV1000, Benchmark Research Products, Roebling, NJ, USA) to homogenize the suspension.
Concentration of the spore suspension was confirmed using the procedure defined in MOP 6535a
(ARCADIS U.S., 2009).
13
-------
2.5.2 Dry Preparation
The same spore preparation method was used for Btk, Bg, and Ba-Ames to be deposited in the dry
chamber. Spores were prepared for dry deposition in the dry chamber by suspending the paste in 90 %
ethanol to achieve a titer on the order of 108 spores/mL. A small amount of sterile Dl water was first
added to the lyophilized paste, and then the suspension was sonicated, vortexed, and left overnight to
hydrate the spores. A solution of 0.07 % Tween® 20 in sterile Dl water was added to the hydrated spores
to create a spore suspension with an overall Tween® 20 concentration of 0.05 % and a titer on the order
of 109 spores/mL. The suspension was sonicated and vortexed. Sterile ethanol was then added, and the
solution was vortexed to create the final preparation with an ethanol concentration of 90 %, Tween® 20
concentration of 0.005 %, and a titer on the order of 108 spores/mL. Concentration of the spore
suspension was confirmed using the procedure defined in MOP 6535a (ARCADIS U.S., 2009).
2.6 Test Surface Characterization
Three types of surfaces—glass, asphalt roofing shingles, and concrete pavers—were investigated by RTI
International using techniques intended to describe the surface features and material characteristics.
These materials are described in Table 2-3. Three subtypes within each category were investigated to
provide a range of roughness within each material type. The three glass types are smooth glass, frosted
glass, and bubbled glass. Glass coupons were purchased from Dan's Glass, Inc. (Raleigh, NC, USA).The
three textures of concrete pavers are smooth, rough, and very rough, based on surface texture produced
by acid-etching. Roofing felt (#30W, Warrior Roofing Manufacturing, Tuscaloosa, AL, USA), economy
shingles (#241008006, Tarco Roofing, Little Rock, AR, USA), and architectural shingles (#0601070, GAP
Corp., Wayne, NJ, USA) were chosen for asphalt materials. Asphalt materials were purchased from a
Home Depot home improvement store in Raleigh, NC. Concrete coupons were purchased from Sonoma
Cast Stone Corp. (Petaluma, CA, USA). The nine materials were subjected to five material
characterization methods.
Table 2-3. Materials/Surfaces Used for Testing
Material
Glass
Asphalt roofing
shingles3
Roughness
Class
Smooth
Rough
Very rough
Smooth
Rough
Very rough
Details
Window glass ("float glass")
"Frosted" or "etched" glass
Architectural textured glass
Asphalt-impregnated "roofing felt"b, same
asphalt chemistry as roofing shingles, no
grit
"Economy" shingles, relatively thin and flat
"Architectural" shingles, relatively thicker
with more texture
7-3/4 in x 7-3/4 in x 1/8 in
coupons of "single-weight" glass
7-3/4 in x 7-3/4 in x 1/8 in
coupons of "single-weight" glass
7-3/4 in x 7-3/4 in x 1/8 in
coupons of "single-weight" glass
7-3/4 in x 7-3/4 in squares cut
with utility knife in laboratory
7-3/4 in x 7-3/4 in squares cut
with utility knife in laboratory
7-3/4 in x 7-3/4 in squares cut
with utility knife in laboratory
Continued on next page
14
-------
Material
Roughness
Class
Details
Smooth
Earth Crete mold release: cement simple
mold release with no sanding, finishing,
paint, wax, or sealants
Rough
Cast concrete
Earth Crete heavy acid etch: cement mold
release acid etched using 8 % muriatic acid
with no sanding, finishing, paint, wax, or
sealants (equivalent to typical sidewalk
finish)
Very rough
Earth Crete extra-heavy acid etch #1:
cement cast mold aggressive acid etched
using 8 % muriatic acid with no sanding,
finishing, paint, wax, or sealants
(equivalent to skid-resistant finish of
parking deck floor)
7-3/4 in x 7-3/4 in x 3/4 in thick
coupons prepared by Sonoma
Stone
7-3/4 in x 7-3/4 in x 3/4 in thick
coupons prepared by Sonoma
Stone
7-3/4 in x 7-3/4 in x 3/4 in thick
coupons prepared by Sonoma
Stone
Asphalt roofing shingles were selected as a proxy for asphalt paving. The shingles are readily available, are of
consistent formulation, and are easily handled in the laboratory. The base material of shingles is asphalt-
impregnated fiberglass. The surface grit is usually boiler slag, similar to the stone aggregate in asphalt paving.
bAsphalt-impregnated roofing felt (also called "felt roof deck protection") is felt saturated with asphalt to form a
composite that is essentially the same as the base layer of the shingles.
2.6.1 Hygroscopicity
The hygroscopicity measurement was used to provide an indication of the ability of the materials to
absorb water molecules. A thermal gravimetric analyzer (TGA) (model Q50, TA Instruments, New Castle,
DE, USA) determined changes in weight of samples relative to change in temperature. A small sample
was immersed in water and then wiped dry. The sample was then placed in the TGA, and the reduction in
mass was recorded throughout the programmed heating cycle. Assuming water loss was the only factor
affecting mass change, the hygroscopicity of the sample was calculated. Three samples of each material
were measured for hygroscopicity.
2.6.2 Surface Tension
The surface tension of water droplets on the test material was used as an indicator of the hydrophobic or
hydrophilic nature of the surface. The angle formed between the solid/liquid interface and the liquid/vapor
interface is referred to as the contact angle. A droplet with an angle greater than 90° indicates a
hydrophobic surface. Likewise, a droplet with a contact angle less than 90° indicates a hydrophilic
surface. Contact angles approaching 0° are possible if the droplet has a strong chemical attraction to the
solid surface, resulting in the liquid spreading thinly across the surface. Contact angle measurements
were obtained with a Rame-Hart 200-F1 goniometer (Rame-Hart Instrument Co., Succasunna, NJ, USA).
A distilled water droplet of 10 microliters (uL) was deposited on each surface, and an image was captured
of the profile. Imaged, a free image analysis tool developed by the National Institutes of Health (NIH), was
used to measure the incident angles. At least two angle measurements were taken on either side of the
droplet for each material.
Using the following equations, the first being Young's Equation and the latter defining surface energy as
the difference between the solid-liquid interface and the sum of the liquid-vapor and solid-vapor interface
(Girifalco and Good, 1957), one can set ysv - VSL and reduce to solve for surface energy, £, in Joules per
square meter (J/m2):
15
-------
YLV cosd = YSV-YSL
E=YLV+ Ysv - YSL
E = YLV (1+cos0), where
YLV = liquid-vapor surface energy (Yiv-water= 0.072 Newtons per meter [N/m])
Ysv = solid-vapor surface energy
ySL = solid-liquid surface energy
£ = total surface energy between particle and surface
To determine surface energy for dry particle deposition, the Hamaker constant for each material was
referenced from literature. The Hamaker constants used are listed in Table 2-4.
Table 2-4. Hamaker Constants
Surface Type
Glass
Concrete
Asphalt
Hamaker Constant, A (J)
6.30E-20
2.41E-19
1.81E-19
Source
Wikipedia(2014)
Lomboy et al. (2011)
Tomas (2006)
The adhesion force between a dry sphere and dry surface was calculated using the relationship between
a smooth sphere and a surface described in Rabinovich et al. (2002):
Adhesion forcedry =
Ar
1
1
6-H02 [1+ 58.14-A--RMS/22 (1 + 1.817 -RMS/H0)2
A = Hamaker constant
r= particle radius
H0 = minimum separation distance between particle and asperity (0.3-0.4 nm)
RMS = root mean square roughness of the surface
A = average peak-to-peak distance between asperities
The Derjaguin-Muller-Toporov adhesion theory equation (Rabinovich et al., 2002), solved for y, was used
to convert the force to surface energy:
Adhesion forcedry = 4 • njr
2.6.3 Surface Roughness
Surface roughness of the materials was measured on two different scales. A micrometer-scale roughness
measurement determines the contact area between the spore and surface. This scale is a factor in the
spore adhesion to the surface. A millimeter-scale roughness determines the airflow boundary layer
characteristics across the surface. This scale is a factor in the lift force acting on the spore and
determines detachment. Surface roughness, Ra, for both scales is reported as the arithmetic mean of the
absolute values of measured distance from the median plane of the surface.
16
-------
Micrometer Scale
Two methods quantified the micrometer-scale surface roughness of the materials. A Micro Photonics
TR200 surface roughness tester (Micro Photonics, Inc., Allentown, PA, USA) was employed first. The
contact profilometer relies on a piezoelectric-operated diamond stylus to measure the contour of the
surface and provides the arithmetic mean profile (in urn) of the surface of interest. All materials except for
the smooth and frosted glass had surface features with scales that exceeded the profilometer range of
motion. Consequently, SEM combined with digital image analysis was employed as an alternative method
of determining roughness on the micrometer scale. Five images were taken from various locations on a
single sample. Three samples of each material were imaged, providing 15 images for each material at a
single magnification. Images of the materials were collected at 300x and 10,000x magnifications.
The SEM images were analyzed individually using Imaged software. All images were preprocessed by
converting them to 8-bit images, enhancing the contrast, applying a topographical threshold value based
on the frequency and intensity of shading in the image, and converting the image to binary. After
preprocessing, the Roughness Calculation plug-in (Chinga and Dougherty, 2006) was used to determine
the surface roughness, Ra. The image-generated roughness value for smooth glass was scaled to the
value measured with the profilometer, thus providing a factor to convert all materials from image-
generated roughness values to actual micrometer-scale roughness values.
Millimeter Scale
To quantify the millimeter-scale roughness of the test materials, optical images were taken at 2x
magnification. The images were processed in the same manner as the SEM images. However, a z-axis
scale proved to be unattainable for optical images at low magnification. Micro-scale SEM images were
able to provide a greater range of depth and clarity at specific locations on the sampling coupons that
could be compared directly to the function of a profilometer. Optical images could not provide this same
basis of comparison. Instead, ImageJ-derived roughness measurement outputs for optical images were
normalized to the lowest measured sample (smooth glass), thus providing a unitless range for roughness
relative to smooth glass.
2.6.4 Surface Area
Actual Surface Area
The surface area of a given material can vary greatly based on the size of the material as well as the
contour of the surface. Aflat surface like smooth glass has very few, if any, ridges or valleys. In contrast,
a rough surface like asphalt roofing shingles has many ridges and valleys. Thus, the surface area of a
smooth surface will be much less than the surface area of a very rough material of the same size. The
amount of surface area can greatly influence the number of attachment points for a contaminant.
The SEM and optical images used for roughness measurements were also used to measure surface area
for all test materials. All images were preprocessed in Imaged by first using the Set Scale function to
measure a known distance on the image. This known distance was either the scale provided on the SEM
image or, for the optical images, the length of the entire coupon. After the scale was set, the images were
converted to 32-bit images and the contrast was enhanced. The SurfCharJ plug-in (Imaged, 2011) was
used to calculate the surface area of the image. The surface area output was normalized with the
projected area of the chosen image and was therefore unitless; however, with knowledge of the pixel
length and scale, the actual surface area was calculated.
17
-------
Specific Surface Area
The specific surface area measurement gave an indication of the microstructure of the surface of interest.
The specific surface area, in square meters (m2)/grarn (g), of small samples of the test coupons were
determined by measuring the physical adsorption of nitrogen gas molecules on the solid surface with a
Micromeritics ASAP 2020 accelerated surface area and porosimetry system (Micromeritics, Norcross,
GA, USA). Using the Brunauer-Emmett-Teller (BET) algorithm to interpret the data from the instrument,
the internal surface area of porous materials was measured. The system made multiple measurements
per sample, and the standard deviation (SD) was calculated automatically. For all types of glass, four
measurements were collected per coupon, and the majority of these measurements were below the
detectable limit for the instrument. For the pavers, four measurements were taken on three different
coupons for each type of paver. Four measurements were taken on a single coupon for each type of
asphalt material. Half of the measurements for roofing felt and rolled roofing were below detectable limits.
2.7 Test Protocol
The flow chart in Figure 2-18 provides a brief description of the test protocol for wet deposition
experiments in the RWT. The same procedure was used for experiments at EPA and DPG. However,
implementing the procedure in the laboratory at DPG was complicated by the restrictions imposed on
working with BSL 3 agent, and additional steps were necessary to provide the required levels of
containment at all times. All procedures were conducted according to MOPs developed specifically for
this project that detail the methods for depositing wet and dry spores onto test coupons (Alion, 2014a and
2014b), conducting reaerosolization tests (Alion, 2014c), and handling filter samples (Alion, 2014d).
The samples generated in each experiment include the following:
• Positive control: an approximately 3-5 ml sample of the particulate solution transferred into a sterile
conical tube as a positive control.
• Reference tube: a reference sample collected for each deposition test by dispensing 0.2 ml of
particulate solution directly into a conical tube containing 10 ml of PBST.
• Filters: four polyester felt filters used for each experiment to collect particulates reaerosolized from
the test coupon and transferred to sterile Stomacher bags for extraction according to MOP 6600
(ARCADISU.S.,2013b).
• Field blank filter: a field blank filter collected for each day of testing in a Stomacher bag as a negative
control.
• Coupon wipes: a wipe sample of the test coupon surface extracted according to MOP 6567
(ARCADIS U.S., 2010) performed as a qualitative check that particulates were deposited on the
coupon surface.
18
-------
> Dust coupon
t Glue coupon to disposable riser, if needed
> Equilibrate coupon at room T and RH
Deposition Chamber
Collect posKive control sample
Collect reference tube sample
Deposit 0.2 ml of surrogate solution on coupon
Place coupon in T/RH chamber to equilibrate
ter Loading
Load autoclaved filters into clean, dry filter holders
J Place filter holders in labeled plastic bags
Reaerosolization Wind Tunnel Test
Place loaded coupon in RWT
Install four filter holders in RWT
Close RWT
Turn on RWT fan with damper closed
Run experiment and simultaneously
>• Open RWT fan damper
>• Open valve lo air jet
>• Turn traverse switch to forward
When traverse completes motion (~ 1 min), simultaneously
^- Close RWT fan damper
>• Close valve to air jet
>• Turn traverse switch to return
Remove filter holders and place in correct, labeled bags
Remove coupon
Run sweeps to clean RWT
Repeat procedure for all scheduled experiments
Sample Collection
Wipe surface of test coupon after experiment
and place wipe in labeled tube
In clean biosafety cabinet, remove filters from filter
holders and place in labeled Stomacher bags
Place field blank filter in labeled Stomacher bag
\
I Cleanup
j Disinfect and dispose of coupon
4 Clean biosafety cabinet
4 Wash disassembled filter holders with amended bleach,
rinse with deionized water, and rinse with isopropyl alcohol
> Leave clean filter holders to dry in clean biosafety cabinet
Figure 2-18: Flow chart describing the test protocol for deposition experiments in the RWT.
19
-------
2.8 Nondimensional Analysis
With a database of surface characteristic measurements and experiment descriptors, including particle
diameter, density, relative humidity, air jet height from the surface, air jet traverse speed, and modeled
shear stress on the surface, a spreadsheet was assembled to provide all known descriptors as
independent variables and the experimentally realized reaerosolization fraction as the single dependent
variable.
To derive relationships between the reaerosolization fraction and the parameters described previously,
dimensional analysis was performed. The parameters were nondimensionalized using Buckingham Pi
Theory (Buckingham, 1914). Particle diameter, particle density, and jet traverse rate were chosen as
parameters with repeating units of mass, length, and time. The parameters of relative humidity and jet
height were left out of the analysis because a single value was used throughout the experiments, thus the
impact of these parameters could not be determined. The result of the nondimensional analysis was
F -
where
F = reaerosolization fraction
cfp = particle diameter
pp = particle density
NR = jet traverse speed
£ = surface energy
T= shear stress
TGA = mass reduction
R-, = microscale roughness height
R2 = macroscale roughness height
A-, = microscale surface area
A2 = macroscale surface area
S1 = microscale roughness spacing between asperities
2.9 Modeling
Two models were developed to predict reaerosolization of particles from surfaces — one for wet particles
and another for dry particles. A similar approach was taken for development of both models.
To develop the predictive model for wet reaerosolization, all wet deposition experiments performed for Btk
and Ba at both laboratories were used. The combined data consisted of 254 observations, 135 of Btk
from the EPA laboratory, 54 of Btk from DPG, and 65 of Ba from DPG performed under varying surface
conditions and shear stresses.
The dry particle data set contained 108 observations, 80 of Btk from the EPA laboratory, 12 of Btk from
DPG, and 16 of Ba from DPG. Due to time constraints, DPG was not able to conduct experiments with
asphalt, thus there was no basis for comparison of experimental results with that material. Therefore the
analysis for the dry model did not include any asphalt experiments, reducing the number of observations
to 82.
20
-------
A multiplicative model was selected to understand the significance of each variable and impact on
predicting reaerosolization. The nondimensional variables were directly log transformed. Descriptive
statistical analyses were performed to understand the distribution of data in each variable and the
relationship between the variables. Data management and statistical analyses were performed using SAS
software, version 9.3 (SAS Institute, Inc., Gary, NC). For individual variables, the summary statistics
(range of data, mean, and SD) and data distribution plots were examined. In addition, Pearson correlation
coefficients were calculated between all variables. Collinearity diagnostics were performed to examine the
degree of involvement and potential impact during the model development.
The transformed terms were then fitted to an additive linear regression model. For the first step, the terms
that include surface energy and shear stress, found as significant factors in reaerosolization in the
literature, were forced to be included in the model. For the second step, all possible additive models that
included surface energy and shear stress were constructed, and the best five models were selected
based on several selection criteria: high adjusted R2, low root mean squared error, low Akaike information
criterion (AIC), and Mallows's Cp. The five models were further compared with regard to the significance
of the contributing variables, and the most meaningful and best-fitting model was selected. Finally, the
model was examined for the possibility of improvement by addition of other remaining variables. A
significance level of 0.15 was used to add and retain the remaining variables in the model.
The final regression model was examined for significance of individual variables in the model. The model
was also verified for the assumptions in regression by checking residuals plots for normality, homogeneity
of variance, and independence of the residuals. Potential outliers and the degree of collinearity were also
investigated.
21
-------
3 Results and Discussion
3.1 Materials Characterization
Samples of each test material were investigated using the techniques described in section 2.6 to describe
the surface features and material characteristics. Table 3-1 gives a summary of the average values for
each test conducted for each material. Note that the specific surface area information was omitted from
the table and from the surface analysis due to lack of results as described previously. The following
sections provide more detailed results including the minimum, maximum, mean, and SD of the
measurements.
Table 3-1. Summary of Surface Measurements Used for Particle Reaerosolization Experiments and
Modeling
Surface
Material
Asphalt
Smooth
Asphalt
Rough
Asphalt
Very Rough
Concrete
Smooth
Concrete
Rough
Concrete
Very Rough
Glass
Smooth
Glass
Rough
Glass Very
Rough
TGA
(% Mass
Loss)
5.57
0.309
0.451
2.09
0.550
1.12
0
0
0
Roughness,
Micro
(Mm)
25.7
29.4
29.6
34.8
33.3
35.5
0.0147
9.74
0.509
Peak-to-
Peak Space
Micro (urn)
2.06
0.545
1.21
0.884
1.49
1.03
0.166
1.28
0.778
Roughness,
Macro
(unitless)
3.28E+04
7.93E+04
9.62E+04
9.51 E+04
1.03E+05
1.17E+05
2.99E+02
4.12E+03
4.13E+04
Surface
Area, Micro
(Mm2)
1.45E+03
1.45E+03
1.13E+03
1 .28E+03
1.37E+03
9.31 E+02
1.28E+03
1.31E+03
2.50E+03
Surface
Area, Macro
(Mm2)
1.23E+13
4.21 E+1 3
3.73E+13
6.69E+12
4.71 E+1 3
4.09E+13
4.58E+12
4.38E+12
1.46E+13
Surface
Energy,
Wet (N/m)
0.0655
0.0995
0.0889
0.135
0.137
0.144
0.143
0.110
0.144
Surface
Energy,
Dry (N/m)
0.0356
0.0356
0.0356
0.0473
0.0473
0.0473
0.0124
0.0124
0.0124
3.1.1 Hygroscopicity
Results from the hygroscopicity tests conducted with the TGA are presented in Table 3-2 in units of %
mass loss. This instrument gives a measurement of the ability of a surface to absorb water. A higher
value indicates a more hygroscopic surface. With the exception of the glass coupons, which are not at all
hygroscopic, the least hygroscopic material was the rough asphalt, and the most hygroscopic material
was the smooth asphalt.
22
-------
Table 3-2. Hygroscopicity Measurements (% Mass Loss)
Material
Asphalt
Asphalt
Asphalt
Concrete
Concrete
Concrete
Glass
Glass
Glass
Roughness
Class
Smooth
Rough
Very rough
Smooth
Rough
Very rough
Smooth
Rough
Very rough
Minimum
3.67
0.100
0.040
1.60
0.260
0.680
0
0
0
Maximum
11.0
0.875
1.56
2.40
0.790
1.67
0
0
0
Mean
5.57
0.309
0.451
2.09
0.550
1.12
0
0
0
SD
3.62
0.378
0.742
0.273
0.181
0.350
0
0
0
3.12 Surface Tension
Goniometry measurements are presented in Table 3-3 with values given as the cosine of the measured
contact angle. This method provides a measurement of the surface tension of water droplets on the test
material surface and was used as an indicator of the hydrophilic or hydrophobic nature of the surface. A
lower mean value, corresponding to a higher contact angle, indicates greater beading of water, greater
capillary force, and greater surface adhesion force. Water droplets on the smooth and rough concrete
surfaces had both zero and nonzero contact angles depending on the hydrophilic or hydrophobic nature
of the specific aggregate particle in contact with the water droplet. Results show that the most
hydrophobic test materials were the three roughness classes of asphalt and the rough glass. The smooth
and very rough glass and all roughness classes of concrete were shown to be hydrophilic.
Table 3-3. Goniometry Measurements Given as the Cosine of the Contact Angle (radians)
Material
Asphalt
Asphalt
Asphalt
Concrete
Concrete
Concrete
Concrete
Concrete
Glass
Glass
Glass
Roughness
Class
Smooth
Rough
Very rough
Smooth
Smooth
Rough
Rough
Very rough
Smooth
Rough
Very rough
Contact
Angle, 9
>0
>0
>0
>0
0
>0
0
0
>0
>0
0
Minimum
-0.12
0.32
0.11
0.78
1.00
0.85
1.00
1.00
0.98
0.47
1.00
Maximum
-0.06
0.43
0.36
0.99
1.00
0.98
1.00
1.00
0.99
0.55
1.00
Mean
-0.09
0.38
0.24
0.89
1.00
0.91
1.00
1.00
0.98
0.53
1.00
SD
1.00
1.00
0.99
0.99
1.00
0.99
1.00
1.00
1.00
1.00
1.00
23
-------
3.1.3 Surface Roughness
Table 3-4 presents the microscopic vertical surface roughness, a measurement of the size of features
that determine the contact area and number of contact points between the surface and a spore. The
values were measured from SEM images captured at 10,000x magnification, with a larger value indicating
a rougher surface. The microscopic vertical surface roughness is proportional to the adhesion force
between a spore and the surface. The mean values for asphalt and concrete were very similar, and the
values for glass were much lower.
Table 3-4. Microscopic Surface Roughness (Mm) Determined from SEM Images
Material
Asphalt
Asphalt
Asphalt
Concrete
Concrete
Concrete
Glass
Glass
Glass
Roughness
Class
Smooth
Rough
Very rough
Smooth
Rough
Very rough
Smooth
Rough
Very rough
Minimum
18.5
20.7
15.9
21.0
23.7
24.4
0.0003
2.99
0.025
Maximum
36.4
39.4
39.0
42.2
42.1
41.6
0.060
17.20
2.40
Mean
25.7
29.4
29.6
34.8
33.3
35.5
0.0147
9.74
0.51
SD
5.58
4.56
5.90
5.41
5.36
5.20
0.0167
3.77
0.61
Table 3-5 presents the microscopic horizontal surface roughness, a measurement of spacing between the
microscopic roughness features (asperities) on the surface, as measured from SEM images captured at
10,000x magnification. The microscopic horizontal surface roughness can be used as a surrogate
measure of the frequency of asperities on the surface, with a smaller value indicating a rougher surface
(with more asperities). On the smooth glass surfaces, very few asperities were counted, but they tended
to be close to each other.
Table 3-5. Microscopic Horizontal Surface Roughness (Peak-to-Peak Space, urn) Determined from
SEM Images
Material
Asphalt
Asphalt
Asphalt
Concrete
Concrete
Concrete
Glass
Glass
Glass
Roughness
Class
Smooth
Rough
Very rough
Smooth
Rough
Very rough
Smooth
Rough
Very rough
Minimum
0.160
0.0149
0.223
0.219
0.320
0.135
0.0509
0.358
0.0913
Maximum
6.55
1.35
2.55
2.48
4.31
2.52
0.280
3.01
2.22
Mean
2.06
0.545
1.21
0.884
1.49
1.03
0.166
1.28
0.778
SD
2.23
0.488
0.794
0.717
1.36
0.718
0.162
0.959
0.621
24
-------
Table 3-6 presents the macroscopic surface roughness measurements that were determined from optical
images of the materials taken at 2x magnification. These values provide a measure of the height of the
surface features that determine the boundary layer velocity profile of the air flow over the material, with a
larger value indicating a rougher surface. This surface roughness is inversely proportional to the lift force
applied to spores on the surface. The rough and very rough asphalt and all concrete types were the
roughest surfaces. The smooth and rough glass were the smoothest surfaces.
Table 3-6. Macroscopic Surface Roughness (unitless) Determined from Optical Images
Material
Asphalt
Asphalt
Asphalt
Concrete
Concrete
Concrete
Glass
Glass
Glass
Roughness
Class
Smooth
Rough
Very rough
Smooth
Rough
Very rough
Smooth
Rough
Very rough
Minimum
1.99E+04
7.61 E+04
8.78E+04
8.49E+04
9.46E+04
1.02E+05
1.00E+00
8.44E+02
2.65E+04
Maximum
4.58E+04
8.35E+04
1.08E+05
1.12E+05
1.09E+05
1.32E+05
1.34E+03
9.65E+03
6.30E+04
Mean
3.28E+04
7.93E+04
9.62E+04
9.51 E+04
1.03E+05
1.17E+05
2.99E+02
4.12E+03
4.13E+04
SD
7.62E+03
2.28E+03
5.45E+03
6.78E+03
3.35E+03
8.37E+03
3.40E+02
2.47E+03
1.10E+04
Three different measurements of surface roughness were used to characterize the test surfaces. All of
the measurements showed that the smooth asphalt, rough asphalt, very rough asphalt, smooth concrete,
rough concrete, and very rough concrete were similar in surface roughness even though they appear very
different to the naked eye. The glass surfaces were much smoother than asphalt and concrete. There
were some significant differences in the glass roughness categories: the rough glass had higher
microscopic roughness than smooth and very rough glass, while the macroscopic roughness increased
by an order of magnitude from smooth to rough glass and another order of magnitude from rough to very
rough glass.
3.1.4 Surface Area
Table 3-7 contains results for surface area determined from the SEM images at 10,000x magnification.
This measurement provides quantification of the microscopic surface features that determine the contact
area and number of contact points between the surface and a spore. The microscopic surface area is
proportional to the adhesion force between the surface and a spore. The values obtained for all surfaces
were very similar.
25
-------
Table 3-7. Microscale Surface Area Measurements (um ) from SEM Images
Material
Asphalt
Asphalt
Asphalt
Concrete
Concrete
Concrete
Glass
Glass
Glass
Roughness
Class
Smooth
Rough
Very rough
Smooth
Rough
Very rough
Smooth
Rough
Very rough
Minimum
8.02E+02
6.40E+02
5.65E+02
5.66E+02
5.29E+02
5.64E+02
1.01E+03
6.74E+02
7.91 E+02
Maximum
3.71 E+03
2.09E+03
1.74E+03
2.07E+03
2.02E+03
1.38E+03
2.08E+03
2.69E+03
4.09E+03
Mean
1.45E+03
1.45E+03
1.13E+03
1.28E+03
1.37E+03
9.31 E+02
1.28E+03
1.31 E+03
2.50E+03
SD
6.72E+02
4.40E+02
2.89E+02
4.90E+02
5.25E+02
2.63E+02
3.82E+02
6.03E+02
9.48E+02
Surface area results determined from the optical images at 2x magnification are given in Table 3-8. This
measurement provides quantification of the macroscopic surface features that determine the boundary
layer velocity profile of the airflow over the material. This measure of surface area is inversely
proportional to the lift force applied to spores on the surface. The results obtained for all surfaces were
very similar, but 5 x 1013 urn2 appears to be the upper limit of the method.
Table 3-8. Macroscale Surface Area Measurements (um ) from Optical Images
Material
Asphalt
Asphalt
Asphalt
Concrete
Concrete
Concrete
Glass
Glass
Glass
Roughness
Class
Smooth
Rough
Very rough
Smooth
Rough
Very rough
Smooth
Rough
Very rough
Minimum
8.10E+12
3.78E+13
2.88E+13
6.18E+12
4.29E+13
3.72E+13
4.09E+12
3.81E+12
1.34E+13
Maximum
1.68E+13
4.60E+13
4.24E+13
7.53E+12
5.47E+13
4.46E+13
5.17E+12
6.09E+12
1.62E+13
Mean
1.23E+13
4.21E+13
3.73E+13
6.69E+12
4.71E+13
4.09E+13
4.58E+12
4.38E+12
1.46E+13
SD
2.99E+12
2.42E+12
3.30E+12
3.58E+11
2.97E+12
1.94E+12
3.08E+11
4.60E+11
7.57E+11
3.2 Scouting Tests
Many scouting tests, listed in Table 3-9, were conducted with wet-deposited Btkto determine the final
experimental procedure and experimental details such as surface loading levels and jet velocities. Fifty-
four individual reaerosolization experiments were run with various loaded coupons at fixed jet velocity
with the purpose of fine-tuning the experimental procedure, assessing suitability of surface loading levels,
and estimating reproducibility and precision of results. Surface loading levels were varied over three
orders of magnitude by changing the spore concentration of the inoculant and the amount of inoculant
26
-------
applied to the test coupon. The goal was to have the majority of spores deposited as single spores. For
this reason, the spore concentration of the inoculants could not be of a higher order than 108 CFU/mL, or
many spores would be contained in each droplet from the ultrasonic nozzle. The amount of inoculant
deposited on the surface varied between 0.1 and 0.6 ml. The volume chosen was 0.2 ml to avoid any
puddles forming on the coupon surface. Therefore, the surface loading level chosen for the wet
deposition experiments was 0.2 ml of 108 CFU/mL.
An additional 18 reaerosolization experiments were performed with loaded coupons at multiple jet
velocities with the primary purpose of determining the low, medium, and high jet velocity settings for each
surface type. The jet velocity is set by adjusting the pressure in the compressed air line supplying the
slotted nozzle. Six of the varying jet velocity experiments were conducted at 11 nozzle pressures, from 0
to 2.5 atm, and 12 later experiments were conducted at eight nozzle pressures, ranging from 0 to 3.5 atm.
When the nozzle pressure is set to 0 atm, there is no air flowing through the nozzle, and the
reaerosolization force is the wind velocity of 5.8 miles per hour (mph) generated by the RWT fan.
Table 3-9. Number of Scouting Tests Completed by Coupon Type
Material
Glass
Asphalt
Concrete
Roughness Class
Smooth
Rough
Very rough
Smooth
Rough
Very rough
Smooth
Rough
Very rough
Tests with Single
Jet Velocity
3
17
0
6
6
0
11
11
0
Tests with Multiple
Jet Velocities
3
3
3
1
1
1
2
2
2
After samples from the reaerosolization tests were extracted, plated, enumerated, and verified by the
microbiology laboratory, the microbiological data were entered into a spreadsheet for analysis.
Calculations from the raw microbiological data include total measured CFU for each sample (filter, wipe,
reference sample tube, or positive control),
CFIL
plate counts x extraction volume .
sample p|ated volume x dilution factor '
total CFU collected on the four filters for each test,
CFUtotal =CFUfilter1 +CFUfilter2 +CFUfilter3 + CFUfilter4 ;
and fraction reaerosolized, F, for each test,
F= CFUtotal
CFUref x deposition efficiency
CFUref is the total number of CFU in the reference tube sample, and the deposition efficiency was
determined from the SS coupon tests described in section 2.3.1.
27
-------
Figure 3-1 presents results from wet-deposited Btk experiments with single jet velocities. The bars
represent the average F from three separate tests for each surface type and jet velocity combination, and
the error bars represent the CV. The overall trend in the results was that very little was reaerosolized from
glass coupons, while significant numbers of spores were reaerosolized and collected from both asphalt
and concrete. In general, the number of spores reaerosolized from materials asphalt and concrete were
one to two orders of magnitude greater than the number of spores reaerosolized from glass.
2.5E-2
2.0E-2
8 1.5E-2
l.OE-2
5.0E-3
O.OE+0
Surface type and jet velocity
Figure 3-1: Scouting test average Ffor wet-deposited Btk by surface type and jet velocity. Each
bar represents the average of three repetitions, and the error bars represent the CV.
Figures 3-2 through 3-4 present F with varying jet velocity for each surface type. The error bars represent
the CV calculated from the replicate tests presented in Figure 3-1. There are no error bars in Figure 3-4
because replicate experiments were not conducted for very rough surfaces in the scouting tests. For the
varying jet velocity experiments, only filter #1 (the filter on the bottom right in Figure 2-5) was collected
and analyzed for each jet velocity to minimize the number of microbiological samples for processing and
plating to conserve time and resources while still providing valuable data for determining the low,
medium, and high jet velocities to be used for each test surface in the final wet deposition experiments.
The F shown in Figures 3-2 through 3-4 was calculated as
F =
amount collected on filter #1 / 0.42
amount deposited on coupon
where 0.42 is the average fraction collected on filter #1 calculated from results of 33 individual
reaerosolization experiments. A great deal of variability was encountered in the fraction collected on each
of the four filters, and this variability was likely a major contributor to the variability in results in Figures 3-2
through 3-4.
28
-------
1.4E-2
1.2E-2
1.0E-2
O.OE+0
-Smooth asphalt
-Smooth glass
Smooth concrete
50
100 150 200
Jet velocity (mph)
250
300
Figure 3-2: F for wet-deposited Btk calculated from scouting tests with varying jet velocity for
smooth surfaces.
•o
01
N
"5
V)
-*— Rough asphalt
- -Rough glass
Rough concrete
50
100 150 200
Jet velocity (mph)
250
300
Figure 3-3: F for wet-deposited Btk calculated from scouting tests with varying jet velocity for
rough surfaces.
29
-------
2.5E-2
2.0E-2
O.OE+0
Very rough asphalt
-Very rough glass
-Very rough concrete
50
100 150 200
Jet velocity (mph)
250
300
Figure 3-4: F for wet-deposited Btk calculated from scouting tests with varying jet velocity for very
rough surfaces.
3.3 Reaerosolization Experiments
Reaerosolization tests completed include wet-deposited Btk and Bg at EPA, wet-deposited Btk and Ba-
Ames at DPG, dry-deposited Btk and Bg at EPA, and dry-deposited Btk and Ba-Ames at DPG. The
reaerosolization experimentation was designed as a 33 full factorial experiment for each
laboratory/particle type/deposition type combination. The three factors were material (asphalt, concrete,
glass), roughness level (smooth, rough, very rough), and jet velocity (low, medium, high). Thus, three
factors, each at three levels, yielded 27 experimental combinations.
In addition, as many replicates as logistically feasible, given the constraints of time and resources, were
collected for each laboratory/particle type/deposition type combination. Each replicate was treated as a
block, and the experimental order was completely randomized within each replicate. Time constraints at
DPG did not permit the execution of the 33 full factorial experiment for dry deposition testing. Instead, two
material/roughness combinations were selected for testing at two jet velocity levels, and as many
replicates as possible were completed of that 22 experiment.
The results expressed as F for each material/roughness/jet velocity combination were calculated and
averaged, and the CV was calculated. The statistician compared appropriate data sets to determine
whether there were differences in F between the laboratories, between the particle types, and between
replicates of the same laboratory/particle/deposition combination (Table 3-10). Outlier testing was used to
answer the question of whether the two separate laboratories were generating equivalent results. The
outlier testing was checked with Dixon's r10 ratio for Btk (the only spore type common to both laboratories)
for every deposition type/material/roughness/jet velocity combination.
30
-------
Table 3-10. Summary of Findings from Statistical Analysis of Reaerosolization Data
Lab
EPA
EPA
EPA
EPA
DPG
DPG
DPG
DPG
Spore
Btk
Btk
Bg
Bg
Btk
Btk
Ba
Ba
Deposition
Type
Wet
Dry
Wet
Dry
Wet
Dry
Wet
Dry
Significant Effects
Material
Asphalt > glass
Concrete > glass
Concrete > asphalt
Concrete > glass
Concrete > asphalt
Concrete > glass
Concrete > asphalt
None
None
None
None
Roughness
Rough > very
rough
None
None
None
None
None
None
None
Speed
High > low
High > medium
Medium > low
None
High > low
High > medium
High > low
High > medium
None
None
None
None
Comments
All two-way and three-
way interactions
significant
Replicate effect
observed
Significant
surface/speed
interaction; replicate
effect observed
Material and roughness
confounded
Replicate effect
observed
Material and roughness
confounded
Statistical analyses of the experiments were conducted as multiway ANOVAs with F as the response
variable, and all factors, their interactions, and replicates entering the model. Residuals were used to
check statistical assumptions of the tests, and no problems were found. Comparisons between the
different levels within each factor were done using least-squares means, which allows comparisons to be
made cleanly within the factors of interest without interference from another effect. For example,
comparisons between particle types could be made without being concerned about a surface effect
interfering with the particle type comparison. Replicate effects were found in three data sets, as noted in
Table 3-10. Such effects might be due to several causes, including differences in spores or spore
preparation, different personnel conducting the tests, or different personnel analyzing the samples.
The wet deposition data for Btk spores showed no evidence of significant differences between the two
laboratories. The dry deposition data for Btk spores also did not indicate a difference between the
laboratories, thus validating the equivalence of experimental equipment and methods at EPA and DPG.
When F results from different particle types were compared, no statistically significant differences were
found between Btk spores and Ba spores for either wet or dry deposition. Table 3-11 presents the overall
mean F calculated from all tests conducted for a particular particle type/deposition combination. Bg
spores yielded significantly lower wet-deposition F results than for Btk and Ba spores. The overall mean F
for wet-deposited Bg spores was 79 % lower than for Btk and Ba spores.
31
-------
Table 3-11. Overall Mean F Values Calculated from Reaerosolization Tests
Lab
EPA
EPA
DPG
DPG
EPA
EPA
DPG
DPG
Spore
Bg
Btk
Btk
Ba
Bg
Btk
Btk
Ba
Deposition
Type
Wet
Wet
Wet
Wet
Dry
Dry
Dry
Dry
# of Data
Points
81
135
54
67
81
81
12
16
Average F
33 Test Matrix
4.09E-4
1.21E-3
2.07E-3
3.03E-3
5.79E-4
1.28E-2
-
-
Btk-Ba
Combined F
33 Test Matrix
-
1.96E-3
-
-
-
-
Average F
22 Test Matrix
-
9.26E-4
9.60E-4
2.38E-3
-
2.15E-2
5.11E-4
1.10E-3
Btk-Ba
Combined F
22 Test Matrix
-
1.34E-3
-
7.05E-3
Another finding of the study was that a significantly lower fraction of wet-deposited particles reaerosolized
compared to dry-deposited particles. The combined mean F for wet-deposited Btk and Ba spores from
the limited 22 experiment was 81 % lower than the combined mean F for dry-deposited Btk and Ba spores
calculated from the same subset of experimental conditions (Table 3-11).
Results from experiments conducted in the EPA laboratory showed significant differences in Ffrom
different material types, with concrete consistently showing higher F values than asphalt and glass. No
significant effect of surface roughness level was found within any part of the experiment with the
exception of the EPA laboratory wet-deposited Btk data. This result was not surprising based on the
surface roughness characterization data presented in Tables 3-4 and 3-6, as there was not generally a
large difference in the measured surface roughness within each material type even though the surfaces
appear different to the naked eye. The wet deposition data did show statistically significant differences in
F related to the jet velocity, with the high velocity consistently yielding greater reaerosolization than low
and medium velocity.
No significant effects of material type, roughness level, or jet velocity were seen on F in the DPG
laboratory results because a small number of tests were conducted in that facility and the material and
roughness were confounded. It is not surprising that conducting more tests, in particular multiple
replicates of the full 27-test experiment, leads to better understanding of the factors that affect
reaerosolization.
3.4 Propagation of Error Analysis
A propagation of error analysis was performed to determine the cumulative effect of uncertainties and
limitations in methods and measurements used throughout the experiments and modeling effort. The error
provides the standard deviation for each experimental and modeled data point and can be described by
To determine the standard deviation of each measurement, the sources of potential error need to be
identified. For the reaerosolization experiments, the potential sources of error per trial were determined to
be microbiological plate count variability in the reference samples and the test samples (A/p), sample
32
-------
extraction efficiency (Vex), dilution error (DF), plated error (VP), collection filter efficiency (Effmer), gel filter
efficiency (Effgej), and particle deposition efficiency (Depeff). The general equation takes the form of
+
\dND
if
id
(
x2
( dF
\dDepEff
9F\2 2
™J V
dF \2
^Efffnter-'
NP \dVe><
)2sU
+ (SF'
S2 +
EfffHter
1 (3F
ff \dNt
1 \2c-2
(dDFj °DFJ
)X
J^P \dDFj
Depos/
SDF
ted
to represent the total error associated with each experiment.
The uncertainty in the deposition efficiency was found to be the largest source of error in the experimental
results. The deposition efficiency was determined to be 48 % ± 31 % from the average of eight trials. It is
possible that the deposition efficiency varied between replicates and laboratories.
The uncertainty in the model was determined by establishing a standard deviation for each parameter
present in the final model equations. In addition, the partial derivative of the model equation was determined
with respect to each variable in the equation. Each partial derivative was squared and multiplied by the
variance of its respective variable. The standard deviation of each modeled data point was determined as
the square root of the sum of squares. The error equations for the wet and dry models are, respectively,
VI I n I VI \ n
^r^l S2 +(T—I S^ +
ap
dF
c2 + _ 92+ — Q2 + /^_ 92
SNa+\ a. ] Sr + lgEJ SE+lao J SR?
and
dA2
For the wet particle reaerosolization model, the largest contributor to error for predicted reaerosolization
from smooth glass was the macroscale roughness height. The measurements of roughness for this
material were unitless and considered relative to other material measured, but microscale roughness
gives an indication that this surface is nearly flat. Thus, any variability in this measurement results in a
large standard deviation. The largest contributor to error for predicted reaerosolization of wet particles
from all other materials was the particle size. The particle size used in the model was 1 urn with a
standard deviation of 0.2 urn based on the particle deposition distribution determined by SEM images of
characteristic deposition.
33
-------
For the dry particle reaerosolization model, the largest contributor to error was the particle size. The dry
model uses a particle size of 2.41 urn with a standard deviation of 0.66 urn, determined by measuring size
distributions of particles deposited on SEM stubs after a deposition run. The effect of particle diameter on
error can be attributed to the weight that the variable carries in the model equations. Error bars determined
through the propagation of error analysis are displayed on the model plots in the following section.
3.5 Modeling
Predictive models were developed from the reaerosolization data for wet- and dry-deposited spores W
and Y. The model for wet-deposited spores was found to have the form
F = f\
The equation for predicting reaerosolization fraction of wet-deposited particles from the surface is
0.862 I
M
e3.46
E 0.750 /i 1.35 /i 0.255 Q0.000292 00.434 0.112ft;0.224
A A2 R, S1 pp NR j
The R2 value for the predictive model for wet spore reaerosolization from urban surfaces is 0.67.
Figure 3-5 shows the fit of the model data versus the measured values. For each point on the plot, the
x-coordinate is the measured reaerosolization fraction and the y-coordinate is the value that would be
predicted for that data point by the model. The resuspended fraction generally falls within two orders of
magnitude of actual measured values for reaerosolization of particles deposited in a wetted state. In
addition, there are no discernible differences in predicted reaerosolization fractions based on laboratory
and particle tested. There were four potential outliers within the Btk data that were not excluded from the
model data but were not found to adversely affect the model statistics. However, two potential outliers in
the Ba data were excluded from the model because they were more than two standard deviations from
the observed range of reaerosolization fractions within the respective data set.
The accuracy of the model indicates that the surface energy, applied shear stress, and surface
characteristics are adequate for prediction of reaerosolization fractions. Characteristics of the particles
that are being removed from the surface are secondary in importance for the prediction, primarily due to
the particle being encapsulated in liquid. The adhesion force of wet particles on the surface is dominated
by the capillary force between the particle and the surface. The surface energy calculated from the
contact angle measurement was able to capture the importance of this phenomenon. If the
reaerosolization behavior for any given wet particle on the test surfaces used varies from this model, it
would then prove beneficial to investigate specific particle characteristics that might have an impact on
adhesion and removal forces from the surfaces. It is important to note that the wet-deposited particles
were observed in singlet, doublet, and triplet formations.
Figure 3-6 shows the fit of the reaerosolization prediction versus the measured values coded by surface
type with no regard to particle type. Higher reaerosolization fractions (per surface type) are a function of
higher applied shear stress on the material, seen in the way that each surface type is grouped into three
different planes on the y-axis. The figure reveals a gap in data between 1E-5 and 1E-4 in the predictions.
Future efforts should include testing materials at additional shear stress values to fill the curve and
improve the predictions.
34
-------
10E+0
1.0E-7
• EPA wet Btk
4 DPG wet Btk
DPG wet Sa
1 OE-7 1 OE-6 1.0E-5 1.0E-4 1 OE-3 1 OE-2 1.0E-1 1 OE+0
Measured Resuspension Fraction (F)
Figure 3-5: Fit of a model to predict reaerosolization of wet particles from various urban surfaces
categorized by laboratory and test particle.
10E+0
1.0E-7 1.0E-6 1.0E-5 1.0E-4 1.0E-3 1.0E-2 1.0E-1 1 .OE+0
» Smooth asphalt
• Rough asphalt
A Very rough asphalt
• Smooth glass
• Rough glass
A Very rough glass
Smooth concrete
Rough concrete
Very rough concrete
Measured Resuspension Fraction (F)
Figure 3-6: Fit of a model to predict reaerosolization of wet particles from various urban surfaces
categorized by surface material.
35
-------
The dry-deposited particle model was found to have a similar form to that described for the wet-deposited
particle reaerosolization model; that is, the dry-deposited particle model is mostly a function of shear
stress and surface energy, with roughness and surface area also having some impact on the prediction.
The model for dry-deposited spores was found to have the form
The equation for predicting the reaerosolization fraction of wet particles from the surface is
0. 604 p 3. 60 o3.37 A
= (a-'
o0.524/i 1.15 4.20Ai8.41_/4.15
"1 M2 Pp /VR °p
The R2 value for the predictive model for dry particle reaerosolization from urban surfaces is 0.41. The
shear stress and surface energy were determined to be contributing factors, but some of the surface
features had little impact on the prediction and were eliminated from the function. Figure 3-7 shows the fit
of the model data versus the measured values. For each point on the plot, the x-coordinate is the
measured reaerosolization fraction and the y-coordinate is the value that would be predicted for that data
point by the model. The resuspended fraction falls within three orders of magnitude of actual measured
values for resuspended particles initially deposited on surfaces in a dry condition. Without regard to
statistical analyses performed on these data, Figure 3-7 shows that there is a much wider spread in
predicted values from the EPA laboratory than there is from DPG. Additionally, there is not a discernible
difference between spore types for DPG data. This is not to say that the surface energy values would be
sufficient for all particle types, simply that Btk and Ba spores have similar surface features.
EPA laboratory values for the reaerosolization fraction are mostly underestimated by the model. Figure
3-8 shows the fit of the reaerosolization prediction versus the measured values coded by surface type
with no regard to spore type. Higher modeled reaerosolization fractions (per surface type) are a function
of higher applied shear stress on the material, seen in the way that each surface type is grouped into
three planes on the y-axis. From this view, the majority of underestimated reaerosolization fraction values
are from concrete experiments performed at EPA. Future testing to provide a larger data set for analysis
could improve the fit of the model and help account for the differences.
A complexity of the dry spore reaerosolization model was determining the surface energy of a dry particle
on the surfaces of interest. For this effort, a rough estimation was used, described in the methods section,
which provided one value for each material, thus there was no differentiation within the classes of glass,
concrete, and asphalt. Resolving the surface energy for dry particle and dry surface interaction proved
difficult, as it is a function of van der Waals and electrostatic forces complicated by surface asperities that
determine contact surface area. Chung et al. (2009) describe a method of measuring adhesion force of
dry Bacillus spores on a planar gold surface via atomic force microscopy. This method could be adapted
to determine adhesion forces, and thus surface energy, of spores on the surfaces of interest. It would be
essential, however, to obtain electrostatic force measurements of the test surfaces because in the
absence of a meniscus, the electrostatic force is prevalent. The addition of the surface interactions
between dry particles and surfaces would be beneficial for improving the model predictions for
reaerosolization in this case.
36
-------
1.0E+0
1.0E-1
i. 1.0E-2
£
c 1 .OE-3
|
g 1.0E-4
0 1.0E-5
1.0E-6
1.0
TT
4t
Ul!
n n i
ii u i
ii
T -if
JrUlL
I
Til I 1
toTJ, & 1
yfFaot1
W27
1
4.
%
*
Jt
1 ,
n
1
t
» EPA dry Bfk
« DPG dry Bf/<
DPG dry Ba
E-6 1.0E-5 1.0E-4 1.0E-3 1.0E-2 1.0E-1 1.0E+0
Measured Resuspension Fraction (F)
Figure 3-7: Fit of a model to predict reaerosolization of dry particles from various urban surfaces
categorized by laboratory and test particle.
1.0E+0
1.0E-1
I
0.
§ 10E-2
£
•§ 1 .OE-3
fc 1 OE-4
1
01
1.0E-5
1.0E-6
1.0
n
A
I -111
n
1 ii 1
H
IfflJ
I
li
il
r
41
i
•*Ti«>
fl 1
5r
I
t-
*T3t
. 1 V
•I
i
i_
» Smooth glass
• Rough glass
A Very rough glass
Smooth concrete
Rough concrete
* Very rough concrete
E-6 1.0E-5 1.0E-4 1.0E-3 1.0E-2 1.0E-1 1.0E+0
Measured Resuspension Fraction (F)
Figure 3-8: Fit of a model to predict reaerosolization of dry particles from various urban surfaces
categorized by surface material.
37
-------
4 Quality Assurance
This project was conducted under the approved Category III quality assurance project plan (QAPP)
Determination of the Difference in Resuspension of Spores off Outdoor Materials, originally QAPP-ZED-
12-02 (Alion, 2012) and updated by QAPP-LB-13-02 (Alion, 2013).
4.1 Equipment Calibration
Before beginning experiments, all of the following monitoring equipment was checked against either a
primary or a secondary standard to ensure that the equipment was operating within acceptance criteria:
• Model HT205 probe (Rotronic Instrument Corp., Hauppauge, NY, USA) was used to monitor and
control temperature and RH inside the test chamber (last calibration 12/30/11).
• Model DA 410 vane anemometer (Pacer Instruments, Keene, NH, USA) was used to measure wind
tunnel velocity. The anemometer is returned to the factory for calibration (last calibration 1/30/12).
• National Institute of Standards and Technology (NIST)-traceable temperature probe was used to
record temperatures in the incubators.
Class A volumetric glassware was used where possible. All equipment used in the project is maintained
and calibrated according to operation manual specifications and/or previous investigations.
4.2 Data Quality Objectives
Precision and accuracy goals were established for each measurement parameter based on (1) scientific
requirements needed to achieve the primary objectives of the study, (2) knowledge of the measurement
system, (3) in-house experience with the sampling and measurement methods, and (4) other similar
research studies. Data quality objectives (DQOs) for each major measurement parameter are listed in
Table 4-1.
Table 4-1. Data Quality Objectives for Experimental Conditions
Parameter (Method)
Velocity uniformity (vane anemometer)
Analytical measurements (plate cultures)
Temperature/RH (capacitance)
Completeness
(%)
90
50
90
Accuracy
(%)
±10
±10
± 5 (temp)
±10(RH)
Precision
(%)
±20
±15
±2
±5
Care has been taken to ensure that samples and measured parameters are representative of the media
and conditions being measured. All data are calculated and reported in units that are consistent with
similar measurements from partnering organizations to allow for comparability of data among
organizations. DQOs for precision and accuracy were based on prior knowledge of the measurement
system employed and method verification studies, which include the use of replicate samples and
duplicate analyses. Definitions of DQOs are given below.
38
-------
Completeness: a measure of the amount of verified data obtained from a measurement system compared
to the amount of data that was expected to be obtained under normal conditions. Completeness is
assessed in this project by reviewing laboratory data logs and laboratory logbooks to ensure that all data
are verified within established objectives.
Accuracy: the degree of agreement of measurements (or an average of measurements) with an accepted
reference or true value. Accuracy is a measure of the bias or systematic error in a system.
Precision: a measure of mutual agreement among individual measurements of the same property, usually
under prescribed similar conditions. Precision is best expressed in terms of the standard deviation.
Various measurements of precision exist depending on the prescribed similar condition.
Representativeness: the degree to which data accurately and precisely represent the characteristics of a
population, process, or environmental condition, or parameter variations at a sampling point.
Representativeness is assessed by the collection of appropriate numbers of samples and the use of a
verified sampling design.
Comparability: the confidence with which one data set can be compared to another. Comparability of
experimental and numerical data is ensured by using standard comparison and reporting methods. All
data are presented in specified and documented units. Comparability also is ensured by the use of
approved SOPs for all instrumentation.
4.3 QA/QC Checks
The quality assurance/quality control (QA/QC) checks that are performed are provided in Table 4-2. Each
time spores are deposited onto a batch of test coupons, two positive controls and one negative control
are performed. The sample for the first positive control is pulled directly from the spore stock solution. The
second positive control sample (called the reference tube) is collected directly from the deposition
chamber ultrasonic nozzle and plated to ensure that the ultrasonic nozzle is not damaging the spores.
Each test coupon was wipe sampled after the experiment was complete as a qualitative check that
spores were deposited on the coupon surface. A negative control (blank) of the phosphate buffer that is
being used is prepared by plating onto an agar plate and spreading evenly using a sterile spreader. In
addition, a wind tunnel blank or background test is run after every third or fourth test coupon experiment
to provide a measurement of the contamination level of the RWT. The background sample is run and
collected onto filters just as the test runs with a clean smooth glass coupon in place. A field blank filter is
also transferred to a Stomacher bag in the laboratory where filters are loaded into the filter holders as a
negative control.
39
-------
Table 4-2. QA/QC Checks
Measurement
Temperature/
RH
Pressure drop
Microbiological
Microbiological
Temperature
Laboratory
contamination
Matrix
RWT
RWT
Spore
inoculant
PBS* without
microbes
Incubator air
Filter
QA/QC Check
Chamber
temperature probe
Pressure gauge
Positive control
Negative control
NIST-traceable
thermometer
Field blank
Frequency
Continuous
Continuous
2/run
1/run
Continuous
1/day
Acceptance
Criteria
±10%
±5%
Growth
No growth
±5%
No growth
Corrective Action
Stop sampling and
correct as necessary
Check filter holders,
RWT seal
Investigate viability of
microbes
Clean and disinfect
systems
Repair incubator
Dispose of questionable
samples
Repeat sampler test
Clean and disinfect
laboratory
*Phosphate buffered saline
4.4 Acceptance Criteria
The following acceptance criteria tables detail the experimental goals, sources of error, and means of
minimizing error. Tables 4-3 and 4-4 list the established acceptance criteria for the surrogate spores and
their deposition, respectively. The acceptance criteria for the reaerosolization tests are given in Table 4-5.
The DQO is less than a 100 % standard error in the measurement of reaerosolization efficiency (amount
reaerosolized/amount deposited on surface). Table 4-6 presents the acceptable variation among replicate
experiments.
Table 4-3. Spore Acceptance Criteria
Goal: Btkand Bg have physical and chemical characteristics similar to Ba-Ames
Controllable Error
Uncontrollable Error
Error Source
Spore size
Spore shape
Spore purity
Minimization Strategy
Compare SEM images with reference images
Compare SEM images with reference images
Confirm endotoxin is absent from Btk and Bg
Spore coating chemistry
Spore surface morphology
40
-------
Table 4-4. Deposition Acceptance Criteria
Goal: Minimize loading variability across test coupon
Controllable Error
Uncontrollable Error
Error Source
Deposition chamber operation
(e.g., duration, flow, pressure)
Spore concentration (CFU/mL,
CPU/mass)
Extraction efficiency from filter or
wipe
Spore plating efficiency
Spore loss between deposition
and reaerosolization experiment
Spore agglomeration as liquid
drops or puddles dry, or dry
powder is dispersed
Minimization Strategy
Make sure operation is consistent
between deposition events
Make sure sample preparation and
extraction from bulk material are
consistent
Make sure efficiency is consistent (most
important); high extraction efficiency is
desired
Make sure plating procedure is
consistent
Handle and transport coupons in
covered tray
Limit liquid application to avoid puddle
formation
Minimize drop size to yield one spore per
drop
Make sure powder dispersal breaks dry
agglomerates into discrete particles
Spore agglomeration due to random deposition pattern
Spore migration into surface pores
Goal: Generate monolayer of individual spores or consistently sized aggregates across test
surface
Controllable Error
Uncontrollable Error
Error Source
Deposition chamber operation
(e.g., duration, flow, pressure)
Spore concentration (CFU/mL,
CPU/mass)
Spore agglomeration as liquid
drops or puddles dry, or powder is
dispersed
Minimization Strategy
Make sure operation is consistent
between deposition events
Make sure sample preparation and
extraction from bulk material are
consistent
Limit liquid application to avoid puddle
formation
Minimize drop size to yield one particle
per drop
Make sure powder dispersal breaks dry
agglomerates into discrete particles
Particle agglomeration due to random deposition pattern
Particle migration into surface pores
Agglomeration of particles in bulk material/container
41
-------
Table 4-5. Reaerosolization Acceptance Criteria
Goal: Minimal variability among test surface coupon characteristics
Controllable Error
Uncontrollable Error
Error Source
Surface chemistry analytical error
and surface variability
Surface porosity analytical error and
surface variability
Surface hygroscopicity analytical
error and surface variability
Contact angle analytical error and
surface variability
Surface roughness analytical error
and surface variability
Minimization Strategy
Characterize chemistry of representative
sample of all surfaces by X-ray
fluorescence
Characterize porosity of representative
sample of all surfaces with BET
Characterize water absorption capacity
of surface by TGA
Characterize a representative number of
all surfaces with a goniometer
Characterize roughness height and
asperity frequency of a representative
sample of all surfaces by profilometry
None but variability could be large and unable to be minimized
Table 4-6. Acceptable Variation among Replicate Experiments
Goal: Minimal variance in reaerosolization among replicate experiments at same conditions
Controllable Error
Uncontrollable Error
Error Source
Wind tunnel operation (carrier air
speed, jet speed, test duration, etc.)
Particle losses within wind tunnel
Filter collection efficiency
Particle extraction efficiency from
filter media
Spore plating variability
Particle loss between reaerosoliza-
tion experiment and sample analysis
Minimization Strategy
Make sure operation is consistent within
and among tests
Ensure proper design to maximize
fraction transported to sample filters
Make sure filtration efficiency is greater
than 99 % for particles larger than 1 urn
Make sure efficiency (most important) is
consistent; high extraction efficiency is
desired
Use consistent plating procedure
Handle and transport coupons in
covered tray
Random variation in test surface characteristics
Particle migration into surface pores
Particle and surface chemistry at point of contact
Evaporation or formation of water meniscus between particle and surface
42
-------
5 Summary
This project report describes the research effort to characterize the reaerosolization of Ba surrogate
spores from outdoor surfaces under different environmental conditions and compare the data with data
from identical experiments carried out using Ba-Ames. A large part of the effort was put into design,
construction, and validation of equipment and development and validation of methods. Reaerosolization
tests were conducted with the biological agent Ba-Ames and surrogates Btk and Bg. Testing conducted at
both laboratories with the bar-coded Btk spores showed no evidence of differences in the results between
the two laboratories for wet-deposited or dry-deposited spores, validating the equivalence of experimental
equipment and methods at EPA and DPG.
The experimental data showed that there were no statistically significant differences between Btk and Ba-
Ames spores for either wet or dry deposition, while Bg spores yielded results 79 % lower than the
average for Btk and Ba-Ames. Wet-deposited Btk and Ba-Ames spores reaerosolized 81 % less than dry-
deposited Btk and Ba-Ames spores, and wet-deposited Bg spores reaerosolized 29 % less than dry-
deposited Bg spores. Other significant findings of the study were concrete consistently showed higher
reaerosolization than asphalt and glass, the roughness level within each material type did not have a
significant effect on reaerosolization, and high jet velocity consistently reaerosolized more particles than
the low and medium jet velocities.
A predictive model was developed for determining particle reaerosolization fractions of wet- and dry-
deposited spores from manmade outdoor surfaces. The predictive model was built on experimentally
determined reaerosolization data. The model incorporated multiple physical measurements that helped to
characterize the surfaces on which particles were deposited.
43
-------
6 References
Abramovich, G.N. 1963. The Theory of Turbulent Jets, L.H. Schindel, ed. The MIT Press, Massachusetts
Institute of Technology, Cambridge, MA.
Akansu, Y.E., Sarioglu, M., Kuwet, K., and Yavuz, T. 2008. Flow field and heat transfer characteristics in
an oblique slot jet impinging on a flat plate. International Communications in Heat and Mass Transfer
35:873-880.
Alion Science and Technology (Alion). 2012. Determination of the Difference in Resuspension of Spores
off Outdoor Materials, QAPP-ZED-12-02. Durham, NC. Available upon request from the corresponding
author.
Alion Science and Technology (Alion). 2013. Determination of the Difference in Resuspension of Spores
off Outdoor Materials, QAPP-LB-13-02. Durham, NC. Available upon request from the corresponding
author.
Alion Science and Technology (Alion). 2014a. Coupon Loading Procedure -Wet Deposition, ATF
MOP 1, Revision 0. Durham, NC. Available upon request from project contact.
Alion Science and Technology (Alion). 2014b. Coupon Loading Procedure - Spray-Dry Deposition, ATF
MOP 2, Revision 0. Durham, NC. Available upon request from project contact.
Alion Science and Technology (Alion). 2014c. Resuspension Wind Tunnel Testing for SPORE, ATF
MOP 3, Revision 0. Durham, NC. Available upon request from project contact.
Alion Science and Technology (Alion). 2014d. Filter Holder Loading and Unloading, ATF MOP 4,
Revision 0. Durham, NC. Available upon request from project contact.
ARCADIS U.S. 2009. Serial Dilution: Spread Plate Procedure to Quantify Viable Bacterial Spores, MOP
6535a, Revision 3.0. Durham, NC. Available upon request from the corresponding author.
ARCADIS U.S. 2010. Recovery of Bacillus Spores from Wipe Samples, MOP 6567, Revision 0. Durham,
NC. Available upon request from the corresponding author.
ARCADIS U.S. 2013a. Recovery of Bacillus Spores from Stainless Steel Coupon Samples, MOP 6601,
Revision 0. Durham, NC. Available upon request from the corresponding author.
ARCADIS U.S. 2013b. Recovery of Bacillus Spores from Polyester Felt Filters, MOP 6600, Revision 0.
Durham, NC. Available upon request from the corresponding author.
Beitelmal, A.M., and Saad, M. A. 2000. Effects of surface roughness on the average heat transfer of an
impinging air jet. International Communications in Heat and Mass Transfer 27:1-12.
Buckingham, E. 1914. On physically similar systems; illustrations of the use of dimensional equations.
Physical Rev;ew4(4):345-376.
Carrera, M., Zandomeni, R.O., Fitzgibbon, J., and Sagripanti, J.L. 2007. Difference between the spore
sizes of Bacillus anthracis and other Bacillus species. Journal of Applied Microbiology 102(2):303-312.
Chin, D.T., and Agarwal, M. 1991. Mass transfer for an oblique impinging slot jet. Journal of the
Electrochemical Society 138:2643-2650.
Chinga, G., and Dougherty, R. 2006. Roughness Calculation,
http://rsb.info.nih.gov/ij/plugins/roughness.html, last accessed March 29, 2014.
44
-------
Chung, E., Kweon, H., Yiacoumi, S., Lee, I., Joy, D.C., Palumbo, A.V., and Tsouris, C. 2009. Adhesion of
spores of Bacillus thuringiensis on a planar surface. Environmental Science and Technology 44(1): 290-
296.
Environmental Protection Agency (EPA). 2012. On the Use of Bacillus thuringiensis as a Surrogate for
Bacillus anthracis in Aerosol Research. Technical Report, EPA/600/R-12/596. Research Triangle Park,
NC.
Glauert, M.B. 1956. The wall jet. Journal of Fluid Mechanics 1:625-643.
Greenberg, D.L., Busch, J.D., Keim, P., and Wagner, D.M. 2010. Identifying experimental surrogates for
Bacillus anthracis spores: a review. Investigative Genetics 1(1):4.
Girifalco, L.A., and Good, R.J. 1957. A theory for the estimation of surface and interfacial energies. I.
Derivation and application to interfacial tension. Journal of Physical Chemistry 61(7):904-909.
Hogg, A.J., Huppert, H.E., and Dade, W.B. 1997. Erosion by planar turbulent wall jets. Journal of Fluid
Mechanics 338:317-340.
Hollander, M., and Wolfe, D.A. 1973. Nonparametric Statistical Methods. John Wiley and Sons, New
York, NY, 503 pages.
ImageJ. 2011. SurfCharJ plug-in, Version 1c, http://www.gcsca.net/IJ/SurfCharJ.html, last accessed
March 29, 2014.
Lomboy, G., Sundararajan, S., Wang, K., and Subramaniam, S. 2011. A test method for determining
adhesion forces and Hamaker constants of cementitious materials using atomic force microscopy.
Cement and Concrete Research 41:1157-1166.
Loureiro, J.B.R., and Freire, A.P.S. 2009. Flow Features of Confined Jet Impinging onto Rough Walls.
Proceedings of COBEM, 20th International Congress of Mechanical Engineering, November 5-9,
Gramado, RS.
Myers, G.E., Schauer, J.J., and Eustis, R.H. 1961. The Plane Turbulent Wall Jet, Part I—Jet
Development and Friction Factor. Technical Report No. 1, Department of Mechanical Engineering,
Stanford University, Stanford, CA.
Rabinovich, Y.I., Adler, J.J., Esayanur, M.S., Ata, A., Singh, R.K., and Moudgil, B.M. 2002. Capillary
forces between surfaces with nanoscale roughness. Advances in Colloid and Interface Science 96:213-
230.
Rajaratnam, N. 1967. Plane turbulent wall jets on rough boundaries. Water Power 149-153.
Schwartz, W.H., and Cosart, W.P. 1961. The two dimensional turbulent wall jet. Journal of Fluid
Mechanics 10:481-495.
Tomas, J. 2006. Mechanics of Particle Adhesion. Department of Process Engineering and Systems
Engineering, Otto-von-Guericke-University, Magdeburg, Germany, http://www.mvt.ovgu.de/mvt_media/
Vorlesungen/Lecture_SFPS/Mechanics_Particle_Adhesion_full.pdf., last accessed June 2, 2014.
Wikipedia: The Free Encyclopedia. 2014. Fused quartz. Wikimedia Foundation, Inc. http://
en.wikipedia.org/wiki/Fused_quartz, last accessed June 1, 2014.
45
-------
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
------- |