|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
CONSISTENCY AND REPRODUCIBILIT Y OF BIOAEROSOL DELIVERY FOR INFECTIVITY STUDIES ON MICE By BRENTON R. STONE A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORID A IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2010 1
2010 Brenton R. Stone 2
To my family and friends 3
ACKNOWLEDGMENTS First and foremost, I thank Professor ChangYu Wu for accepting me as a student and accommodating my unusual challenges as a distance student. I also thank Dr. Joseph Wander for bringing me onto this project and helping me to enter the graduate program that this thesis partially fulfills Professor Myoseon Jang chose to round out my thesis committee, joining Dr. Wander and Pr ofessor Wu, and I thank her for her time and effort. My employer, Applied Research Associ ates Inc., generously provided monetary support for my graduate coursework. I would like to thank Brian Heimbuch, Kimberly Kinney, April Lumley, and Rashelle McDonal d of Applied Research Associates for providing laboratory su pport on some of the experiments that appear in this thesis, and Bob Nichols (also of Applied Research Associates) for supporting the design and construction phase of the project. The data from this thesis appear as part of the US Air Force Research Laboratory technical report AFRL-RX-TY-TR-2009-4593. Funding from US Air Force Project DODT0049 is gratefu lly acknowledged. 4
TABLE OF CONTENTS page ACKNOWLEDG MENTS .................................................................................................. 4LIST OF TABLES ............................................................................................................ 7LIST OF FI GURES .......................................................................................................... 9LIST OF ABBR EVIATION S ........................................................................................... 10ABSTRACT ................................................................................................................... 13 CHAPTER 1 INTRODUC TION .................................................................................................... 15Motivati on ............................................................................................................... 15Bioaerosols ....................................................................................................... 15Creating Bioaer osols ........................................................................................ 16Measuring Bioa erosols ..................................................................................... 17Filtration ............................................................................................................ 19Antimicrobials for Aero sol Filtra tion .................................................................. 23The Antimicrobial Poly(styrene-4-[tri methylammonium]met hyl triiodide) .......... 25Review of Animal Inhala tion Exposure System s ..................................................... 27Factors Influencing an Animal I nhalation Exposure System ................................... 30Objectiv e ................................................................................................................. 342 MATERIALS AND METHODOL OGY ...................................................................... 35Materials ................................................................................................................. 35Controlled Aerosol Test System ....................................................................... 35Sampling Instru mentati on ................................................................................. 39All-glass im pingers ..................................................................................... 39Particle sizers ............................................................................................. 39Challenge Micr oorganism s ............................................................................... 40MS2 coli p hage .......................................................................................... 40Bacillus atrophaeus .................................................................................... 41Filter Media ....................................................................................................... 42Safe Life T-5000 ........................................................................................ 423M 1860S................................................................................................... 43Isopropyl alcohol-t reated 1860S ................................................................ 43Methods .................................................................................................................. 43Leak Chec k ...................................................................................................... 43Flow Rate, Relative Humidity, and Temperature Consistency .......................... 44Correlation of Sampling Ports ........................................................................... 44Bioaerosol Consis tency Tria ls .......................................................................... 45 5
3 RESULT S ............................................................................................................... 50Leak Check and Flow Rate, Relative Humi dity, and Temperature Consistency ..... 50Correlation of Sa mpling Po rts ................................................................................. 50Bioaerosol Consistenc y Trials wit h MS2 ................................................................. 50Bioaerosol Consis tency Trials with B. atrophaeus .................................................. 514 DISCUSSI ON ......................................................................................................... 57Flow Rate, Relative Humidity, and Temperature C onsisten cy ................................ 57Correlation of Sa mpling Po rts ................................................................................. 57Bioaerosol Consis tency Tria ls ................................................................................. 58Particle Size Distribution ................................................................................... 58Viabilit y ............................................................................................................. 59MS2 ........................................................................................................... 59B. atrophaeus ............................................................................................. 60Filter Physical Remo val Efficiency .................................................................... 62Extrapolating to the Delivered Do se ....................................................................... 645 CONCLUSION ........................................................................................................ 66 APPENDIX A OPERATING SEQUENCES ................................................................................... 68Leak Chec k ............................................................................................................. 68Pre-Nebulization Pr eparati ons ................................................................................ 68Aerosol Consist ency Tria ls ..................................................................................... 69B FLOW RATE, TEMPERATURE, A ND RELATIVE HUMIDI TY DATA ..................... 72C PORT CORRELAT ION DATA ................................................................................ 73D BIOAEROSOL CONSIS TENCY RAW DATA .......................................................... 75MS2 ........................................................................................................................ 75B. atrophaeus ......................................................................................................... 76LIST OF RE FERENCES ............................................................................................... 80BIOGRAPHICAL SKETCH ............................................................................................ 89 6
LIST OF TABLES Table page 3-1 Mean relative humidity (RH), temper ature, and particle size distribution (PSD) moments, and coefficients of variation (CVs) of PSD moments for MS2 experiments ................................................................................................ 553-2 Filter used, mean temperat ure, RH, and PSD moments for Bacillus atrophaeus experimen ts ..................................................................................... 563-3 Viable concentrations and CVs of PSD moments and upstream airborne viable concentration for B. atrophaeus experim ents ........................................... 56B-1 Temperature, RH, and flow consist ency dat a ..................................................... 72C-1 Readings of particle concentration at ports on Controlled Aerosol Test System (CATS) while nebulizing 250-nm beads ................................................. 73C-2 Readings of particle concentration at ports on CATS wh ile nebulizing 1-m beads .................................................................................................................. 74D-1 PSD data for Exper iment 724 (M S2) .................................................................. 75D-2 PSD data for Exper iment 728 (M S2) .................................................................. 76D-3 PSD data for Exper iment 730 (M S2) .................................................................. 76D-4 PSD data for Exper iment 811 (M S2) .................................................................. 76D-5 PSD data for Exper iment 812 (M S2) .................................................................. 76D-6 PSD data for Exper iment 813 (M S2) .................................................................. 76D-7 PSD data for Experiment 819 ( B. atrophaeus ) ................................................... 76D-8 Viability data for Experiment 819 ( B. atrophaeus ) .............................................. 77D-9 PSD data for Experiment 820 ( B. atrophaeus ) ................................................... 77D-10 Viability data for Experiment 820 ( B. atrophaeus ) .............................................. 77D-11 PSD data for Experiment 827 ( B. atrophaeus ) ................................................... 77D-12 PSD data for Experiment 901 ( B. atrophaeus ) ................................................... 78D-13 Viability data for Experiment 901 ( B. atrophaeus ) .............................................. 78D-14 PSD data for Experiment 903 ( B. atrophaeus ) ................................................... 78 7
D-15 PSD data for Experiment 908 ( B. atrophaeus ) ................................................... 78D-16 Viability data for Experiment 908 ( B. atrophaeus ) .............................................. 79D-17 PSD data for Experiment 909 ( B. atrophaeus ) ................................................... 79D-18 PSD data for Experiment 910 ( B. atrophaeus ) ................................................... 79D-19 Viability data for Experiment 910 ( B. atrophaeus ) .............................................. 79 8
LIST OF FIGURES Figure page 2-1 Photograph of Controlled Aerosol Te st System (CATS), with key components labele d ................................................................................................................ 482-2 Process-flow diagram of CATS. .......................................................................... 493-1 Representative particle size distri bution (PSD) from MS2 nebulization and 95% confidence intervals for each individual diameter ....................................... 533-2 Particle removal efficiency (PRE) of T-5000 medium as a func tion of particle size and 95% confiden ce interv als ..................................................................... 543-3 Representative PSD from Bacillus atrophaeus nebulization and 95% confidence in tervals ............................................................................................ 543-4 Downstream measurements from Experiment 910 with isopropyl alcohol (IPA)-treated 1860S medium and 95% confidence in tervals .............................. 543-5 PRE of IPA-treated 1860S medium as a function of particle size and 95% confidence in tervals ............................................................................................ 55C-1 Representative PSD from nebulizing 250nm beads .......................................... 73C-2 Representative PSDs from nebulizing 1m beads ............................................ 74 9
LIST OF ABBREVIATIONS AGI All-Glass Impinger APS Aerodynamic particle sizer ATCC American Type Culture Collection BSA Bovine serum albumin BSL Biosafety level C Concentration (typically of particles, with units #/m3, or of viable microorganisms, with units PFU/m3 or CFU/m3) Cdown Concentration upstream downstream of a filter Cup Concentration upstream of a filter cm Centimeter CATS Controlled Aerosol Test System CFU Colony-forming unit CMD Count median diameter CV Coefficient of variation DI Deionized EPA (United States) Envir onmental Protection Agency p Pressure drop F Fractional deposition FFR Filtering facepiece respirator GSD Geometric standard deviation HEPA High efficiency particulate air HVAC Heating, ventilati on, and air conditioning in H2O Inches of water pressure IPA Isopropyl alcohol (2-propanol) L Liter 10
MERV Minimum Efficiency Reporting Value MID50 Median infective dose (Minimum infective dose for 50% of the population) mL Milliliter, cubic centimeter mm Millimeter MPPS Most-penetrating particle size m Micrometer N Count of PFUs or CFUs n Dilution factor N95 A FFR with a PRE of 95% or gr eater for 300-nm salt particles ND No data NIOSH (United States) National In stitute of Safety and Health NIST (United States) National In stitute of Standards and Technology nm Nanometer P95 An oil-resistant FFR with a PRE of 95% or greater for 300-nm salt particles PBS Phosphate-buffered saline PFU Plaque-forming unit PRE Physical removal efficiency PSD Particle size distribution psi Pounds per square inch pressure psig Pounds per square inch pr essure over gauge pressure PSL Polystyrene latex PSTI poly(styrene-4-[trimethyl ammonium]methyl triiodide) Q Flow rate Qa Flow rate of aerosol collected 11
R2 Coefficient of determination for a linear regression RH Relative humidity SARS Severe acute respiratory syndrome SMPS Scanning mobility particle sizer St. dev. Standard deviation T Temperature t Time; duration of exposu re; duration of sample TPC Total particle count TSA Tryptic Soy Agar TSB Tryptic Soy Broth VEE Venezuelan equine encephalomyelitis Vi Volume of liquid in impinger Vm Minute volume of an animal (units mL/min) Vp Volume of liquid plated VRE Viable removal efficiency VSF Viable spray factor 12
Abstract of Thesis Pres ented to the Graduate School of the University of Florida in Partial Fulf illment of the Requirements for t he Degree of Master of Science CONSISTENCY AND REPRODUCIBILIT Y OF BIOAEROSOL DELIVERY FOR INFECTIVITY STUDIES ON MICE By Brenton R. Stone May 2010 Chair: Chang-Yu Wu Major: Environmental Engineering Sciences Questions about the clinical significance of an antimicrobial resin used on personal respirators led to the need for a system to generate consistent test bioaerosols for use in animal studies. The hypothesis was propos ed that an aerosol delivery system based on the Collison nebulizer can be designed an d engineered to provide, at selectable concentrations, a respiratory challenge of bi oaerosol particles that is verifiably consistent in time and that can be fed in separate experiments through treated and untreated control filters to de liver a consistent challenge to a small-animal model of human respiration. To verify this hypothesis, such an expe rimental filtration system was designed and built. Challenge trials were performed with MS2 bacteriophage and Bacillus atrophaeus Over 30 to 40 minutes, the particle size di stribution (PSD) was measured, and viability of microorganisms collected in All-Gl ass Impingers (AGI-4s) was determined. Concentrations of particles and microorganisms dow nstream of the filter were too low to measure, and the viable counts for MS2 bacteriophage were not measured at all owing to problems with the assay method. However, in each experiment, the coefficients of variations (CVs) of time-series measurements of the total particle count, count median 13
14 diameter, and geometric standard deviation of the upstream PSD were less than 10%. From five B. atrophaeus experiments with viability dat a, all CVs of time-series measurements of upstream viabl e airborne concentration were less than 26%. This CV is somewhat higher than has been reported in the literature for tests with airborne Bacillus spores, but the plati ng method used to measure the viability may have introduced additional variation that was not caused within the system itself. It can be reasoned based on this data that the system can pr ovide a sufficiently steady aerosol challenge to be used for later studies using a small-animal model of human respiration. The system provides a design for an animal exposure system incorporating aerosol filtration, which is a capability prev iously unreported in the literature.
CHAPTER 1 INTRODUCTION Motivation Bioaerosols An aerosol is any sort of finely divided ma terial suspended in a gas. Bioaerosols are aerosols made up of particl es of biological origin. Included in the category of bioaerosols are airborne viruses, bacteria, fu ngi, pollen, and organic material produced by biological processes (such as dust mite waste, a common household allergen).1 Bioaerosols are known to be a transmission mechanism for many disease-causing organisms, including Legionella smallpox, severe acute re spiratory syndrome (SARS) coronavirus, and rhinovirus.2 Whether bioaerosols are an important natural transmission mechanism for influenza is hotly debated.30 Besides the huge impact of naturally occurring infectious disease in humanity, th ere is concern that infectious organisms could be weaponized in a bioaerosol form and used for biological warf are or terrorism. A great deal of research has gone into examin ing the weapon potential of bioaerosols and how to defend against potential threats. A distinction is drawn bet ween bioaerosols that are viable, or capable of being cultured, and those that are nonviable. A non-viable organism cannot infect a host. The viability of bioaerosols can change depending on a number of fact ors, including relative humidity (RH), temperature, oxygen content, airborne ions, radiation, and open air factors (a set of ambiguous in fluences that cause faster i nactivation in outdoor air than in clean laboratory air).11 Different organisms react different ly to stresses in the aerosol state: the influenza virus is strongly affected by humidity,12 and the stability of the bacterium Escherichia coli depends on RH, temperatur e, oxygen content, and the 15
aerosol generating met hod, while the bacteria Bacillus subtilis does not show a strong dependence on RH, temperatur e, or oxygen content.11 For some microorganisms, an unfavorable humidity can decrease the airbor ne viable concentration by several orders of magnitude.13 For viruses, the presence or absence of a lipid coating on the virions a quality that varies by strain alters whether the virus is more stable at high or low RH.11 Among bacteria, the airborne viability and its dependence on RH and oxygen content differs greatly between Gramnegative and Gram-positive species.13 Bioaerosol particles vary in size dependi ng on what microorganisms they contain. Individual virions tend to be 0.02 to 0.3 m in physical diameter, bacteria 0.3 to 10 m, fungal spores 0.5 to 30 m, and pollen 10 to 100 m. However, the individual particles may agglomerate into larger ones, or combi ne with non-biological airborne particles, increasing their dimensions and changing their behavior in the aerosol state.14 Creating Bioaerosols Artificial bioaerosols are gener ated in the laboratory to simulate naturally occurring bioaerosols, for example to simulate a cough or an intentional release of an infectious agent as a biowarfare agent. The Collison nebulizer is often used to create a bioaerosol from a fluid containing microor ganisms. In the Collison, a pr essurized (typically 25 to 30 psig) stream of air draws up a liquid by the Venturi effect and jets it as a stream of droplets against the wall of it s container. Of these droplets, those that are small enough are swept out of the nebulizer in the aerosol state; the rest re turn to the liquid reservoir.15,16 Because Collison nebulizers are re circulating systems and impose large shear forces, microorganisms in suspens ion accumulate metabolic damage as a Collison continues to operate, and may lose viability.17 The rate of loss of viability is typically not rapid enough to prevent an experiment from being pe rformed. A quantity 16
used in evaluating the effectiveness of deliver ing a viable bioaerosol with nebulization is the viable spray factor (VSF). The VSF is def ined as the ratio of the concentration in the aerosol state produced by the nebulizer to that in the nebulization liquid. On average, VSFs are of magnitude 10-7, depending on the hardiness of the particular strain.18 Other methods of producing bioaerosols in clude other modes of atomization, such as ultrasonic nozzles that use high-frequency vibrations to produce an aerosol, and electrostatic nebulizers that use electrical forces. Dry powder dispersion techniques are used to produce an aerosol from a powder sour ce, such as dry bacterial spores, and powder scrapers are also used for fungal bioaerosols.16 These methods all have advantages, but are more co mplicated than a Collison, which has no moving parts. Measuring Bioaerosols Bioaerosols can, like any other aerosol, be measured with a particle sizer. Particle sizers measure the distribution of particles in a sample as a function of aerodynamic diameter. (The aerodynamic diam eter of a particle is the equivalent diameter of a spherical particle of density 1 g/mL that has the same aerodynamic behavior, and is sometimes casually referred to as the size.) From this particle si ze distribution (PSD), the particles in a certain size range can be counted, or moment s of the PSD can be measured. The distribution c an be measured in terms of particle count concentration, particle mass concentration, or a number of other ways. Aerosols in particular aerosols produced by nebulizing a liquid or nebul izing a solution of dissolved solids and then drying the produced droplet s often follow a log-normal size distribution, and in those cases can be defined by a total parti cle concentration (TPC) (if mass-based, a total mass concentration), an average (for in stance, count median diameter (CMD), mass mean diameter, etc.), and a geometric standard deviati on (GSD) quantifying the 17
spread of the distribution.19 If the nebulization liquid contains particles (such as microorganisms) larger than the mode for dissolved solids in a concentration sufficiently small that they do not agglomer ate, the particle sizer detects them as a sharp peak near the diameter of the individual particles.19 A number of particle sizers can be purchased, each model relying on different operating principles and having different capabilities. Researchers also want to collect bioaerosol s for later analysis. In practice, one of the most common methods is to collect bi oaerosols in impingers. Impingers jet the bioaerosol into a liquid medium that traps a size-dependent fracti on of the particles. Other methods include impacting onto a medium and collecting onto filters that can be weighed to measure the mass of aerosol or dissolved to recover the sample.20 No collection method is a perfect collector, and a ll collection methods impose some stress on the bioaerosol and cause a fraction of the collected microorganisms to become nonviable. For instance, Hogan et al .21 measured the viable collecti on efficiency of all-glass impingers (AGIs) for particles of 30 to 100 nm diameter as being below 10%, increasing for larger and smaller particles. Grinshpun et al.22 measured the bounce and reaerosolization from impingers: a signific ant quantity of particles escape from the impinger. Viable collection efficiency of impingers can depend on sampling time and RH.23 The longer microorganisms are held in the collection medium before incubation, the more viability may be lost.24 The efficiency and loss in a sampler may even depend on individual strains.25 The degree of loss is difficult to quantify. However, animal experiments often compare dose-dependent responses to deduce the relative reduction in the dose. The values of these parameters do not need to be exactly known as much as they need to remain constant for the duration of timed-expos ure experiments. 18
Microorganisms collected by these methods can be cultured to measure viability. Viable bacteria can be quantified by perfo rming a plate assay and counting colonyforming units (CFUs); viruses, by performing a plate assay in a host microorganism and counting plaque-forming units (PFUs). Other methods exist to quantify bioaerosols. Assays can be performed for endotoxins specific to a organism, which is useful when the endotoxin causes disease.26 A polymerase chain reaction assay can be performed to measure the amount of Dor RNA charac teristic to the organism that appears in a sample: real-time versions of this me thod are in development specifically for bioaerosols.27 Neither of these methods is sensitive to viability. Filtration Filtration by flowing an aerosol through a fibrous porous medium is a well-known and accepted method used in respiratory protection systems to remove unwanted particles, such as infectious bioaerosols, from breathing air. Filters require lower pressure drops ( p ) than other particle control syst ems and are efficient for a wide range of particle challenges, including very small particles and low particle concentrations. Filters are also relatively simple to use.28 Other particle control systems, such as cyclones, electrostatic precipitators, or wet scrubbers, are typically large pieces of machinery that require industrial blow ers, or require a large degree of upkeep.28 In non-industrial situations, like a heating, vent ilation, and air condit ioning (HVAC) system in the home or office, or a personal resp irator, filtration is a reasonable and common choice. The process of filtrati on relies on five basic mechanisms:16 Gravitational settling Gravity draws the particle o ff its streamline and onto a surface, where it is captured. In filters, this mechanism is unimportant compared to 19
others unless the face velocity through the filter is extr emely low or particles are extremely large. Interception A particle travelling along a stream line makes tangential contact with a filter fiber and is captured. The parti cles must have large enough dimensions to contact the filter fiber while remaining on the streamline. Inertial impaction Because of its momentum, a particle deviates from a streamline, makes contact with a filter fiber, and is capt ured. This is the dominant mechanism for particles with large inertia. Diffusion A particle travelling along a stream line experiences random deviations from its path due to Brownian mo tion. If these deviations caus e the particle to contact a filter fiber, it is captured by diffusion. This is the dominant mechanism for small particles. Electrostatic attraction The electrostatic force on par ticles from charged filter fibers can move a particle off a streamline on to a surface. Many common filters use an electrically charged (electret) filter medium. The contribution of electrostatic attraction to filtration efficiency can be very large, but quantifying it require s knowing the charge on the particles and filter materi al at a microscopic level, wh ich is difficult to measure. The efficiency of a filter is typically quantified as the physical removal efficiency (PRE), which is the fraction of aerosol removed by the filter relative to the feed material. PRE is calculated as Equation 1-1, where Cup is the concentration of particles upstream and Cdown is the concentration downstream. This e fficiency can be at a specific size or for a range of particles. One can also quantif y the viable removal efficiency (VRE), the fraction of viable particles remo ved by the filter relative to the feed. VRE is calculated the same as PRE, except with concentrations of viable microorganism s. For filters with 20
no special antimicrobial capability the VRE is cl ose to the PRE at the particle size of the microorganism. All filters have a range of par ticle aerodynamic diameters for which the dominant process transitions fr om diffusion to impaction, and this range is the window of dimensions in which a given f ilter medium captures least efficiently. The PRE decreases with increasing face velocity, as does the most-penetrating particle size (MPPS).29 (1-1) Efficiency can be increased by increasing the thickness of the layers of the material, but p across the media increases proportionally. The PRE and p of a mechanical filter increase as more particles are l oaded onto the filter.28 However, excessive loading can damage a filter and reduce its PRE, and certain particles, such as dioctyl phthalate and NaCl, can reduce the electric charge on electret filters and thereby decrease their PRE. Barret and Rousseau30 showed that this reduction in PRE can differ between filter media made from the same substance with different fiber production methods. In their article, some f ilters made from polypropylene show minimal change in p while their PRE reduces, and some show a large increase in p High Efficiency Particulate Air (HEPA) f ilters are used in building ventilation systems where biological isolation is desir ed. HEPA filters are defined to be 99.97% efficient at filtering 300-nm particles at a specified face velocity. Heimbuch et al .31,32 showed that during challenges with practically attainable particle concentrations, biological pathogens with aerodynamic diameters in the range of 100 to 300 nm penetrate HEPA filters at t he predicted fraction of 0.03%. The MS2 coli phage used in their test is not infective in humans, but a number of infect ious microorganisms such as Francisella tularensis whooping cough, SARS coronavirus, Venezuelan equine 21
encephalomyelitis, and the influenza virus could allegedly form particles in the 100to 300-nm range by accumulating salts and organi c matter on their surface. The median infectious dose (MID50) for F tularensis is 10 to 50 organisms, and while the MID50s of the viruses are not known exactly, many are believed to be fewer than 10 organisms, possibly as few as a single organism for SARS.33,34 These MID50s are low enough that they could be surpassed by par ticles penetrating a HEPA filter.32 Filtration is also used for individual respiratory protection. A common piece of personal protective equipment for use with hazar dous aerosols is the filtering facepiece respirator (FFR), a filter that covers a persons nose and mouth. The US National Institute for Occupational Safety and Health (NIOSH) approves FFRs in classes by PRE classes 95, 99, and 100 denote f ilter media with at least 95% 99%, or 99.97% PRE, respectively, at the most-penetrating parti cle size (MPPS) when tested with a NaCl aerosol and by oil resistance: N, R, and P for non-resistant, somewhat resistant, and strongly resistant, respectively.35 N95 and P95 respirators ar e the most commonly used, and have accordingly been studied extensively.34,36,37 Without proper fit to a persons face, a respirator can leak around its seal, causing the level of protection provided to fall drastically below its PRE classification.38 Even with proper fit, Ba azy and colleagues39,40 showed that particles can penetrate a nominal N95 medium at a fraction greater than 5% (although they declined to name the specif ic models of FFR te sted). The protection provided by the mask while worn is not necessa rily equivalent to the filtration efficiency of the filter f abric on its own. Another risk in the use of filters is fomites, which are inanimate objects capable of transmitting infectious organisms. Because infectious particles are trapped within the 22
fibers of a filter, the filter may become a fomite. The filter may protect a person from infection until he handles it and acquires an infectious organism by contact transmission.41,42 The threat from fomites is seve re enough that a National Academy of Science committee recommended that per sonal FFRs not be reused at all.43 Designing and choosing a respirator involv es balancing the risk of infection with the suitability of the respirator for the specific work environment and user.34 Creating a more mechanically efficient FFR is difficult because the p across the filter cannot be so high that breathing is uncomfortable whil e wearing the mask. Hence, simply adding more layers of fabric is often not an option. Antimicrobials for Aerosol Filtration Antimicrobials have been examined as a way to enhance air purification systems by killing microorganisms in addition to and in lieu of capturing them while minimally affecting the mechanical properties of the filter. Such filters are typi cally impregnated or coated with an antimicrobial substance, al though the surface chemistry of fibers can also be modified.44 Generally antimicrobials serve to chemically cause metabolic or structural damage to the microorganism and cause it to become non-viable: the mechanism differs for different antimicrobi als, and can vary with the presence or absence of other substances, such as water vapor in an airstream.45 Marchin et al.,46 in work on water disinfect ants, draw a distinction among antimicrobials that are constant-release, which release a background of antimicrobial into the fluid; demand-release, which rel ease antimicrobial only in the presence of microorganisms; and contact, which require the microorganisms to make physical contact with the antimicrobial. In the aerosol state, contact antimicrobials would require some form of capture, bounce, or reaerosolization to occur: while this would help 23
prevent the filter fr om becoming a fomite, it would mi nimally affect an aerosol passing through the filter. The comm on test in which a microorganism is streaked across an antimicrobial fabric in a Petri dish does not tell the whole story of the antimicrobials mechanism when used in an aerosol filter. Common antimicrobials used in aerosol f iltration include quaternary ammonium compounds,47 N -halamines, iodine compounds, and silver.45,48 Foarde et al .45 studied what were, at the time, the three antimicr obials registered with the US Environmental Protection Agency (EPA) for use on aerosol filters. They f ound those antimicrobials to be effective and noted that the application of the antimic robials did not appreciably affect the filters PRE. Verdenelli et al .48 showed that if fiberglass filters were loaded with bacteria or fungi, the f ilters treated with quaternar y amines had (depending on microorganism) much lower or zero counts of viable microorganisms, which prevented their becoming fomites. The same group st udied a range of antimic robials on fiberglass filters.49 They found that not all antimicrobial substances they examined, including a formulation of quaternary amines, are chemically compatible with the fiberglass material. For the bacteria and fungi they te sted, a different quaternary amine was more successful than the others. However, Marchin et al.46 state that in water quaternary amines are not effective against viruses and cysts. Antimicrobial silver is beginning to fa ll out of favor because of overuse and subsequent evolving bacterial resistance.51 Also, silver is not effective against viruses in water and suspected to be ineffective in air as well.46 Sullivan found minimal antimicrobial capacity in a commerci al air filter containing silver.50 Some researchers are still examining its use.52 In more-recent work silver is examined in combination with 24
other antimicrobials, such as titanium dioxide.53 Other metals have some antimicrobial action, but data describing their effectiveness in the aerosol state are rare. Byeon et al .54 suggest that copper-coated activated car bon could be used as an antimicrobial and also for its adsorptive capabilities, alt hough they do not use an aerosol challenge. While a good deal of literature on aqueous antimicrobials exists, information on antimicrobials for aerosol filtration is surp risingly sparse, especially considering the commercial availability of antimicrobial aeros ol filters. Sometimes researchers assume that the mechanisms in water and air are th e same, but this is not always an easy or valid assumption to make, as shown in the next section. The Antimicrobial Poly(styrene-4-[trimethylammonium]methyl triiodide) Safe Life Corporation (San Diego, CA, the parent company of Triosyn Corporation) produces filters, for both i ndividual and collective protecti on systems, containing the antimicrobial resin poly(styrene4-[trimethylammonium]methyl tr iiodide) (PSTI). Taylor et al.55 in 1970 showed PSTI to be an effective, broad-spectrum disinfectant in waterbased solution. PSTI has been shown effective against threats including E. coli Giardia muris and G. lamblia Newcastle virus, polyomavirus, and a number of phages.46,56 In-vitro results have suggested that air purification products incorporating PSTI provide a 99% increase in VRE compar ed to standard filtration systems.31,60 PSTI air filters do appear to have some sensit ivity to temperature and RH.63 Ratnesar-Shumate et al.64 have proposed that the mechanism of displacement of I2 from the surface-bound I3 complex proposed in water by Taylor et al.55 applies as well to bioaerosols undergoing near-contact with treated fibers during passage through an air filter medium. In this mechanism, microorganisms which generally carry a negative charge on their outer membrane or coat pass sufficiently cl ose to the resin to displace an I2 molecule from 25
the I3 complex. The I2 molecule sticks to the microor ganism and damages it, causing it to become non-viable. Since bioaerosols have higher surface charges than inert aerosols,65 theoretically PSTI is a demand-release antimicrobial in air. However, the damage realized may depend on a number of external factors. The measurement of the viability of mi croorganisms after passage through a PSTI filter is not entirely straightforward. Bioaerosols are often collected in impingers, but chemical species that off-gas from the filter also collect in the impinger fluid and may build up to toxic levels. Lee et al .63 showed that PSTI can off-gas enough I2 to cause the impinger liquid to become toxic to microor ganisms. (PSTI is thus not exclusively a demand-release antimicrobial.) The test sys tem cannot discriminate between killing of microbial agents in the aerosol state or in the impinger system. A proposed solution to this problem is the addition of reagents to the collection medium in the AGI ( e.g., sodium thiosulfate) to inactivate chemical species deposited in the collection fluid. Lee et al.63 used this strategy when performing MS2 aerosol challenges of samples of PSTI media. They found that collecting in a thio sulfate solution effe ctively eliminated the increase in VRE over PRE, and collecting into a moderate excess (3%) of bovine serum albumin (BSA) caused only a 90% increase in VRE. Eninger et al.66 performed tests in which bioaeros ols passing through a PSTI filter were collected into gelatin plates and measured no reduction in viable concentration caused by the PSTI. The absence of a reduct ion is likely a consequence of successful competition for aerosol-bound I2 by the gelatin matrix. Rengasamy et al .42 measured the survivability of MS2 virus captured within a PST I filter to determine if it prevented the filter from becoming a fomite. They found that the PSTI filter did not cause a significantly 26
larger reduction in viabilit y than a non-antimicrobial filt er at low RH and temperature.42 The lack of a difference may be because the dry capture surface of the antimicrobial filters lacks the activating agent needed by the deactivation mechanism, such as water. Given this information, what would happen when a microorganismiodine complex impacts the mucous membrane of a living being is not at all clear. The question remains: does a PSTI filter give an advantage in protection against infection by airborne pathogens compared to a non-tr eated filter? This questi on will be addressed by performing an animal exposure study in which challenges of a microbial agent will be delivered to age-, sexand weight-matched test subjects in parallel experiments through a PSTI medium and through a mechanically ma tched inert medium. In this design the animal replaces the impinger as the detecto r. This planned experiment is expected to provide conclusive evidence that the incorporation of available I2 at the surface of the air filter fiber does (or does not) convey clinical effectiveness to the medium. Review of Animal Inha lation Exposure Systems Reviews of experimental inhalation exposure system s have been performed by Drew and Laskin,67 MacFarland,68 Cheng and Moss,69 Jaeger et al.,70 Roy and Pitt,18 Wong71 and others. Inhalat ion exposures are generally performed when a researcher is studying an otherwise unquantifia ble biological response like measuring the MID50 of a microorganism for which dete rmination there is no useful in-vitro surrogate. Usually an animal is used as a model of human respiration, the most common being the common laborator y rat (strains of Rattus norvegicus ) and mouse ( Mus musculus ). Animal respiratory systems are im perfect approximat ions of the human respiratory system. For inst ance, the deposition efficiency in rat lungs reaches a minimum near 1 m, while in the vastly different physiology of human lungs it is closer 27
to 100 nm.71 Because of the risk of death and disease, high-risk experiments on humans are usually unethical if not outri ght illegal, and burdensome to perform when they can be morally justified. The Nuremberg Code of ethics for human experimentation, developed in response to atrocities committed in experiments on concentration-camp prisoners by German scient ists within the Nazi regime, requires among other things that experiments on hum ans be based on the results of previous animal experiments.72 Standard procedures for the ethica l selection, care, and use of laboratory animals exist and are accepted by most (but by no means all) scientists.73 These procedures mandate the use of the least-sentient ani mal that is appropriate for the experiment, thus the widespread use of laboratory rats and mice. Laboratory rodents are also, compared to other test ani mals, inexpensive to acquire and maintain. Animal inhalation exposures of vapors appear in the scientific lit erature as far back as the late nineteenth century.67 These early exposures were performed by putting the animal in a large container with room to mo ve around and flowing the challenge into the animals ambient atmosphere for it to breathe: this is called a whole-body exposure and is still in use. Whole-body exposures are a natural way to expose the animals and do not stress the animals by restraining them. Also, the animals can be housed in their exposure chamber, reducing contamination that may occur while transporting the animal. On the other hand, the animal can be exposed by other routes, such as dermal (the substance makes contact with the skin) or oral (the substances lands on or is absorbed into its fur and ingested when the ani mal grooms itself). Good air mixing in the chamber and a comparatively large amount of material is required, which may be problematic if the test ma terial is highly hazardous.71 Whole-body exposures have been 28
used on animals as small as rodents and as large as dogs,67 and in the famous Operation Whitecoat, lasti ng from 1955 to 1973, the US Army performed experiments where humans were exposed to infectious bioaerosols in a 1-million-liter whole-body exposure chamber.72 Henderson, in 1952, made the important innovation, in the eponymous Henderson apparatus, of exposing only the nose and mouth of the animal to infectious aerosols, rather than the ani mals whole body.74 This kind of exposure is called a nose-only exposure or nose-and-mouth-only exposure. Nose-only exposures have the benefit of reducing the amount of materi al needed and eliminati ng the other routes of exposure in the animal. However, because the animals must be restrained, this method is not suitable for long exposures. The restraints st ress the animals, and t hey may attempt to turn around in their restraints and accidentally suffocate themselves.71 A further distinction is made between noseonly exposures in which the animals breathe from a common plenum of air and exposures to the animals through individual air sources that are drawn out through a sepa rate plenum. The first are called flow-past and the second are called directed-flow.70 In flow-past systems the exhalations of animals earlier in the line are breathed by those further down: this may introduce undesirable variability.71 An example of directed-fl ow exposures is the highly sophisticated system Baumgart ner produced for the study of tobacco smoke in 1980.75 Baumgartner is also responsible for desi gning the Battelle restraint tubes commonly used in nose-only exposure systems. Rihn et al.76 validated a system for exposing mice to aerosolized asbestos fibers: he commented that studies on nose-only procedures for mice were, at the time he wa s writing in 1995, still fairly rare. Other nose-only aerosol 29
exposure systems have been produced for expos ure to aerosols of radioisotopes,77 asphalt fumes,78 and pharmeceuticals,79,80 as well as for general purposes.70,81 Experiments exposing animals to aeros ols are common, as are experiments passing bioaerosols through filters; however, experiments using animal models to study the infectivity of filter ed aerosols are not. Studies have been done on the reduction in infection that occurs when filters are added to the cages of pigs82 and chickens85,86 in an agricultural setting, but these studi es examined casual transmission between animals and did not expose the animals to a metered challenge of aerosol. No reports of controlled exposures of an animal model of human respiration to an infectious bioaerosol penetrating a filter were found in the literature. Without an interaction like the PSTIiodine chemistry problem described above, simple collection in impingers is sufficient and animal experiments are not nece ssary: that the PSTI animal study is the first to require such a system is not a surprise. Exposure of a bioaerosol to the PST Iiodine chemistry described above introduces the complications of timeand environment-dependent effe cts on the viability of microbial components of the aerosol. Bec ause these effects can affect the claimed antimicrobial capability of the PSTI component of the fi ber, measurement of the protective impact exerted by PSTI can be acco mplished only with a biological indicator. Therefore, developing and characterizing a syst em to perform such an animal exposure experiment was necessary. Factors Influencing an Animal Inhalation Exposure System Before building the system, it was necessary to learn what to consider in the design of such a system. Wong71 identified four key factors in which variability can adversely affect an ani mal inhalation study: 30
individual response of the animal, animal environment, inhaled dose, exposure atmosphere. The individual response of the animal is outside the scope of this work, as is a major component of the anima l environment, the housing envir onment of the animals. The remainder of the animal environment is the exposure device used to expose the animals to an aerosol and their surroundi ngs during the experiment. As described earlier, exposure devices can be wholebody or nose-and-mouth only. For the PSTI animal study, in which the MID50 can be reached after a relatively small period of time of exposure, a directed-flow nos e-only system is appropriate. The inhaled dose received by the animal can be calculated as Equation 1-2, wherein C is the concentration of test mate rial in the animals breathing air, Vm is the minute volume of the anima l (breathing rate [in min-1] times tidal volume), F is the fraction of material deposited, and t is the duration of exposure in minutes.87 In a test of infectious bioaerosol, C is a concentration of airborne viable microorganisms (PFU/m3 or CFU/m3). The constants C Vm, and F may vary in time. In reality, measuring these three quantities in real time is often not an option and instead these quantities are assumed to be constant. In the st eady state, Equation 1-2 simplifies71,87 to Equation 1-3. When making the steady-state assumption, as was done in this work, F, Vm, and C and must be kept as constant in time as aa. chievble (1-2) (1-3) The fractional deposition F depends on the PSD, and the sites in the animals respiratory tract on which the particles deposit can affect infectivity. Even particles that 31
are not infectious can cause irritation or ot her effects that may affect the animals respiratory system, resulting in swelling of membranes or increased mucus production: particles of one size may affect the deposit ion of particles of another size and the animals immune response.71 Variations in the PSD of the exposure atmosphere are a possible source of error in the dose. The breathing rate Vm of an animal in laboratory conditions can vary wildly from its textbook values and be a source of error. Real-time measurements of respiration have been made on large animals, but no instances of real-time respiration measurements on mice can be found in the literature. Fairchild88 reported that each mouse has a mean tidal volume of 0.18 mL and breathes 255 times a minute. However, in a laboratory situation, the breathing rate of an animal may vary widely from values recorded in less stressful situations, and flows from 1.5 to 10 times the total minute volume of the exposed animals have been recommended for nose-only systems.71 Obviously, variation in the airborne viable concentration C of infectious microorganisms is a source of variation in the dose. As mentione d in the introduction, the RH and temperature of a bioaerosol can a ffect its viability. T herefore, the system must keep the loss of viable particles due to humidity and temperature constant by keeping those factors consistent. In an animal exposure system moderate loss of viability within the system can be tolerated as long as t he viable concentration is consistent and the desired concen trations are attainable. The aerosol source is another important pot ential source of variation in the PSD and viability. Wong71 states that maintaining a stable concentration of aerosols is notoriously more difficult than other inhalat ion challenges. While that is certainly true 32
for fungi and dry powders, Collison nebulizer s are often used to aerosolize a steady airborne concentration of viruses and bacteri a from a liquid source in bioaerosol experiments: Hogan describes their use in animal tests as almost exclusive.21 A steady Collison output depends on a steady feed pressure, as variations in pressure can alter the PSD.16 In practice, evaporation of the nebu lization liquid over time causes the aerosol concentration to increase with time at a slow rate. Also, because Collison nebulizers are recirculating systems that impose large shear forces, viability may decrease slowly.17 Significant variation in output occurs among different models of Collisons.89 This variation is not enough to change an experimental protocol, but using only one model of Collison in any series of experiments is wise. Another factor that can affect the PSD of the challenge is the loading on the filter: as mentioned earlier, increased loading can increase p and can alter PRE. However, if the cumulative loading during the period of experimentation is low enough, changes in PRE and p are negligible. All these factors mu st be considered to keep an aerosol challenge consistent. The gold standard for validation of an expos ure system is to perform an exposure of animals and measure the consistency of the dose by measuring how much is deposited in the animal. This validation is mo re suitable for some exposures, where the dose is a static quantity that remains in the animal, than others. Directly performing this sort of validation is less helpful for cha llenges of infectious bioaerosols, because the microorganisms multiply when they reach a host. Henderson measured the time-based coeffi cient of variation (CV) of viable concentration in the aerosol cloud in his app aratus and showed that within nine trials 33
34 using aerosolized Bacillus subtilis and 15 trials using Chromobacterium prodigiosum the mean CV was 5.73% and no measured CV was above 15%.74 Henderson remarked that this was very consistent and a num ber near Hendersons result was used as an objective in this work. Objective The hypothesis was proposed that an aerosol delivery system based on the Collison nebulizer can be designed and engineered to provide, at selectable concentrations, a respiratory challenge of bi oaerosol particles that is verifiably consistent in time and that can be fed in separate experiments through treated and untreated control filters to de liver a consistent challenge to a small-animal model of human respiration. The goal of this work was to build an appropriate controlled animal exposure system, characterize the aeros ol challenge delivered by the system, and validate the hypothesis that the challenge is sufficiently uniform to support statistically reliable animal infectivity testing. (Performing the animal exposure study was not a goal of this work; instead, it is enabled by this work .) The criteria for success initially set out were that within a number of individual experiments, from the time-dependent PSDs of the aerosol penetrati ng the filter, the TPCs, CMDs, and GSDs all have CVs less than 20%, and from time-series aerosol samples collected in the impingers, plated, and counted, the downstream airborne viable concentrations have a CV less than 20%.
CHAPTER 2 MATERIALS AND METHODOLOGY Materials Controlled Aerosol Test System An aerosol delivery system based on the Collison nebu lizer was designed and built. The system, called the Controlled Aeroso l Test System (CATS) and illustrated in Figures 2-1 and 2-2, enables experiments measuring infection rates of a common laboratory mouse to discriminate the extent, if any, to which a treated air filter medium diminishes the exposure risk from an aeros olized pathogen challenge compared to the same challenge delivered through a mechanically equivalent untreated filter medium. The CATS generates a stream of biolog ical aerosol at a range of constant concentrations, passes the aerosol thr ough a filter, and delivers the penetrating particles to a mouse model of human respir ation. Accommodation of the mouse model was a necessary aspect of the design and construction processes: it was also necessary that all components carrying aerosol flow fit within the biological safety cabinet where it will reside for the animal exposure trials. This cabinet is a SterilGARD III Advanced Animal Transfer Station (Bak er Company, SG603-ATS), which has interior dimensions of 27 in H 20 in D 68 in W. The largest sash opening allowed when an infectious agent is present is 8 inches, wh ich limits the reach of the operator(s). The convenience of the operator wa s considered in the design. Tubing used to connect components containing aerosol flow is -inch stainless steel. All curves in the tubi ng containing the main aerosol flow are gradual and smooth, with an inner curvature radius greater than 1 inch. All valves carrying aerosol flow are -inch stainless ball valves. Flows of ma keup and purge air are controlled by -inch 35
needle valves, with toggle valves before the needle valves to enable the operator to turn flows off and on without having to readjust the needle valves. The needle valves are followed by rotameters to ve rify the flow rate. The Collison nebulizer is preceded by a toggle valve and rotameter. (The rotameter before the Collison is at pressure and will read lower than its actual flow: it can be corrected using Equation 2.51 in Hinds.90 At a Collison pressure of 25 psig the rotameter can be corrected by multiplying its reading by 1.64: at 30 psig, 1.74.) In the CATS, air is supplied to the system by an air compressor. For this work the lab air line was filtered first through an oil trap and then a DFC-21 HEPA canister particle trap (Porous Media Corp., St. Paul, MN) to feed the nebulizer and porous tube diluters. The animal studies require a source of breathable air free of both particles and toxic gases and vapors, to be provided onsite. The incoming air is then regulated to a pressure of 50 psig and flowed through a porous tube humidifier (PermaPure LLC, Toms River, NJ; model MH-070), which contains a Nafion membrane tube. Water on the outside of the tube is transported through the membrane into the air flow. T he humidity can be controlled via the water temperature, although water at a different temperature than ambient was not needed in this work. If a low RH is needed the tube can be removed, althoug h evaporation from the Collison nebulizer increases the hum idity of the aerosol flow somewhat. The airflow is then manifolded. Some of the airflow is regulated to 30 psig and flowed into the system later on. The rest of the flow is regulated to 25 to 30 psig and flowed to a Collison nebulizer (BGI Inc., Wa ltham, MA), which gene rates the bioaerosol. A single-jet Collison nebulizer is used becau se it does not produce excessive flow. 36
Make-up air supplies the rest of the flow. A 1-psig pop valve to prevent overpressure and a pressure gauge (Dwyer Instruments, Houston, TX; Magnehelic series 2000) tee off directly after the Collis on. A porous tube diluter (Mott Corp., Farmington, CT; model #7610105-020) is used to deliver make-up air from Valve A to adjust the flow rate after the nebulizer. The porous tube lets the two air flows mix in a non-turbulent fashion. The nebulization process puts charges on the created particles. A charge neutralizer (TSI Inc., Shoreview, MN; Model #3012A) is necessary to neutralize that electrical charge. The 3012A c harge neutralizer uses a 370-MBq 85Kr beta-emitting source. It can be used with flows as high as 50 L/min. After the charge neutralizer, a length of tubing guides the flow to an inters ection where the first sampling point in the system, Valve and Port 1, tees off and can be connected via -inch conductive silicone tubing (TSI, part #3001789) to sampling instru mentation. A type of O-ring compression fittings known as Ultra-Torr hose connectors are present at the sa mpling ports to allow the operator to easily connect and disconnect instrumentation. After this tee is a custom-built samp le holder (Triosyn Corp, Williston, VT) comprising an inner and outer sleeve holding a 47-mm diameter disc (40-mm exposed) of filter medium compressed (by bolts around the edges) between elastomeric annular seals. The sleeve has a tapered chamber 10 cm long before the f ilter to allow the aerosol to spread, and then a tapered chamber 10 cm long after the filter to return to the tubing. The holder can accommo date other sizes of filter with the use of reducers, although reducers were not used in these experiments. A second sampling point is directly downstr eam of the sample ho lder at Valve and Port 2. A differential pressure gauge ( Dwyer, Magnehelic series 2000) is connected 37
before and after the sample holder to measure p across the filter. Valve 3, following the tee for Valve 2, is necessary to isolat e the aerosol flow from the animal subjects during post-exposure samplings of the aerosol. Downstream of Valve 3, flow from Valve B can be supplied immediately after the exposure is term inated to give the animals clean breathing air before their remo val from the exposure system. Next, the exposure system, a JaegerNYU Modular Nose-Only Directed-Flow Rodent Inhalation Exposure Unit (CH Technologies, Westwood, NJ),91 hereinafter referred to as a mouse tree, is used to ex pose the mice to the aerosolized agent. A nose-only system was, among other reasons, chosen to prevent cutaneous and enteric infections to the mice. The capacity of the mouse tree to deliver infectious aerosol to mice was not tested in this work. The m ouse tree itself has been validated and verified in the literature by Jaeger, so rep eating that process is not necessary.70 Each mouse is placed in a polycarbonat e holder and constrained with a sealed restraint inserted in the rear opening of the holder so that the ti p of the mouses nose projects out of an opening in the front of t he holder. The holder inse rts securely into a socket on the mouse tree. Vents inside t he body of the tree blow an airstream containing the filtered aerosol at the nares of the mouse as its only source of breathing air. Exhaled air and excess flow are drawn away from the mouse.70 The mouse tree is a directed-flow system and no mouse rebreathes flow from other mice. The mouse tree can hold up to 12 mice at one time. A rotating joint is present at the inlet to the mouse tree to allow it full range of rotation and make all the sockets accessible. At the effluent of the m ouse tree, the relative hum idity and temperature are measured by a National Inst itute of Standards and Technolog y (NIST)-traceable digital 38
hygrometer (Control Company, Friendswood, TX; Model 35519-020). The flow may either be sampled at Valve and Port 4 or exhausted through another HEPA canister filter, after which a flow meter (TSI, Model #4143D) m easures the flow rate. Sampling Instrumentation All-glass impingers Built into the CATS is a hook-up for sampling with impingers. Sampled aerosol flow is combined with flow from needle Valve C in another porous tube diluter. This combined flow is drawn through Valve and Port 5a or 5b into AGI-4 impingers (Ace Glass Inc., Vineland, NJ). When a vacuum is drawn on AGI-4s they draw 12.5 L/min of air.21 Therefore the rate of sa mpling from the system is c ontrolled by the make-up air from Valve C, which is usually set to 10 to 11.5 L/min (i.e., 1 to 2.5 L/min of aerosol sampling). The efficiency of an impinger de creases with time because of evaporation and aerosolization of the sampling liquid.21 Therefore, two impingers must be present in the system so that t he operator can switch to a fresh impinger after a time and replace the used impinger base with a fresh one. Only one impinger is in use at any given time, although when switching between the two, Valves 5a and 5b may be open simultaneously. A vacuum pump is used to dr aw air through the impingers, and thence through a HEPA filter to c apture uncollected aerosol. T he pump is a generic component and, being downstream of the HEPA trap, it is not a potential source of contamination. Particle sizers The operator can also make measurement s using particle sizers. Two different instruments are used to measur e PSDs. For particles in the micrometer range, such as most bacteria, an aerodynamic particle sizer (APS) is used. For particles in the nanometer range, such as viruses, a scanning mobility particle sizer (SMPS) is used. 39
The APS (TSI, Inc.; Model 3321) operates by measuring the time of flight of particles accelerated through a nozzle. The a cceleration is measured by parallel lasers. Particles from 0.5 m to 20 m can be si zed by the APS. The APS samples at a flow rate of 5 L/min and can sample continuously. TSIs SMPS consists of a Model 3080 elec trostatic classifier with a 3081 long differential mobility analyzer and a 3785 con densation particle counter. The 3080 and 3081 separate particles based on their electric al mobility. The separated particles pass into the 3785, which grows the particles by condensation of water vapor and counts the resulting droplets optically. Particles fr om 10 nm to 1 m can be sized with the SMPS. The length of the particle sizer s sampling interval depends on the total concentration of particles in the air; lower concentrations require more sampling time to get sufficient particle counts. The parti cle range depends on the sampling rate. At 0.6 L/min, particles with diameters from 10 to 410 nm can be measured, and other size ranges require a sampling rate of similar magnitude. Challenge Microorganisms Using a challenge of bioaerosol was necessa ry to validate the system, but as the system was not contained in a biological sa fety cabinet for this work, infectious bioaerosols were not an option. Two nonpathogenic microorganisms were chosen for this work: MS2 coli phage virus and Bacillus atrophaeus bacterial spores. MS2 coli phage MS2 bacteriophage is a small RNA virus that lives on male cells of E. coli bacteria. It has an icosahedral virion with a diameter92 of about 27 nm. Its coat does not have a lipid layer, thus in theory making it more stable at high RH.11,92 Trouwborst et al.93 showed that if MS2 is nebuliz ed from a fluid with appropriat e concentrations of protein 40
and salts, over an interval of 30 minutes it loses less than one order of magnitude of viability, regardless of RH. Agai n, that the challenge stay c onsistent is more important than that it not lose any vi ability, and since this MS2 stock grew at a titer of 1011 to 1013 PFU/mL, one order of magnitude of lo ss poses no problem. MS2 is a common simulant for infectious viruses. Because keeping MS2 viable in the aerosol state is comparatively easy, it has been used in a large number of studies.21,25,31,32,37,39,42,50,61,63,66,93 Stock of MS2 virus, from American Ty pe Culture Collection (ATCC) 15597-B1, was grown in E. coli (ATCC 15597) in tryptic soy br oth (TSB) according to standard EPA protocols.94 To determine viability, a single-layer plaque assay was performed.94 In this assay, a 1-mL portion of a serial dilu tion of the impinger aliq uot was combined with 250 L of an E. coli stock and 9 mL of warm, liquefie d tryptic soy agar (TSA). This mixture was poured into an empt y Petri dish and left to cool and solidify, and then the plates were incubated overnight. Plaques were counted the next day. Bacillus atrophaeus The genus Bacillus makes up a variety of endos pore-forming, Gram-positive rod-shaped bacteria. Bacillus spores are resistant to ai r-drying and other stresses, and therefore can be found in a wi de variety of environments.95 B. anthracis is the causative agent of the disease anthrax and is of c oncern in bioterrorism defense. Some noninfective Bacillus species are often used as simulants in bioaerosol tests, because of their hardiness and their similarity to B. anthracis .22,61,62,65,74,96 B. atrophaeus is virtually identical to the common species B. subtilis, or hay Bacillus except that B. atrophaeus produces a black pigment on media containing an 41
organic nitrogen source.97 Bacillus cells range from 0.5 m D 1.2 m L to 2.5 m 10 m: B. subtilis and B. atrophaeus are on the small end of this range.95 B. atrophaeus spores, from ATCC 9372 stock, were grown in TSB according to standard methods,98 at a titer of approximately 108 CFU/mL. Samples containing B. atrophaeus were applied with a spiral plater (Mi crobiology International, Frederick, MD) onto TSA plates and then incubated overnigh t. Colonies were counted the next day. Filter Media Three different filter media, extracted from FFRs available on the market, were used in these consistency tests. Samples were taken from these filters using a 47-mm circular punch and mallet. An off-the-shelf HVAC filter rated at Minimum Efficiency Reporting Value (MERV) 8 was also used, but it was found to have PRE near zero in the size range considered in this work, so it was discarded. Safe Life T-5000 The Safe Life T-5000 FFR is a NIOSH-cert ified P95 respirator. This filter was chosen because it contains the PSTI resin and is therefore similar to the filters that will be used in the animal exposure study. A flow rate through the filter of 5.3 L/min was used because Safe Life Corp. specified for t he animal experiment a testing face velocity for their material of 7.08 cm/s. At the 85-L/min flow rate used by NIOSH for testing FFRs,35 this face velocity scales to an FFR with a 200-cm2 surface area, which is a reasonable estimate of the true area. (A rough measurement wit h a ruler of a T-5000 mask gives 160 cm2.) Measurements of the T5000, either the PRE of the fabric or t he protection of the FFR when worn, do not seem to exist in the literature. Rat nesar-Shumate et al .64 measured the PRE and VRE of Safe Life-produc ed P95 filter fabric, although their fabric 42
was designed for respirator cart ridges, not FFRs. In their work, fluorescent particles with a mass mean diameter of 0.27 m were removed with PRE near 99%. The bacteria E. coli and Micrococcus luteus were removed with almost five nines of VRE.64 The T-5000 consists of a covering outer layer, a layer of electrically charged polypropylene filtration mate rial embedded with particles of PSTI, a carbon layer to reduce organic vapors, and a supporting inner laye r. Only the filtration layer was used. 3M 1860S The 3M Corporation produces N95 particula te respirators that are commercially available in most hardware stores. Model 1860S was used in this work, as an alternative to the T-5000. Coffe y reports this respirator as, when worn with a correct fit, filtering with 95% PRE among 90% of wearers.38 However, measurements of the efficiency of the filter fabric it self do not seem to exist in the literature. The filter media of the 1860S is an electret m ade of polypropylene fibers. Isopropyl alcohol-treated 1860S Exposing electrically charged filter materi al to vapors of isopropyl alcohol (IPA) removes the electrical charge and decreases the filters PRE. One swatch from an 1860S was exposed to vapors of IPA to enhance penetration. Methods Leak Check After the systems construction was finis hed, it was leak tested by replacing the Collison nebulizer with a plug and pressurizing the system to a few inches of water, then observing it for an hour. If no significant c hange occurred, the system was deemed leakfree. A full description of this leak check procedure is in Appendix A. 43
Flow Rate, Relative Humidity, and Temperature Consistency To determine the consistency of flow ra te, RH, and temperature, the system was operated with deionized (DI) water as the Co llison liquid. The flow rate at the exhaust and the RH and temperature were recorded over a period longer than an hour. No filter was used in this test. Correlation of Sampling Ports To determine the loss of particles during flow in the system, and to make sure samples from different ports could be compared with one another, a correlation of sampling ports on the instrument wa s performed by nebulizing two separate suspensions, one of 250-nm polystyrene latex ( PSL) beads (Duke Scientific, Palo Alto, CA; G250) and one of 1-m PSL beads (Duke Sc ientific, 4009A). Beads were dispersed in DI water (on the order of one unit of bead solution to ten units of water) as the Collison nebulization liquid. The makeup flow was adjusted to deliver a total flow of 5.3 L/min, and the system was allowed to equilibrate. No filter was used in this test. Each of the sampling ports (1, 2, and 4) and the ports on the impinger hook-up (5a and 5b) were sampled repeatedly with the par ticle sizer, as were ports on each quadrant of the mouse tree. For readings fr om the mouse tree, the sampling tube was inserted into a mouse restraint device and inse rted into the sockets of the tree at four different quadrants. For Ports 5a and 5b on t he impinger hook-up no dilution air was added: Valve C was closed. The dilution air w ould be too much for the particle sizer to sample and some would need to be vented by opening the alternate port to prevent backflow. Whether venting would entirely prevent backflow was unclear, and when the test was attempted with venting, the reading s deviated wildly from what was expected. 44
For the 250-nm beads, readings were taken in triplicate at each port and the mean of those three readings was used. For 1m beads, only one reading was taken at a time. The consistency was calculated based on the combined concentration at the aerodynamic diameter where the peak occu rred and the two surrounding data points. Bioaerosol Consistency Trials To test the consistency of the challenge delivered, a bioaerosol was created and flowed through the system. Microo rganism stocks were indivi dually diluted in filtersterilized water and delivered into the Collis on nebulizer. The Collison spray was started and the make-up flow was adjusted to deliver a total flow of 5.3 L/min. The system was allowed to equilibrate for 15 minutes bef ore 5-minute impinger samples into 1X phosphate-buffered saline (PBS) were taken at sampling ports 1 and 2, and particle size measurements were taken at port 1. Particle counts at port 2 were too close to the instrument error to be useful for m easuring consistency, though downstream measurements were made to ve rify filter integrity. p was observed throughout the experiment. A step-by-step operating sequenc e for these experiments are in Appendix A: a specific sequence of movements was used to prevent splash from the impingers. The first series of tests nebulized 30 mL suspensions of MS2 virus with nominal titers ranging from 108 to 1012 PFU/mL. (The nominal titer is the concentration of viable microorganisms in the liquid, calculated based on the original titer of the stock and the dilution ratio.) For each test in this seri es, T-5000 medium was used. Pairs of particle sizer measurements upstream of the filter concurrent with an impinger collection downstream alternated with downstream samp ling into impingers: thus, two upstream PSD measurements and a downstream impinger sample were made during the first, third and fifth 5-minute sampling period, and ups tream collections into impingers were 45
made during the second, fourth and sixth periods, resulting in a total of six values for PSD and three each for viable counts befor e and after the filter. The SMPS scanning period was consistently 135 seconds long. Im pingers sampled 1.5 L/min of aerosol flow with makeup air to increase the flow rate to 12.5 L/min. Total sampling time was a bit more than 30 minutes because switching be tween impingers was labor intensive. A second series of tests nebulized 16 to 30 mL suspensions of B. atrophaeus spores with nominal titers of 107 to 87 CFU/mL. Media used were T-5000, 1860S, and IPA-exposed 1860S. The impinger s sampled at a rate of 2. 5 L/min of aerosol flow during the first three experiments and at 1.5 L/min during the final three, with makeup air to increase the flow rate to a total 12. 5 L/min. Because the APS draws more air than the SMPS, it could not be operated an the same time as an impinger; therefore, particle size measurements were taken before and a fter impinger measurem ents, resulting in seven particle size measurements (with so metimes an extra initial measurement) and three viable counts at each sampling point per experiment. The APS sampling period was 20 seconds. The total length of an experiment was about 40 minutes. After the experiments, the im pinger media were serially diluted and plated in triplicate, and incubated overnight at 37 C. Plaques or colonies were counted the next day. The remaining liquid in the Collison nebulizer was also pl ated in triplicate and counted. The dilution series deemed to be the smallest dilution that was not too overgrown to be reliable (generally, fewer than 60 CFU/plate or PFU/plate) was used to calculate the concentration. The titer of the Collison liquid was calculated by Equation 2-1, where N is the count of CFU or PFU on the plate, Vp is the volume of liquid plated (always 1 mL), and n is the dilution factor. The airborne viable 46
47 concentration was calculated from t he plates by Equation 2-2, where Vi is the volume of impinger liquid (in this work always 20 mL), Qa is the flow rate of aerosol flow collected (1.5 to 2.5 L/min) and t is the duration of sampling (in this work always 5 minutes). (2-1) (2-2) Penetration curves were calculated for select experiments; for each individual aerodynamic diameter, the mean of upstream readings and the mean of downstream readings were plugged into Equation 1-1. C onfidence intervals were calculated based on the combined standard deviation, assumi ng a normal distribution. The combined standard deviation is calculated as per Equation 3.18 in Taylor99 for uncertainties of ratios, using the standard deviation of upstream and downstream readings. From each time step, the mean and standard deviation of the TPC, CMD, and GSD of the aerosol distribution and the ai rborne viable concentration was calculated. For each individual experiment, the CV is calculated as the ratio of standard deviation of the time-based data points to the mean. CVs are calculated for each individual experiment: data is not pooled between experiments. The aerosol need only remain consistent within an experimen t, not between experiments. Fo r the experiments with the largest and smallest GSD, the CV of the regular standard deviation is also calculated, to demonstrate that the standard deviation does not grow wildly compared to the GSD.
Figure 2-1. Photograph of C ontrolled Aerosol Test System (CATS), with key components labeled. Not pictured: control panel and impinger hook-up. 48
49 Air Compressor Main flow dilution air (A) Mouse tree flush (B) Impinger dilution air (C) 30 psi >25 psi Pop valve Charge Neutralizer Porous Tube Diluter Collison Nebulizer Filter Holder Mouse Tree Particle Sizer Porous Tube Diluter Vacuum Pump12 3 5a 5b Toggle & Needle Valve Ball/Toggle Valve Ultra-Torr Hose Connector HEPA Filter Impingers 1 2 5 L / m i n 50 psi Porous tube humidifier P RH/ T P 4 Flow Meter Collison pressure Manifold pressure Collison flow (D) Conductive Sampling Hose (for connecting to hose connector) Blue lines contain aerosol flow. Arrows indicate direction of flow. Figure 2-2. Process-flow diagram of CATS.
CHAPTER 3 RESULTS Leak Check and Flow Rate, Relative Hu midity, and Temperature Consistency The system was leak checked: it held 3 in H2O of pressure for an hour. Then the consistency of the flow rate, humidity, and temperature was measured. Deviations from the mean for the temperatur e and the exhaust flow rate were lower than 1% for observations during a 90-minute period, and RH stayed within 5%. Toggling Valve C to turn flow to the impingers on or off caused a dev iation in flow of about 2%. Data for this are presented in Appendix B. Correlation of Sampling Ports Data for the port consistency trials performed with inert beads are presented in Appendix C. For 250-nm beads, t he worst case difference between ports is 15% of the overall mean, but the minimum value was taken at the beginn ing of the system (Port 1) and the maximum at the end (Port 4), which is the opposite of what would be expected if particles were being lost along the length of the system, and readings that large did not occur more than once. All deviations re mained within 10% of the overall mean. For 1-m beads, the worst-case difference betwe en ports was 4.8% of the overall mean, and the worst deviation from the mean wa s 2.7%, which is very consistent. Bioaerosol Consistency Trials with MS2 The PSD was observed to be approximately log-normal: a repres entative plot is given in Figure 3-1. Although the PSD meas urements on the downstream side of the filter were not usable for consistency, they could be compared to the upstream measurements to calculate a penetration curve, as in Figure 3-2. With 95% confidence, it can be said that the PRE of the T-5000 in the 10 to 400 nm range is between 99.77% 50
and 99.97%, and the MPPS is likely somewher e between 100 and 300 nm. This PRE is higher than that measured by Ratnesar-Shumate et al.,64 who were using P95 material from cartridge respirators, not FFRs: whether the sa me material is used in Safe Lifes FFRs and their cartridges is unknown. Ho wever, given the high titer of the microorganism, viable penetration was assu med to be possible. Plated viable MS2 counts were not measurable, likely because of problems executing the assay method or contamination in the laboratory workspace. Mean values of RH, T and the TPC, CMD, and GSD of the PSD are presented in Table 3-1, as well as the coefficients of va riation (CVs) of the PSD moments within each experiment. Raw data for this series of ex periments are in Appendi x D. The PSD varied very little over the 30 minutes observed, as reflected in the very low CVs of the moments, all 6% or less. The PSD moment s were not observed to trend upwards or downwards in time during the 30 minutes of observation. No noticeable change in p was observed over the course of the experiments. Bioaerosol Consistency Trials with B. atrophaeus The PSD produced by aerosolizing B. atrophaeus was observed to be bimodal, with particles in the peak near 1 m containi ng bacteria, and a hump of smaller particles presumed to contain only dissolved solids fr om the aerosolization medium. Based on the representative distribution in Figure 3-3, the dividing point between the two modes was taken as 0.8 m, the concentration of part icles larger than 0. 8 m was calculated, and the CV of that concentra tion was measured as well. For the T-5000 medium, penetration in the range measured by the APS was so small as to be indistinguishable from inst rument noise. Because of this, the 1860S medium was tried: the dow nstream data were still unusabl y small. The IPA-treated 51
1860S medium was less efficient: the PSD measur ed downstream of this filter is shown in Figure 3-4, and its curve of PRE versus par ticle size is reproduced in Figure 3-5. With 95% confidence, it can be said that its PRE near 1 m was still no lower than 99.95%. Regardless of this, these experiments were attempted anyways, in the hopes that enough bioaerosol would penetrate over the sampling period to be measurable. Viable counts of microorganisms were measur ed upstream, but no viable microorganisms were detected downstream even past the IPA-treated 1860S because the challenge concentrations were not large enough to overcome the high filtration efficiency of these filters at particle sizes near and above 1 m. Plates from downstream samples showed an occasional lone colony, not enough to calculate from reliably. Table 3-2 identifies the filt er used in each test, lists the mean values of RH and temperature, and reports values of the TPC, CMD, and GSD of the PSD. Table 3-3 contains the nominal titer in the Collison pre-experiment, the post-experiment Collison titer, the mean upstream viable concentra tion, and the CVs of the PSD moments and upstream airborne viable concentration m easured within the experiment. No trend upwards or downwards was observed in the PSD moments. Raw data for this series of experiments are repor ted in Appendix D. No change in p was observed over the course of any experiment in this series. Downstream measurements were performed bef ore and after the exposure of the IPAtreated 1860S medium, as shown in Figure 3-4. While there appears to be more penetration after the exposure, the incr ease is within the error of the APS. By performing a linear regression between the post-experiment tite r of the Collison liquid and the airborne viable concentration, the VSF for the CATS was determined to 52
be 7.8-7. This linear regression had a R2 of 90%. Note that this VSF does not account for the viable collection efficiency of the impingers: if it did, it would be somewhat larger. Also note that this spray fa ctor is for the entire system: the VSF at the Collison nozzle can be ca lculated to be about 2-6 by scaling by the ratio of total flow to Collison flow (5.3/2). The TPC correlated well with the TPC larg er than 0.8 m, with an R2 higher than 99% and a regression constant of 0.4996. In Experiment 901 and the ones following, the post-exposure Collison tite r seems to be depressed by an order of magnitude compared to the nominal titer, while in experiments before 901 the concentrations are of the same magnitude. Neither the TPC nor TPC above 0.8 m showed an obvious correlation with the post-ex periment Collison titer or the airborne viable concentration, although both TPC and TPC above 0.8 m correlate with the nominal Collison titer with R2 near 99% (with regre ssion constants of 10-5 and 7-6 #/CFU, respectively). Figure 3-1. Representative particle size di stribution (PSD) from MS2 nebulization and 95% confidence intervals for each indi vidual diameter. Based on six samples in Experiment 724. 0.0E+00 1.0E+12 2.0E+12 3.0E+12 4.0E+12 5.0E+12 10 100 1000dN/dlndp (#/m3)Aerodynamic diameter (nm) Measured PSD Fitted log-normal curve 95% Confidence interval 53
Figure 3-2. Particle removal efficiency (PRE) of T-5000 medium as a fu nction of particle size and 95% confidence intervals. Ca lculated from six upstream samples and three downstream samples in Experiment 811. 99.75% 99.80% 99.85% 99.90% 99.95% 100.00% 10 100 1000PREAerodynamic diameter (nm) Figure 3-3. Representative PSD from Bacillus atrophaeus nebulization and 95% confidence intervals. Taken from seven samples in Experiment 819. 0.0E+00 1.0E+08 2.0E+08 3.0E+08 4.0E+08 5.0E+08 6.0E+08 7.0E+08 0 5124dN/dlogdp (#/m3)Aerodynamic diameter (m) Figure 3-4. Downstream m easurements from Experiment 910 with isopropyl alcohol (IPA)-treated 1860S medium and 95% conf idence intervals. Three samples each were taken before and after the 40 minutes of experimental interval. 0.0E+00 1.0E+06 2.0E+06 3.0E+06 4.0E+06 0.5 1 2dN/dlogdp (#/m3)Aerodynamic diameter (m) Before After 54
Figure 3-5. PRE of IPA-treat ed 1860S medium as a functi on of particle size and 95% confidence intervals. Calculated fr om seven upstream samples and six downstream samples in Experiment 910. 99.70% 99.75% 99.80% 99.85% 99.90% 99.95% 100.00% 0.5 1 2PREAerodynamic diameter (m) Table 3-1. Mean relative humid ity (RH), temperature, and particle size distribution (PSD) moments, and coefficients of va riation (CVs) of PSD moments for MS2 experiments Experiment RH T (C) TPC (1012 #/m3) CMD (nm) GSD CV of TPC CMD GSD 724 65% NDa 2.65 81.41 1.70 4.66% 1.23% 0.19% 728 61% 22.7 4.51 74.22 1.69 5.21% 3.09% 1.02% 730 64% 22.4 4.16 75.72 1.68 3.31% 1.20% 0.21% 811 55% 27.0 5.13 75.62 1.72 3.16% 1.93% 0.56% 812 53% 26.5 4.51 76.53 1.70 6.00% 0.49% 0.08% 813 56% 26.0 4.11 77.96 1.70 5.15% 0.37% 0.29% Minimum 53% 22.4 2.65 74.22 1.68 3.16% 0.37% 0.08% Maximum 65% 27.0 5.13 81.41 1.72 6.00% 3.09% 1.02% Mean 59% 24.9 4.18 76.91 1.70 4.58% 1.38% 0.39% a No data. The maximums are simply the la rgest entry of data in the above column. Minimums are, similarly, the smallest entry The means are simply the mean of the data in the corresponding column. CVs are calc ulated within each experiment. The CVs of standard deviation for Experiments 728 and 812 are 0.59% and 0.47%, respectively. 55
56 Table 3-2. Filter used, mean tem perature, RH, and PSD moments for Bacillus atrophaeus experiments Exp. Filter RH T (C) TPC (106 #/m3) TPC > 0.8 m (106 #/m3) CMD (m) GSD 819 T-5000 58% 24 123 .0 73.30 1.080 1.33 820 T-5000 57% 24 59.2 49.30 1.100 1.21 827 1860S 63% 23 30.8 6.56 0.705 1.35 901 1860S 50% 23 254 .0 121.00 0.953 1.33 903 1860S 48% 23 259 .0 120.00 0.942 1.34 908 1860S 47% 22 528 .0 254.00 0.948 1.34 909 1860S 47% 23 1208 .0 580.00 0.951 1.34 910 IPA 1860S 45% 23 1174 .0 590.00 0.960 1.34 Minimum 45% 22 30.8 6.56 0.705 1.21 Maximum 63% 24 1208 .0 590.00 1.100 1.35 Mean 52% 23 455 .0 224.00 0.955 1.32 The lower portion of the table is ca lculated the same as Table 3-1. Table 3-3. Viable concentrations and CV s of PSD moments and upstream airborne viable concentration for B. atrophaeus experiments Exp. Collison (106 CFU/mL) Upstream (106 CFU/m3) CV of Nominal Post-exp. TPC TPC > 0.8 m CMD GSD Upstream 819 10 26.3 0 23.30 5.21% 3.12% 0.69% 0.42% 19.21% 820 10 14.0 0 7.15 7.42% 5.26% 0.60% 0.58% 24.02% 827 10 NDa_ NDa_ 9.77% 9.66% 6.18% 0.77% NDa __ 901 16 2.80 3.04 6.81% 3.64% 0.70% 0.15% 18.23% 903 16 NDa_ NDa_ 5.00% 4.58% 0.35% 0.08% NDa __ 908 40 5.57 5.10 3.84% 4.21% 0.42% 0.22% 25.54% 909 80 NDa_ NDa_ 9.21% 4.97% 0.91% 0.19% NDa __ 910 80 18.0 0 12.40 3.52% 4.94% 0.30% 0.14% 5.00% Min. 10 2.80 3.04 3.52% 3.12% 0.30% 0.08% 5.00% Max. 80 26.3 0 23.30 9.77% 9.66% 6.18% 0.77% 25.54% Mean 32 13.3 0 10.20 6.35% 5.05% 1.27% 0.32% 18.40% a No data. The lower portion of the table is calculated the same as Table 3-1. CVs are calculated for each individual experi ment. The CVs of standard deviation for Experiments 827 and 903 are 7.03% and 0.94%, respectively.
CHAPTER 4 DISCUSSION The CATS was designed as an ensemble to deliver a constant challenge of aerosolized pathogens through a test filter to a panel of mice who serve as biological indicators of net viable penetration through tw o categories of test filters. The ensemble comprises a Collison nebulizer, a particle char ge neutralizer, a filter holder and filter, and an animal exposure apparatus. Located th roughout the system are ports from which the aerosol can be sampled with parti cle sizers and impingers. Pressure gauges and RH and temperature sensors are includ ed to measure environmental conditions. The hypothesis tested in this work is that the challenge delivered to the animals is consistent. Variation in the challenge could arise from variations in the PSD of the upstream challenge, the airborne viable concentration, and the PRE or VRE of the filter. The criteria to confirm the hypothesis are that for each individual experiment the TPC, CMD, and GSD of the PSD and the airborne viable concentration have CVs below 20%. Flow Rate, Relative Humidity, and Temperature Consistency Precedents show that variations in flow rate, RH, and temperatur e affect both the PSD and the airborne viable concentration. The flow rate, RH, and temperature data show that the flow rate and temperature of the CATS remain within a range of 1% and RH is consistent within a range of 5%. This compares well wi th the literature: Bonnet et al.78 maintained RH within % in their system. Because the flow rate, RH, and temperature are consistent, they are not significant sources of variation. Correlation of Sampling Ports To prove that samples taken at different sampling ports on the CATS are equivalent, aerosols of 250-nm and 1-m beads were created and the measurements at 57
different ports were compared. The devia tions from the mean of particle readings between ports on the CATS are within % of the mean with aerosolized 250-nm beads and the deviations with 1 m beads are mu ch lower. These measurements were performed with different instruments (SMPS for 250-nm beads, APS for 1-m beads) and the difference in consistency may be a difference between the repeatability of readings on each instrument. The particle readi ngs in this work did not decrease further along the flow path, suggesting minimal particle loss occurs in the CATS. The largest of these deviations is smaller than that meas ured on the system built by Oldham et al.,81 although Oldham et al. were studying an ani mal exposure chamber much larger than the CATS. Measurements taken at differ ent sampling ports on the CATS can be compared to each other with confidence. Bioaerosol Consistency Trials Particle Size Distribution To measure variation in the PSD inside the CATS, measurements were taken with APS and SMPS particle sizers. The low CVs in Tables 3-1 and 3-3 for the statistical upstream quantities of the PSDs are all less than 10%, often very much less. The statistical properties of the PSDs showed no discernible trend upw ard or downward in time during the observation period. Previous work in which qualities of particle size distributions were measured can be compared. Raabe et al.77 created a uranine aerosol to validate his animal exposure system, co llected it on filters, and measured the change in filter weight. Their data can be transforme d to give data points of mass concentration for individual time steps, and CVs can be ca lculated from that transformed data. The CVs for his two experiments are around 27%. Rihn et al.76 measured mass concentration of aerosolized asbestos fibers within their system. From a single 58
experiment measuring mass collected on filt ers, they reported a CV for their mass concentration of 15%. Bonnet et al.78 reported total mass concentrations of particles from fumes of bitumen collect ed on filters; among their th ree experiments, CVs ranged from 17% to 32%. Nadithe et al.79 used an Aerosizer particle sizer to measure an aerosol of radiolabeled human serum albumin They measured a CV for mass median aerodynamic diameter near 13% and a CV fo r GSD of 4%. Based on the CVs measured in this work, the PSD within the CATS was gener ally more consistent than that of other systems in the literatur e, and varied only slightly during the periods of observation. The literature shows that the method used to measure the PSD can introduce a great deal of error into the measurement, an d that some methods are more consistent than others. Particle sizers are more consis tent than other methods, although particle sizers are significantly more expensive and require more upkeep than filters or cascade impactors. When particle sizers are available, their use is a good way to obtain consistent measurements of the bioaerosols PSD. The Collison is also a more consistent method of creating an aerosol than others, although the Collison cannot a ccommodate the smoke particles examined by Bonnet et al.78 or fibers as studied by Rihn et al.76 Since the CATS will only be used with bioaerosols, the Collison is a good choice to cr eate a consistent bioaerosol challenge. Viability MS2 Experiments were performed with aeroso lized MS2 coli phage to determine the consistency of a viral challenge. The PR E graph shown in Figure 3-2 suggests that viable particles were capable of penetrating. MS2 was collected in impingers upstream and downstream of a T-5000 medium and assa yed, but all MS2 plates showed 59
contamination or were other wise unusable, so viability data were not measured. Inexperience with the plating method likel y contributed to lack of success producing viability data with MS2. As well, because no biol ogical safety cabinet was available, the MS2 plating was performed on an open bench. Wh ile, for a BSL-1 organism, plating on the bench poses no hazard of infection, it in creases the contamination risk from ambient air. Since MS2 is a small particle it is dwar fed by the particles produced from the dissolved solids, and any evidence of MS2 in the particle size distribution is obscured by the dissolved solids mode. The MS2 particles ar e invisible in the PSDs, which, in the absence of viability data, ar e the only data collected. Ther e is little difference between the experiments performed with MS2 and an equi valent set of ex periments done with an inert challenge. However, given the success of other researchers in the literature in producing MS2 aerosols over durations sim ilar to the length of these experiments, aerosolization of viable MS2 should be achiev able in this system. Eventually MS2 work was halted and this work moved on to mo ved on to the next or ganism. Measurements of viability were not achieved for MS2 in th is work, but its properties should not differ from the literature. B. atrophaeus Work with B. atrophaeus was begun after MS2 proved problematic. Experiments were performed with aerosolized B. atrophaeus to determine the consistency of the viability of a bacterial challenge. B. atrophaeus was collected in impingers and successfully assayed. Out of seven exper iments performed, five experiments had usable plates. No viable penetration of B. atrophaeus through any filter was observed. A brief example calculation shows why the PREs of the filters used were too large to 60
observe viable penetration. If a nebulization li quid consisting of once-diluted stock (at a titer of 107 CFU/mL) were sprayed, then calcul ating based on the VSF for the entire system, an aerosol of 7.86 CFU/m3 would be measured upstream of the filter. Using Equation 2-2, and assuming that the minimum number of col onies counted from a plate to have reliable data is N = 30, that the sampled aerosol flow rate is Qa = 2.5 L/min, that the initial plates are counted ( n = 0), and that everything else is the same as for the other experiments, the minimal detection limit of the impingers is about 54 CFU/m3. Plugging a minimal downstr eam concentration of 54 CFU/m3 and a maximal upstream concentration of 7.86 CFU/m3 into Equation 1-1 gives a VRE of 99.4%. Therefore one should be able to detect aerosol with an impinger downst ream of a filter with 99.4% VRE or lower, provided a lar ge challenge concentration on the order of 107 CFU/m3. PREs for all of the f ilters tested were all much higher than 99.4% in the size range of B. atrophaeus Therefore, t he penetration of B. atrophaeus can only be measured on a less-efficient filter; alternat ely, a smaller microorganism could be used. Out of the five experiments, the largest CV for upstream airborne viable concentration in a single B. atrophaeus experiment was 26%, and another CV lay slightly outside the goal of 20%. The variabilit y in viability is worse than observed by Henderson,74 who was also using a Bacillus spore and measured a worst-case CV of 10.4%. The data lie outside the criteria to validate the hypothesis, but in retrospect that criterion was overly ambiti ous and not necessary for validating the CATS. The absence of a clear trend suggests that the few high CV s for viability are because of experimental noise rather than systematic decrease in vi ability. Inexperience with microbiological methods also likely contributed to the variabili ty observed in the bacterial spore tests. 61
The airborne viable concentration is st eady enough for use in an animal exposure, and the Collison is suitable fo r creating that challenge. The VSF of the CATS estima ted at the Collison nozzle for B. atrophaeus (2-6) is comparable to the VSFs measured by Henderson74 for B. subtilis at the end of his spray tube (3.5-6 to 4.1-6), suggesting reasonably low loss due to the nebulization method. The regression used to determine the VSF has an R2 higher than 90%, suggesting that the challe nge atmosphere is fairly repeatable as well. Again, the Collison is a suitable method of creating a bacterial bioaer osol for an animal exposure. With B. atrophaeus the TPC or TPC above 0.8 m did not show a correlation with the post-experiment Collison ti ter or the airborne viable concentration. The lack of correlation is unexpected: if there are more microorganisms, there should be more particles, especially at the si ze of that microorganism. The lack of correlation may be because the particles that make up the TP C above 0.8 m are not necessarily viable. As shown in Table 3-2, in Experiment 901 and after, nominal titers were an order of magnitude smaller than the tite r measured post-experiment. Data from this work show the aerosol does not signific antly decrease in viability ov er an experiment, so the difference is not due to losses in the Collison. Experiment 901 and the ones following used a different lot of B. atrophaeus stock than the ones precedi ng. Stocks were titered by the laboratory staf f where this work was performed, but something may have caused loss of viability between when it was first titered and when it was used. The lack of correlation is an inconsistency in laboratory methods rather than a flaw in the CATS. Filter Physical Removal Efficiency Unfortunately, no viable penetration was meas ured through the filter s. Filter media tried in this work were either too effici ent to measure penetration or achieved no 62
removal in the range of interest. For the la ter animal experiment, media with a low PRE (near 97% at a size of ~500 nm) has been specially manufactured by Safe Life, and a smaller microorganism (H1N1 Influenza A, with a particle size on the order of 100 nm) that can be nebulized at higher titers will be used. Not enough pieces of the specialorder media were available for it to be used in this work as well as the animal trials. Because penetration was not measur ed, and the downstream PSD was not consistently measured before and after t he exposure period, the only observable parameter of the filter was its p which did not observably change over the course of any experiment. The filter media used in this work were electrets, composed of polypropylene. Barret and Rousseau30 showed that the behavior of electret polypropylene filters varies widely dependi ng on how the fibers of the media were made, and that some lose PRE without showing a change in p However, Barret and Rousseau were using NaCl and dioctyl pht halate aerosols specifically intended to reduce the PRE of electrets. Dioctyl phthalat e is a strong plasticizer. It and other plasticizers do not appear in a bioaerosol te sts. While salt may appear in a microbial stock, Barret and Rousseau were al so using a challenge of 15 mg/m3 of NaCl particles a far larger mass concentration than encounter ed in a bioaerosol test at similar face velocity for nearly 3 hour s. In Experiment 20090910, fo r instance, the total mass concentration of the chall enge was only about 0.6 mg/m3, and very little of that mass was salt. The bioaerosol challenges that th e CATS is used with likely do not have quite the capacity to reduce PRE t hat Barret and Rousseaus aeros ol challenges did, and no previous studies on PSTI electret medi a have showed such a reduction in PRE. Bioaerosols likely do not reduce the PRE of electret filters significantly. 63
Measurements of PSD were taken downstream before and after loading in only Experiment 20090910 on the IPA-tr eated 1860S medium, which had had its electric charge removed by the IPA treatment. Since t hat filter was no lon ger an electret, it should not be expected to have the reduction in PRE with loading that some electrets have, and the difference between upstr eam and downstream measurements in Figure 3-4 likely is due to experimental error. There is no reason to believe that the PRE changed over the course of the exper iments performed in this work. It should be noted that all of the test cases in this wo rk used a nebu lization liquid that had relatively low concentrations of di ssolved solids, thus causing a lower loading on the filter. The microorganism initia lly chosen for the animal exposure was Francisella tularensis which is not as stable in water and requires a larger amount of dissolved protein content in its n ebulization liquid. Secti on 3.2 of Heimbuch et al .100 details the preliminary work that determined that F. tularensis was not an appropriate challenge for the animal study. When F. tularensis was aerosolized through the test filter, in a setup similar to the CATS, the filter medium wa s rapidly loaded by dissolved solids and its PRE approached 100% quickly. That the PRE of the fabric increased instead of decreased under heavy bioaerosol loading give s credence to the idea that bioaerosols do not have the capacity to significantly reduce the PRE of Safe Lifes filters. Extrapolating to the Delivered Dose That viable penetration by a micrometer-s ized bacterium was not measured does not invalidate the performance of the CATS. Data in Tables 3-1 and 3-3 s how that the upstream challenge is consistent for the duration of the experiments. There is no reason to believe that the filters in this work had any change in their PRE, so one can predict 64
65 that the downstream challenge would also be consistent in tests with filters with lower PREs and not contribute va riability to the dose. As said earlier, simply measuring the ac tual dose received by the animal is, when possible, the best metric for validation. Kaur et al.80 state that in th eir system, in which mice were exposed to aerosolized dry powders of anti-tubercu losis drugs, the dose received had a CV of 13.5% or lower, and the dose was accurate enough that no significant difference was observed between mice dosed intravenously and by the aerosol route. Raabe et al.77 exposed mice to 137Cs aerosol particles and measured a CV for the lung burden among 80 Syrian ham sters of 25%. While an animal exposure was not performed in this work, it can be r easoned that the variability of the dose is driven by the most-variable component of E quation 1-3, which in this work is the airborne viable concentration. The variability of airborne viable concentration in this work is higher than Kaur et al.s variation in dose and slightly high er than Raabe et al.s. Again, what part of the variability in the viable counts in this work is not an artifact of the viability measurement method is unclear. The CVs measured for viable counts are low enough that a dose with that CV or slightly hi gher is acceptable in an animal exposure. The data support a conclusion that the CATS satisfied the key conditions to maintain a consistent challenge: PSDs re mained acceptably constant, airborne viable concentration was fairly consistent, and PRE can be reasoned to have not changed discriminably. One can conclude that CATS is capable of producing an acceptably consistent challenge for dosing animals in an exposure trial. The CATS cannot accommodate every conceivable combination of organism and filter, but its operating envelope is wide enough to enable the PST I-filter animal exposure trials.
CHAPTER 5 CONCLUSION To enable an animal inhalation study t hat will evaluate the effect of an antimicrobial filter on the in fectivity of bioaerosols, an ex perimental system to expose rodents to aerosols that hav e passed through a filter was designed and built, and its mechanical performance was validated. Aeroso l challenges of MS2 coli phage virus and the bacteria Bacillus atrophaeus were created and flowed through the system, and thence through coupons of filter media cut from commercially available FFRs. However, viability of MS2 was not measured because of assay problems, and penetration by B. atrophaeus was too small to quantify. Two commercially available FFRs, the T-5000 and 1860S, have very large PREs near 1 m, too large for use in an animal te st using bacteria. No significant viable penetration was observed in challenge experiments because of these filters high PRE. However, the p across the filters remained const ant, and no sign was found that PRE changed over the course of t he experiments. The upstream PSD was very consistent during these tests, with CVs all less than 10%. The upstream viable airborne concentration of airborne B. atrophaeus was suitably consistent, with CVs of less than 26%, comparable to the literature. This maximu m observed CV is larger than the criteria to validate the hypothesis, but that goal was likely too ambitious. From the data in this work, and r easoning based on the literature, one can conclude that the downstream PSD and viable airborne concentration remain steady for long enough to accurately deliver the challenges needed to perform the animal exposure trials. The CATS can produce a bioaerosol challenge that is sufficiently uniform to support statistically reliable animal infectivity te sting. The CATS provides a 66
67 design for an animal exposure system incorporating aerosol filtration, a capability previously unreported in the literature.
APPENDIX A OPERATING SEQUENCES Leak Check Step 1. Metal stubs of -inch diameter were placed into ports and the fittings were tightened. All ports on the mouse tree were plugged. Step 2. Valve 1 was opened until pressure in the system had reached approximately 0.25 psig. The exact value was not important, but if the pressure was too high, the plugs in the mouse tree would begin to creep out, affecting the reading. The charge neutralizer has a stated maximum pressure of 5 psig. Step 3. The pressure readings were observed. If after one hour or so, the pressure was still what it was initially, the leak check wa s successful. If not, seals and fittings were checked. Pre-Nebulization Preparations Step 1. The system was depressurized and air bl ed from the humidification loop. Air creeps into the humidifying loop over the course of an experiment because of the pressure difference. The air needs to be purged and the tube allowed to completely soak before pressurizing it again. This step was done at least an hour before nebulization, but could be done at any time after the previous experiment. 1. The plug at the end of t he bleed stem was opened. Any water in the line was let to leak out into a small glass. If any ai r was in the line, the flow would stop. 2. A pipetter and a length of tubing were used to draw flow from the bleed stem until siphon pressure caused the water to flow freely. 3. The plug at the end of the bleed stem was closed. Step 2. The filter was inserted into the filter holder. 1. The two halves of the filter holder we re separated by unscrewing the screws using a ball-tipped hex driver (or other hex driver or an Allen wrench). 68
2. The upstream section of the CATS was pulled back, and the downstream section of the filter holder was rotated outwards so that it could be accessed. 3. The filter was inserted. For 47-mm sample s, it was inserted so that the mesh was downstream, the O-ring upstr eam, and the upstream side of the filter material facing the correct direction. From the vantage of looking into the filter holder this appeared as the O-ring in front and the mesh behind. 4. The filter holder was closed. The downs tream and upstream sections were aligned and brought together, and then screwed together with the hex driver. The filter holder was then examined visually to make sure the two halves of the filter holder were level. Step 3. Airflow through the system was begun by turning on the main air to the system Valve A. This air was flowed through the filter to blow off the initial iodine bloom. The system was sampled before the filt er (at 1) with a particle size r to confirm that the CATS was clear of particles. The pressure gauges were checked for correct readings. Aerosol Consistency Trials Step 1. The correct starting valve confi guration was confirmed as so: Off: Valves B, C, D, 1, 2, 4, 5a, 5b. On: Valves A, 3. Step 2. The cap was removed and a filled ne bulizer was attached. The -inch Swagelok nut attaching the nebulizer to the CATS was tightened. The nebulizers pressurized air line was connect ed using its Ultra-Torr fitting. Step 3. The Collison was turned on at Valve D. The impinger dilution air at Valve C was also turned on, as the change in pressure fr om the air going to the impingers can affect the main flow rate by .2 L/min. Valve A was adjusted to produce a flow such that the face velocity through the media was correct. Fo r the tests with PSTI media, this velocity is specified as 7.08 cm/s. For a 47-mm samp le of which 40 mm experiences flow, the corresponding flow rate is 5.3 L/min. Also, t he flow rate of the impinger dilution air was adjusted to an appropriate level fo r the test (10 to 11 L/min). 69
Step 4. The flow was allowed to equilibrate. A period of 15 minutes is standard and appeared to be sufficient. Step 5. After equilibration, upstream (Port 1) and downstream (Port 2) measurements were made with the particle sizer. Step 6. The first impinger sample was begun at the downstream, as follows: 1. The dilution air (Valve C) was turned off. 2. The impinger sampling hose was connect ed to Port 2. Valve 2 was not opened at this step. 3. The impinger was connected to Port 5b and the Ultra-Torr fitting was tightened. Valve 5b was opened. 4. The dilution air (Valve C) was turned on. 5. The vacuum pump was connected to the impinger. 6. Valve 2 was opened, beginning the samp ling. A 5-minute timer was started. Step 7. Sampling with the impingers was perfo rmed for a period of 5 minutes per impinger, alternating upstream and downs tream. To reduce contamination and backflow, the following procedure was used when switching. For simplicity, this sequence is couched as switching from downstream to upstream: for t he other way, just swap Valves and Ports 1 and 2, and 5a and 5b. Particle sizer measurements were taken at points during this step but did not require a special sequence to prevent contamination. 1. Valve 2 was closed. 2. The vacuum pump tube was removed from the impinger. 3. The dilution air (Valve C) was turned off. 4. The impinger sampling hose was disconnec ted from Port 2, and inserted into a HEPA capsule. 5. Valve 5b was closed. 6. The dilution air (Valve C) was brie fly turned back on to purge the impinger sampling hose. A purge was performed for at least 15 seconds to remove remaining aerosol. 7. The dilution air (Valve C) was turned back off. 8. The impinger sampling hose was connec ted to port 2. Valve 2 was not opened. 9. The impinger was connected to Port 5b and the Ultra-Torr fitting was tightened. Valve 5b was opened. 70
71 10. The dilution air (Valve C) was turned on. 11. The vacuum pump was connected to the impinger. 12. Valve 2 was opened. T he timer was started. Step 8. Once impinger sampling was finish ed, upstream (Port 1) and sometimes downstream (Port 2) measurem ents were made with the particle sizer. (This step was not always performed.) Step 9. Collison flow was turned off, the nebulizer removed and the cap replaced. Air was flowed to purge the system. Step 10. If sampling a biological c hallenge, the impinger aliquot and remaining Collison liquid were assayed.
APPENDIX B FLOW RATE, TEMPERATURE, A ND RELATIVE HUMIDITY DATA Temperature, relative humid ity, and flow rate data measured on the CATS over a period of more than an hour are presented in Table B-1. Data begin from the absolute start of beginning flow: there is no equilibration period. Note that the relative humidity starts off low, illustrating the importance of letting the system equilibrate before beginning measurements. The low initial RH measurement is ex cluded from the mean. Data after 65 minutes are taken to show the e ffect of toggling the flow to the impingers, which causes a deviation in flow rate of about 2%. Table B-1. Temperature, RH, and flow consistency data Time (min) T (C) RH Flow (L/min) % deviation from mean T RH Flow rate 0 20.7 44.7% 5.28 0.47% -0.50% 5 20.7 71.5% 5.31 0.47% 1.70% 0.18% 15 20.6 72.4% 5.32 0.22% 3.01% 0.33% 20 20.6 71.5% 5.31 0.13% 1.70% 0.16% 25 20.6 70.8% 5.31 0.08% 0.76% 0.12% 30 20.6 70.4% 5.30 0.03% 0.23% 0.05% 35 20.6 70.1% 5.30 -0.07% -0.25% -0.03% 40 20.6 69.9% 5.30 -0.12% -0.59% -0.05% 45 20.6 69.8% 5.30 -0.16% -0.72% -0.03% 50 20.5 69.6% 5.30 -0.21% -1.02% -0.07% 55 20.5 69.2% 5.30 -0.21% -1.50% -0.03% 60 20.5 69.1% 5.30 -0.31% -1.63% -0.07% 65 20.5 69.1% 5.30 -0.31% -1.67% -0.07% Mean of above 20.6 68.3% 5.30 70 20.6 68.1% 5.43 -0.16% -3.10% 2.35% 75 20.5 68.5% 5.42 -0.21% -2.59% 2.20% 80 20.5 68.9% 5.41 -0.21% -2.02% 1.95% 85 20.5 69.0% 5.40 -0.21% -1.80% 1.86% 90 20.5 68.9% 5.40 -0.26% -1.93% 1.86% 72
APPENDIX C PORT CORRELATION DATA Figure C-1. Representative PSD from nebuliz ing 250-nm beads. Taken from the first set of samples at Port 1. Lines are 95% c onfidence intervals for each diameter. The peak consistently occurred at 241.4 nm and was fairly sharp. 0.0E+00 5.0E+10 1.0E+11 1.5E+11 2.0E+11 2.5E+11 10 100 1000dN/dlogdp (#/m3)Aerodynamic diameter (nm) Table C-1. Readings of particle concentration at ports on Cont rolled Aerosol Test System (CATS) while nebulizing 250-nm beads Sampling port Concentration, 232.9 to 250.3 nm (106 #/m3) Sample mean Deviation from overall mean Port 4 9712 9833 9739 9761 3.16% Port 1 8666 9061 9220 8982 -5.08% Port 2 9384 9228 9295 9302 -1.69% Port 4 9123 9429 9411 9321 -1.50% Port 1 8382 8402 8935 8573 -9.40% Port 2 9205 9418 9288 9303 -1.68% Port 4 9903 10073 10129 10035 6.05% Port 1 9351 9371 9432 9385 -0.82% Port 5a 9836 9575 9570 9660 2.09% Port 5b 9573 9668 9398 9546 0.89% Mouse tree, quad 1 9663 9505 9664 9611 1.57% Mouse tree, quad 2 9009 9975 10052 9679 2.28% Mouse tree, quad 3 9529 9949 10050 9843 4.02% Mouse tree, quad 4 9138 9740 9540 9473 0.11% Overall mean 9463 (Max-min) /mean Minimum 8573 Maximum 10035 15.45% The entire experiment lasted four hour s after aerosol equilibration. 73
74 Figure C-2. Representative PSDs from nebulizing 1-m beads. Data is based on samples taken at quadrants of the m ouse tree. Lines are 95% confidence intervals. A broad, slightly unsymmetr ical peak consistently occurred at 1.037 m. 0.0E+00 5.0E+08 1.0E+09 1.5E+09 2.0E+09 2.5E+09 3.0E+09 3.5E+09 4.0E+09 4.5E+09 0.5 1 2 4d N /dlogd p (#/m3)Aerodynamic diameter (m) Table C-2. Readings of particle concentration at ports on CATS while nebulizing 1-m beads Port Concentration, 0.965 to 1.114 m (106 #/m3) Deviation from mean Port 4 246 -0.68% Port 2 246 -0.39% Port 1 249 0.60% Port 5b 251 1.68% Port 5a 249 0.63% Quad 1 247 -0.14% Quad 2 245 -0.83% Quad 3 242 -1.95% Quad 4 242 -2.11% Quad 4 246 -0.53% Port 2 250 1.04% Port 1 254 2.68% Overall mean 247 Minimum 242 (Max-min)/mean: Maximum 254 4.79% This correlation took only 30 minutes after aerosol equilibration.
APPENDIX D BIOAEROSOL CONSIS TENCY RAW DATA In the tables of PSDs, elaps ed time is the time at the end of the sampling period minus the time the run was started; TPC, CMD, and GSD are the moments calculated by the particle sizer software; mean is t he mean across each row; St. dev. is the standard deviation across each row; and CV is the ratio of st. dev to mean. In the tables for the PSD of B. atrophaeus TPC >0.8 is the concentration of particles with aerodynamic diameter larger than 0.8 m. Note that the magnitude of TPC and TPC >0.8 varies from table to table for B. atrophaeus In the tables of viability the Dilution column indicates n in the dilution ratio 10-n. Because the plating method adds another 1:10 dilution, the reading s from the plates were multiplied by 10, which is the reason all the raw counts end in zero. The collected aerosol flow rate is denoted Qa. Nominal titer is the tite r of the nebulization liquid calculated from its dilution ratios and the ti ter of the undiluted st ock. Except where specified, the volume of nebulizer liquid wa s 30 mL. The lot number is an in-laboratory identifier for each batch of freezer stock of B. atrophaeus All MS2 plates were contaminated or otherwise unusable. No viability data are recorded for B. atrophaeus experiments 827, 903, and 909 because the plat es were contaminated: where viability was not measured, the nominal titer appears in the notes to the PSD table. NR indicates data not recorded. MS2 Table D-1. PSD data for Experiment 724 (MS2) Elapsed time 19:17 21:36 32:04 34:23 45:04 47:23 Mean St. dev. CV TPC (1012 #/m3) 2.74 2.69 2.50 2.51 2.68 2.80 2.65 0.124 00 4.66% CMD (nm) 79.8 0 81.40 81.50 81.90 80.90 82.80 81.40 1.00000 1.23% GSD 1.70 1.70 1.70 1.70 1.71 1.70 1.70 0.00327 0.19% Mean T: NR. Mean RH: 65%. p : 0.8 in H2O. Nominal titer: NR. 75
Table D-2. PSD data for Experiment 728 (MS2) Elapsed time 19:42 22:01 33:50 36:08 46:29 48:47 Mean St. dev. CV TPC (1012 #/m3) 4.86 4.64 4.57 4.36 4.20 4.40 4.51 0.235 0 5.21% CMD (nm) 71.5 0 71.10 76.20 75.40 75.30 75.80 74.20 2.3000 3.09% GSD 1.71 1.71 1.68 1.68 1.68 1.68 1.69 0.0172 1.02% Mean T: 23 C. Mean RH: 61%. p : 0.84 in H2O. Nominal titer: 1012 PFU/mL. Table D-3. PSD data for Experiment 730 (MS2) Elapsed time 17:54 20:13 30:53 33:11 43:35 45:54 Mean St. dev. CV TPC (1012 #/m3) 4.34 4.28 4.00 4.04 4.21 4.10 4.16 0.138 00 3.31% CMD (nm) 74.8 0 75.00 75.20 76.20 76.00 77.20 75.70 0.90700 1.20% GSD 1.68 1.68 1.68 1.69 1.68 1.68 1.68 0.00360 0.21% Mean T: 22 C. Mean RH: 64%. p : NR. Nominal titer: 1011 PFU/mL. Table D-4. PSD data for Experiment 811 (MS2) Elapsed time 16:54 19:12 30:05 32:24 43:40 45:59 Mean St. dev. CV TPC (1012 #/m3) 0.500 0.544 0.514 0.513 0.503 0.503 0.513 0.0162 0 3.16% CMD (nm) 74.1 00 73.900 75.500 75.700 76.900 77.600 75.600 1.46000 1.93% GSD 1.72 0 1.700 1.720 1.720 1.710 1.720 1.720 0.00960 0.56% Mean T: 27 C. Mean RH: 55%. p : NR. Nominal titer: NR. Table D-5. PSD data for Experiment 812 (MS2) Elapsed time 19:05 21:23 32:34 34:52 45:06 47:24 Mean St. dev. CV TPC (1012 #/m3) 4.91 4.71 4.53 4.46 4.22 4.22 4.51 0.270 00 6.00% CMD (nm) 75.9 0 76.20 76.70 76.80 76.70 76.90 76.50 0.37400 0.49% GSD 1.70 1.70 1.70 1.70 1.70 1.70 1.70 0.00144 0.08% Mean T: 27 C. Mean RH: 53%. p : NR. Nominal titer: NR. Table D-6. PSD data for Experiment 813 (MS2) Elapsed time 18:29 20:47 31:18 33:36 43:28 45:46 Mean St. dev. CV TPC (1012 #/m3) 4.30 4.28 4.19 4.22 3.88 3.81 4.11 0.212 00 5.15% CMD (nm) 77.9 0 78.10 78.30 78.20 77.60 77.70 78.00 0.28600 0.37% GSD 1.70 1.70 1.70 1.69 1.70 1.71 1.70 0.00493 0.29% Mean T: 26 C. Mean RH: 56%. p : NR. Nominal titer: NR. B. atrophaeus Table D-7. PSD data for Experiment 819 ( B. atrophaeus ) Elapsed time 14:54 22:27 30:12 37:37 45:42 53:22 60:39 Mean St. dev. CV TPC (107 #/m3) 12.2 0 11.80 11.40 12.00 12.90 12.90 13.10 12.30 0.64200 5.21% TPC >0.8 (107 #/m3) 7.69 7.34 6.92 7.24 7.40 7.35 7.36 7.33 0.229 00 3.12% CMD (m) 1.09 1.08 1.07 1.08 1.07 1.07 1.06 1.08 0.00742 0.69% GSD 1.32 1.32 1.32 1.32 1.33 1.33 1.33 1.33 0.00560 0.42% Mean T: 24 C. Mean RH: 58%. p : 0.9 in H2O. 76
Table D-8. Viability data for Experiment 819 ( B. atrophaeus ) Source Dilution Plate counts Mean Concentration Nebulizer Liquid 4 1460 1440 1320 1407 5 280 300 210 263 2.637 CFU/mL 0 Upstream sample 1 1 1130 1200 840 1057 2 180 130 170 160 2.567 CFU/m3Upstream sample 2 1 840 1050 1050 980 2 160 140 190 163 2.617 CFU/m3Upstream sample 3 1 990 930 940 953 2 130 150 60 113 1.817 CFU/m3Mean of upstream 2.337 CFU/m3 Standard deviation 4.476 CFU/m3CV 19.21% _________ Qa = 2.5 L/min. Nominal titer: 107 CFU/mL, from Lot 07-08-29. Table D-9. PSD data for Experiment 820 ( B. atrophaeus ) Elapsed time 15:09 16:11 23:36 32:34 39:46 47:56 55:02 66:05 Mean St. dev. CV TPC (107 #/m3) 5.31 5.40 5.70 5.80 6.08 6.15 6.35 6.54 5.92 0.439 00 7.42% TPC >0.8 (107 #/m3) 4.56 4.60 4.84 4.87 5.04 5.08 5.17 5.28 4.93 0.259 00 5.26% CMD (m) 1.11 1.10 1.10 1.09 1.09 1.09 1.09 1.09 1.10 0.00652 0.60% GSD 1.20 1.21 1.21 1.21 1.22 1.22 1.22 1.22 1.21 0.00697 0.58% Mean T: 24 C. Mean RH: 57%. p : NR. Table D-10. Viability data for Experiment 820 ( B. atrophaeus ) Source Dilution Plate counts Mean Concentration Nebulizer liquid 4 980 1100 1020 1033 5 90 160 170 140 1.407 CFU/mL Upstream sample 1 1 580 600 520 567 9.076 CFU/m32 90 110 110 103 Upstream sample 2 1 350 420 470 413 6.616 CFU/m32 30 50 180 87 Upstream sample 3 1 370 400 310 360 5.766 CFU/m32 50 40 80 57 Mean of upstream 7.156 CFU/m3 Standard deviation 1.726 CFU/m3CV 24.02% _________ Qa = 2.5 L/min. Nominal titer: 107 CFU/mL, from Lot 07-08-29. Table D-11. PSD data for Experiment 827 ( B. atrophaeus ) Elapsed time 13:02 22:12 29:22 36:32 43:51 51:24 58:30 Mean St. dev. CV TPC (107 #/m3) 2.55 0 2.920 2.920 3.190 3.400 3.190 3.380 3.080 0.3010 9.77% TPC >0.8 (107 #/m3) 0.743 0.721 0.664 0.658 0.653 0.582 0.574 0.656 0.0634 9.66% CMD (m) 0.789 0.727 0.711 0.691 0.681 0.673 0.660 0.705 0.0435 6.18% GSD 1.36 0 1.360 1.360 1.350 1.350 1.350 1.340 1.350 0.0104 0.77% Mean T: 23 C. Mean RH: 63%. p : 0.2 in H2O. Nominal titer: 107 CFU/mL, from Lot 07-08-29. 77
Table D-12. PSD data for Experiment 901 ( B. atrophaeus ) Elapsed time 01:49 08:45 15:38 22:51 30:20 37:39 44:48 Mean St. dev. CV TPC (108 #/m3) 2.63 0 2.360 2.510 2.340 2.460 2.770 2.740 2.540 0.17300 6.81% TPC >0.8 (108 #/m3) 1.200 1.140 1.220 1.180 1.240 1.250 1.260 1.210 0.04410 3.64% CMD (m) 0.948 0.957 0.957 0.961 0.958 0.943 0.948 0.953 0.00670 0.70% GSD 1.33 0 1.330 1.330 1.330 1.330 1.330 1.330 1.330 0.00203 0.15% Mean T: 23 C. Mean RH: 50%. p : 0.26 in H2O. On this experiment, the equilibration time appears to be highly abbreviated, although this may just be a mistake in noting the time. The short equilibration time did not seem to affect the results. Table D-13. Viability data for Experiment 901 ( B. atrophaeus ) Source Dilution Plate counts Mean Concentration Nebulizer liquid 3 1430 1420 1640 1497 4 360 280 200 280 2.806CFU/mL 5 10 20 40 23 Upstream sample 1 0 620 560 730 637 1 240 200 250 230 3.686 CFU/m3Upstream sample 2 0 860 720 630 737 1 210 140 160 170 2.726 CFU/m3Upstream sample 3 0 750 720 680 717 1 160 180 170 170 2.726 CFU/m3Mean of upstream 3.046 CFU/m3 Standard deviation 5.545 CFU/m3CV 18.23% _________ Qa = 2.5 L/min. Nominal titer: 1.67 CFU/mL, from Lot 09-09-01. Table D-14. PSD data for Experiment 903 ( B. atrophaeus ) Elapsed time 13:49 21:29 29:40 37:08 44:36 51:39 59:49 Mean St. dev. CV TPC (108 #/m3) 2.54 0 2.570 2.560 2.370 2.590 2.750 2.740 2.590 0.12900 5.00% TPC >0.8 (108 #/m3) 1.150 1.180 1.190 1.110 1.240 1.250 1.250 1.200 0.05480 4.58% CMD (m) 0.944 0.944 0.946 0.940 0.945 0.940 0.936 0.942 0.00334 0.35% GSD 1.34 0 1.340 1.340 1.340 1.340 1.340 1.340 1.340 0.00111 0.08% Mean T: 23 C. Mean RH: 48%. p : 0.3 in H2O. Nominal titer: 1.67 CFU/mL, from Lot 09-01-09. Table D-15. PSD data for Experiment 908 ( B. atrophaeus ) Elapsed time 15:59 23:57 31:17 38:26 45:29 52:15 59:18 Mean St. dev. CV TPC (108 #/m3) 5.22 0 5.150 5.080 5.180 5.190 5.530 5.600 5.280 0.20300 3.84% TPC >0.8 (108 #/m3) 2.390 2.450 2.480 2.560 2.580 2.630 2.690 2.540 0.10700 4.21% CMD (m) 0.948 0.951 0.953 0.951 0.948 0.943 0.943 0.948 0.00398 0.42% GSD 1.35 0 1.340 1.340 1.340 1.340 1.340 1.340 1.340 0.00297 0.22% Mean T: 22 C. Mean RH: 47%. p : 0.31 in H2O. 78
79 Table D-16. Viability data for Experiment 908 ( B. atrophaeus ) Source Dilution Plate counts Mean Concentration Nebulizer liquid 3 2510 2730 3160 2800 4 520 670 480 557 5.576CFU/mL Upstream sample 1 1 140 210 380 243 6.496 CFU/m32 20 20 0 13 Upstream sample 2 1 250 130 170 183 4.896 CFU/m32 60 30 0 30 Upstream sample 3 1 150 140 150 147 3.916 CFU/m32 10 30 20 20 Mean of upstream 5.106 CFU/m3 Standard deviation 1.306 CFU/m3CV 25.54% _________ Qa = 1.5 L/min. Nominal titer: 47 CFU/mL in 18 mL, from Lot 09-09-01. Table D-17. PSD data for Experiment 909 ( B. atrophaeus ) Elapsed time 12:54 21:01 28:00 35:17 42:24 49:12 56:11 Mean St. dev. CV TPC (109 #/m3) 1.139 1.120 1.138 1.162 1.172 1.313 1.410 1.208 0.110 00 9.12% TPC >0.8 (109 #/m3) 0.561 0.548 0.560 0.576 0.587 0.596 0.634 0.580 0.0289 0 4.97% CMD (m) 0.960 0.956 0.955 0.954 0.955 0.939 0.938 0.951 0.00862 0.91% GSD 1.34 0 1.340 1.340 1.340 1.340 1.340 1.340 1.340 0.00255 0.19% Mean T: 23 C. Mean RH: 47%. p : 0.25 in H2O. Nominal titer: 8107 CFU/mL in 16 mL, from lot 09-09-01. Table D-18. PSD data for Experiment 910 ( B. atrophaeus ) Elapsed time 14:31 22:57 30:01 37:55 44:52 52:22 59:23 Mean St. dev. CV TPC (109 #/m3) 1.206 1.135 1.097 1.201 1.186 1.192 1.199 1.174 0.0414 0 3.52% TPC >0.8 (109 #/m3) 0.589 0.565 0.537 0.608 0.599 0.609 0.620 0.590 0.0291 0 4.94% CMD (m) 0.961 0.962 0.954 0.962 0.959 0.960 0.962 0.960 0.00291 0.30% GSD 1.34 0 1.340 1.340 1.340 1.340 1.340 1.340 1.340 0.00191 0.14% Mean T: 23 C. Mean RH: 45%. p : 0.26 in H2O. Table D-19. Viability data for Experiment 910 ( B. atrophaeus ) Source Dilution Plate counts Mean Concentration Nebulizer liquid 5 250 180 110 180 1.807 CFU/mL 6 50 10 20 27 Upstream sample 1 1 380 430 590 467 1.247 CFU/m32 40 40 30 37 3 0 10 10 7 Upstream sample 2 1 540 530 400 490 1.317 CFU/m32 70 20 30 40 3 10 10 20 13 Upstream sample 3 1 450 490 390 443 1.187 CFU/m32 40 20 80 47 3 10 0 0 3 Mean of upstream 1.247 CFU/m3 Standard deviation 6.225 CFU/m3CV 5.00% _________ Qa = 1.5 L/min. Nominal titer: 87 CFU/mL in 16 mL, from Lot 09-09-01.
LIST OF REFERENCES 1. Hirst, J. M. Bioaerosols: In troduction, Retrospect and Prospect. In Bioaerosols Handbook 2nd ed.; Wathes, C. M.; Cox, C. S., Eds.; Lewis: Boca Raton, FL, 1995; pp 5. 2. Fiegel, J.; Clarke, R.; Edwards, D. A. Airborne Infectious Disease and the Suppression of Pulmonary Bioaerosols. Drug Discovery Today 2006, 11, 51. 3. Tellier, R. Review of Aeroso l Transmission of Influenza A Virus. Emerg. Infect. Dis. 2006, 12, 1657. 4. Lemieux, C.; Brankston, G.; Gitterman, L.; Hirji, Z.; Gar dam, M. Questioning Aerosol Transmission of Influenza. Emerg. Infect. Dis. 2007, 13, 173. 5. Tellier, R. Questioning Aerosol Tr ansmission of Influenza: In Response. Emerg. Infect. Dis. 2007, 13, 174. 6. Brankston, G.; Gitterman, L.; Hirji, Z.; Lemieux, C.; Ga rdam, M. Transmission of Influenza A in Human Beings. Lancet Infect. Dis. 2007, 7 257. 7. Tang, J. W.; Li, Y. Reflection and R eaction Transmission of Influenza A in Human Beings. Lancet Infect. Dis. 2007, 7 758. 8. Tellier, R. Reflection and Reaction Transmission of Influenza A in Human Beings. Lancet Infect. Dis. 2007, 7 759. 9. Lee, V. R. Reflection and Reaction Transmission of Influenza A in Human Beings. Lancet Infect. Dis. 2007, 7 760. 10. Gardam, M.; Lemieux, C. Reflection and Reaction Trans mission of Influenza A in Human Beings Authors' Reply. Lancet Infect. Dis. 2007, 7 761. 11. Mohr, A. J. Fate and Transpor t of Microorganisms in Air. In Manual of Environmental Microbiology; Hurst, C. J.; Knudsen, G. R.; McIner ney, M. J.; Stetzenbach, L. D.; Walter, M. V., Eds.; American Society for Microbiology: Washington, DC, 2007; pp 641. 12. Schaffer, F. L.; Soergel, M. E.; Straube, D. C. Survival of Airborne Influenza Virus: Effects of Propagating Host, Relative Humi dity, and Composition of Spray Fluids. Arch. Virol. 1976, 51, 263. 13. Cox, C. S. Stabi lity of Airborne Micr obes and Allergens. In Bioaerosols Handbook 2 ed.; Wathes, C. M.; Cox, C. S., Eds .; Lewis: Boca Raton, FL, 1995; pp 77. 14. Hinds, W. C. Bioaerosols. Aerosol Technology 2 ed.; Wiley: New York, 1999; pp 394. 80
15. May, K. R. The Collison Nebulizer: Description, Performance and Application. Aerosol Sci. 1973, 4 235. 16. Hinds, W. C. Producti on of Test Aerosols. Aerosol Technology 2 ed.; Wiley: New York, 1999; pp 428. 17. Reponen, T.; Willeke, K.; Ulevicius, V.; Grinshpun, S. A.; Donnelly, J. Techniques for Dispersion of Microorganisms into Air. Aerosol Sci. Technol. 1997, 27, 405 421. 18. Roy, C. J.; Pitt, L. M., Infectious Disease Aerobiology: Aerosol Challenge Methods. In Biodefense: Research Met hodology and Animal Models; Swearengen, J. R., Ed.; CRC: Boca Raton, FL, 2006; pp 61. 19. Heber, A. J. Bioaerosol Particle Statistics. In Bioaerosols Handbook 2 ed.; Wathes, C. M.; Cox, C. S., Eds.; Lewis: Boca Raton, FL, 1995; pp 55. 20. Grinshpun, S. A.; Willeke, K.; Ulevicius V.; Donnelly, J.; Lin, X.; Mainelis, G. Collection of Airborne Microorganism s: Advantages and Disadvantages of Different Methods. J. Aerosol Sci. 1996, 27 (Suppl. 1), S247S248. 21. Hogan, C. J.; Kettleson, E. M.; Lee, M.-H.; Rama swami, B.; Angenent, L. T.; Biswas, P. Sampling Methodologies and Dosage Assessment Techniques for Submicrometre and Ultrafine Virus Aerosol Particles. J. Appl. Microbiol. 2005, 99, 1422. 22. Grinshpun, S. A.; Willeke, K.; Ulevicius, V .; Juozaitis, A.; Terzieva, S.; Donnelly, J.; Stelman, G. N.; Brenner, K. P. Effect of Impaction, Bounce, and Reaerosolization on the Collection Efficiency of Impingers. Aerosol Sci. Technol. 1997, 26, 326 342. 23. Lin, X.; Willeke, K.; Ulevicius, V.; Grins hpun, S. A. Effect of Sampling Time on the Collection Efficiency of All-Glass Impingers. Am. Ind. Hyg. Assoc. J. 1997, 58, 480. 24. Li, C.-S.; Lin, Y.-C. Storage Effects on Bacterial Concentration: Determination of Impinger and Filter Samples. Sci. Total Environ. 2001, 278 231. 25. Tseng, C.-C.; Li, C.-S. Collection Effi ciencies of Aerosol Samplers for VirusContaining Aerosols. J. Aerosol Sci. 2005, 36, 593. 26. Crook, B.; Sherwood-Higham, J. L. Sampling and Assay of Bioaerosols in the Work Environment. J. Aerosol Sci. 1998, 28, 417. 27. Wen, Z.; Yu, L.; Yang, W.; Wang, J.; Zhao, J.; Li, N.; Lu, J.; Li, J. Detection of Viral Aerosols by Use of Real-Time Quantitative PCR. Aerobiologia 2009, 25, 65. 81
28. Cooper, C. D.; Alley, F. C. Fabric Filters. Air Pollution Control: A Design Approach 3 ed.; Waveland: Long Gr ove, IL, 2002; pp 177. 29. Hinds, W. C. Filtration. Aerosol Technology 2 ed.; Wiley: New York, 1999; pp 182. 30. Barret, L. W.; Rousseau, A. D. Aeroso l Loading Performance of Electret Filter Media. Am. Ind. Hyg. Assoc. J. 1998, 59, 532. 31. Heimbuch, B. K.; Proudfoot, E.; Wander, J.; Laventure, G.; McDonald, R.; Burr, E. Antimicrobial Efficiency of Iodinated Indivi dual Protection Filt ers. Proceedings of the Scientific Conference on Chemical and Biological Defense Research, Hunt Valley, MD, Nov 15, 2004; ECBC-SP-020. 32. Heimbuch, B. K.; Hodge, J. M.; Wander, J. D. Viral Penetration of High Efficiency Particulate Air (HEPA) Filters. Proceedings of the 5th ASM Biodefense and Emerging Diseases Research Meeting, Washington, DC, 27 Feb Mar, 2007; 206 (H). 33. US Army Medical Research Institute of Infectious Diseases. Viral Agents. USAMRIID's Medical Management of Biological Casualties Handbook 5 ed.; Darling, R. G.; Woods, J. B., Eds. ; Fort Detrick, MD, 2004; pp 58. 34. Lange, J. H.; Mastrangelo, G. Resp irators and Other Personal Protective Equipment for Health Care Workers. In Emerging Infectious Diseases and the Threat to Occupational Health in the U.S. and Canada; Charney, W., Ed.; CRC: Boca Raton, FL, 2006; pp 42. 35. National Personal Protective Technology Laboratory. Determinat ion of Particulate Filter Penetration to Test Against Liquid Particulates for Negative Pressure, AirPurifying Respirators St andard Testing Procedure; Procedure No. RCT-APR-STP0051, 0052, 0053, 0054, 0055, 0056, Revision 1.1 ; NIOSH: Pittsburgh, PA, 2005. 36. Qian, Y.; Willeke, K.; D onnelly, J.; Coffey, C. C. Perf ormance of N95 Respirators: Filtration Efficiency for Airborne Microbial and Inert Particles. Am. Ind. Hyg. Assoc. J. 1998, 59, 128. 37. Grinshpun, S. A.; Eninger R. M.; Honda, T.; Adhika ri, A.; Heinonen-Tanski, H.; Reponen, T. Filter Perfo rmance of N99 and N95 Facepi ece Respirators Against Viruses and Ultrafine Particles. Ann. Occup. Hyg. 2008, 52, 385. 38. Coffey, C. C.; Lawrence, R. B.; Cam pbell, D. L.; Zhuang, Z.; Calvert, C. A.; Jensen, P. A. Fitting Characteristi cs of Eighteen N95 Filtering-Facepiece Respirators. J. Occup. Environ. Hyg. 2004, 1 262. 39. Ba azy, A.; Toivola, M.; Adhi kari, A.; Sivasubramani, S .; Reponen, T.; Grinshpun, S. A. Do N95 Respirators Provide 95% Pr otection Level Against Airborne Viruses, and How Adequate Are Surgical Masks? Am. J. Infect. Control 2006, 34, 51. 82
40. Ba azy, A.; Toivola, M.; Reponen, T.; Podgr ski, A.; Zimmer, A.; Grinshpun, S. A. Manikin-Based Performance Evaluation of N95 Filtering-Facepiece Respirators Challenged with Nanoparticles. Ann. Occup. Hyg. 2006, 50, 259. 41. Maus, R. Survival of Bacterial and Mold Spores in Air Filter Media. Atmos. Environ. 2001, 35, 105. 42. Rengasamy, S.; Fisher, E.; Shaffer, R. E. Evaluation of the Survivability of MS2 Viral Aerosols Deposited on Filtering Face Piece Respirator Samples Incorporating Antimicrobial Technologies. Am. J. Infect. Control 2010, 38, 9. 43. Committee on the Development of R eusable Facemasks for Use During an Influenza Pandemic. Findings and Recommendations. Reusability of Facemasks During an Influenza Pandemic: Facing the Flu ; National Academies: Washington, DC, 2006; pp 57. 44. Sun, G.; Worley, S. D. Chemistry of Durable and Regenerable Biocidal Textiles. J. Chem. Educ. 2005, 82, 60. 45. Foarde, K. K.; Hanley, J. T.; Veeck, A. C. Efficacy of Antimicrobial Filter Treatments. ASHRAE Journal 2000, 42 (12), 52. 46. Marchin, G.; Fina, L.; Hoff, J. C ontact and Demand-Release Disinfectants. Crit. Rev. Enviro. Sci. Technol. 1989, 19, 277. 47. Price, D. L.; Simmons, R. B.; Ramey, D. L.; Crow, S. A.; Ahearn, D. G. Assessment of Air Filters Treated with a Broad Spectrum Biostatic Reagent. In Indoor Air '93, Proceedings of the 6th International Conference on Indoor Air Quality and Climate, Helsinki, Finland, July 4, 1993; pp 527. 48. Verdenelli, M. C.; Cecchini, C.; Orpianesi, C.; Dadea, G. M.; Cre sci, A. Efficacy of Antimicrobial Filter Treatm ents on Microbial Colonizati on of Air Panel Filters. J. Appl. Microbiol. 2003, 94, 9. 49. Cecchini, C.; Verdenelli, M. C.; Orpianesi, C.; Dadea, G. M.; Cre sci, A. Effects of Antimicrobial Treatment on Fiberglass-Acrylic Filters. J. Appl. Microbiol. 2004, 97, 371. 50. Sullivan, P. D. Evaluation of Low-Pressure Drop Antimicrobial and Hybrid Air Filters ; Technical report AFRL-ML-TY-T R-2006-4526; Air Force Research Laboratory: Tyndall AFB, FL, 2006. 51. Silver, S. Bacterial Silv er Resistance: Molecular Biology and Uses and Misuses of Silver Compounds. FEMS Microbiol. Rev. 2003, 27, 341. 83
52. Tessier, D.; Radu, I.; Filteau, M. Antimicrobial Fabrics Coated with Nano-Sized Silver Salt Crystals. In the Tec hnical Proceedings of the 2005 NSTI Nanotechnology Conferenc e and Trade Show, Anaheim, CA, May 18, 2005; pp 762. 53. Li, Y.; Leung, P.; Yao, L.; Song, Q. W.; Ne wton, E. Antimicrobial Effect of Surgical Masks Coated with Nanoparticles. J. Hospital Infect. 2006, 62, 58. 54. Byeon, J. H.; Yoon, K.-Y.; Park, J.-H.; Hwang, J. Char acteristics of Electroless Copper-Deposited Activated Carbon Fi bers for Antibacterial Action and AdsorptionDesorption of Volatile Organic Compounds. Carbon 2007, 45, 2313 2316. 55. Taylor, S. L.; Fina, L. R.; Lambert, J. L. New Water Disinfectant: an Insoluble Quaternary Ammonium ResinTriiodide Combination that Releases Bactericide on Demand. Appl. Microbiol. 1970, 20, 720. 56. Gilmour, M. N.; Wicksell, E. K. Bactericidal and Sporici dal Activity of a Quaternary Ammonium Resin-Triiodide Complex. Antimicrob. Agents Chemother. 1972, 2 417. 57. Fina, L. R.; Hassouna, N.; Horacek, G. L. ; Lambert, J. P.; Lamber t, J. L. Viricidal Capability of Resin-Triiodide Demand-Type Disinfectant. Appl. Environ. Microbiol. 1982, 44, 1370. 58. Marchin, G. L.; Fina, L. R.; Lambert, J. L.; Fina, G. T. Effect of Resin DisinfectantsI3 and -I5 on Giardia muris and Giardia lamblia Appl. Environ. Microbiol. 1983, 46, 965. 59. Marchin, G. L.; Silverstein, J.; Br ion, G. M. Effect of Microgravity on Escherichia coli and MS-2 Bacteriophage Disinfection by Iodinated Resins. Acta Astronautica 1997, 40, 65. 60. Heimbuch, B. K.; Wander, J. Bioaerosol Challenges to Antimicrobial Surface Treatments: Enhanced Efficacy against MS2 Coli Phage of Air Filter Media Coated with Polystyrene-4-methyltrim ethylammonium Triiodide; Technical report AFRLML-TY-TR-2006-4527; Air Force Research Laboratory: Tyndal l AFB, FL, 2006. 61. Lee, J.-H.; Wu, C.-Y. Evaluation of the Performanc e of Iodine-Treated Biocide Filters Challenged with Bacterial Spores and Viruses ; Technical report AFRL-MLTY-TR-2007-4510; Air Force Research Laboratory: Tyndall AFB, FL, 2006. 62. Lee, J.-H.; Wu, C.-Y.; Wysocki, K. M.; Farrah, S.; Wander, J., Efficacy of IodineTreated Biocidal Filter Media Ag ainst Bacterial Spore Aerosols. J. Appl. Microbiol. 2008, 105 1318. 84
63. Lee, J.-H.; Wu, C.-Y.; Lee, C. N.; Anwar, D.; Wyso cki, K. M.; Lundgren, D. A.; Farrah, S.; Wander, J.; Heim buch, B. K. Assessment of Iodine-Treated Filter Media for Removal and Inactivati on of MS2 Bacteriophage Aerosols. J. Appl. Microbiol. 2009, 107 1912. 64. Ratnesar-Shumate, S.; Wu, C.-Y.; W ander, J.; Lundgren, D.; Farrah, S.; Lee, J.H.; Wanakule, P.; Blackburn, M.; Lan, M.-F. Evaluation of Physical Capture Efficiency and Disinfection Capability of an Iodinated Biocidal Filter Medium. Aerosol Air Qual. Res. 2008, 8 1. 65. Mainelis, G.; Willeke, K.; Baron, P.; Reponen, T.; Grinsh pun, S. A.; Grny, R. L.; Trakumas, S. Electrical Charges on Airborne Microorganisms. J. Aerosol Sci. 2001, 32, 1087. 66. Eninger, R. M.; Adhikari, A.; Reponen, T.; Grinshpun, S. A. Differentiating Between Physical and Viable Penetrations When Challenging Re spirator Filters with Bioaerosols. Clean 2008, 36, 615. 67. Drew, R. T.; Laskin, S. Envi ronmental Inhalation Chambers. In Methods of Animal Experimentation; Gay, W. I., Ed.; Academic: Ne w York, 1973; Vol. 4, pp 1. 68. MacFarland, H. N. Designs and Operationa l Characteristics of Inhalation Exposure Equipment A Review. Fund. Appl. Toxicol. 1983, 3 603. 69. Cheng, Y. S.; Moss, O. R. I nhalation Exposure Systems. In Concepts in Inhalation Toxicology; McClellan, R. O., Henderson, R. F., Eds.; CRC: Boca Raton, FL, 1995; pp 25. 70. Jaeger, R. J.; Shami, S. G.; Tsenova, L. Directed-Flow Aerosol Inhalation Exposure Systems: Applications to Pathogens and Highly Toxic Agents. In Inhalation Toxicology 2 ed.; Salem, H.; Katz, S. A. Eds.; CRC: Boca Raton, FL, 2006; pp 73. 71. Wong, B. A. Inhalation Exposure Syst ems: Design, Methods, and Operation. Toxicol. Path. 2007, 35 3. 72. Stephenson, J. E.; Anderson, A. O. Ethical and Legal Dillemas in Biodefense Research. In Medical Aspects of Biological Warfare; Lenhart, M. K.; Lounsbury, D. E.; Martin, J. W., Eds.; US Army Office of the Surgeon General: Falls Church, VA, 1987; pp 559. 73. Commission on Life Sciences. Guide for the Care and Use of Laboratory Animals ; National Academies Press: Washington, DC, 1996. 74. Henderson, D. W. An Apparatus for the Study of Airborne Infection. J. Hyg. 1952, 50, 53. 85
75. Baumgartner, H.; Coggins, C. R. E. Description of a Continuous-Smoking Inhalation Machine for Exposing Sm all Animals to Tobacco Smoke. Beit. Tab. Int. 1980, 10, 169. 76. Rihn, B.; Kauffer, E.; Mart in, P.; Coulais, C.; Villa, M. ; Bottin, M. C.; Vigneron, J. C.; Edorh, A.; Martinet, N. Short-term Crocidolite Inhalation Studies in Mice: Validation of an I nhalation Chamber. Toxicol. 1996, 109 147. 77. Raabe, O. G.; Bennick, J. E.; Light, M. E. ; Hobbs, C. H.; Thomas, R. L.; Tillery, M. I. An Improved Apparatus for Acute Inhalat ion Exposure of Rodents to Radioactive Aerosols. Toxicol. Appl. Pharm. 1973, 26, 264. 78. Bonnet, P.; Binet, S.; Brand t, H.; Kriech, A. J.; Lafontai ne, M.; Nunge, H.; Morele, Y.; de Groot, P.; Wissel, H .; Castegnaro, M. Inhalation Study on Exposure to Bitumen Fumes. Part 1: Development and Validation of the Equipment. Ann. Occup. Hyg. 2000, 44, 15. 79. Nadithe, V.; Rahamatalla, M.; Finlay, W. H.; Mercer, J. R.; Samuel, J. Evaluation of Nose-Only Aerosol Inhalation C hamber and Comparison of Experimental Results with Mathematical Simulation of Aerosol Deposition in Mouse Lungs. J. Pharm. Sci. 2003, 92, 1066. 80. Kaur, J.; Muttil, P.; Verma R. K.; Kumar, K.; Yadav, A. B.; Sharma, R.; Misra, A. A Hand-Held Apparatus for "Nose-Onl y" Exposure of Mice to Inhalable Microparticles as a Dry Powder I nhalation Targeti ng Lung and Airway Macrophages. Eur. J. Pharm. Sci. 2008, 34, 56. 81. Oldham, M. J.; Phalen, R. F.; Budi man, T. Comparison of Predicted and Experimentally Measured Aerosol Depositi on Efficiency in BALB/C Mice in a New Nose-Only Exposure System. Aerosol Sci. Technol. 2009, 43, 970. 82. Dee, S.; Batista, L.; Deen, J.; Pijoan, C. Evaluation of an Air-Filtration System for Preventing Aerosol Transmission of Porcine Reproductive and Respiratory Syndrome Virus. Can. J. Vet. Res. 2005, 69, 293. 83. Dee, S.; Batista, L.; Deen, J.; Pijoan, C. Evaluation of Systems for Reducing the Transmission of Porcine Reproductive and Respiratory Syndrome Virus by Aerosol. Can. J. Vet. Res. 2006, 70, 28. 84. Dee, S.; Deen, J.; Cano, J. P.; Batista, L.; Pijoan, C. Further Evaluation of Alternative Air-Filtration Systems fo r Reducing the Transmission of Porcine Reproductive and Respiratory Sy ndrome Virus by Aerosol. Can. J. Vet. Res. 2006, 70, 168. 85. Hopkins, S. R.; Drury, L. N. Efficacy of Air Filters in Preventing Transmission of Newcastle Disease. Avian Disease 1971, 15, 596. 86
86. Burmester, B. R.; Witter, L. Efficiency of Commercial Air Filters Against Marek's Disease Virus. Appl. Microbiol. 1972, 23, 505. 87. Hartings, J. M.; Roy, C. J. The Automated Bioaerosol Exposure System: Preclinical Platform Development and a Re spiratory Dosimetry Application with Nonhuman Primates. J. Pharm. Tox. Meth. 2004, 49, 39. 88. Fairchild, G. A. Measurem ent of Respiratory Volume fo r Virus Retention Studies in Mice. Appl. Microbiol. 1972, 24, 812. 89. Gussman, R. A. Note on the Particle Size Output of Collison Nebulizers. Am. Ind. Hyg. Assoc. J. 1984, 45 (3), B-8. 90. Hinds, W. C. Properties of Gases. Aerosol Technology 2 ed.; Wiley: New York, 1999; pp 15. 91. Jaeger, R. J. Modular Re versible-Flow-Past Nasopulm onary Respiratory Dosing System for Laboratory An imals. U.S. Patent 5,297,502, March 29, 1994. 92. Liljas, L.; Valegrd, K. The Structure of Bacteriophage MS2. Semin. Virol. 1998, 1 467. 93. Trouwborst, T.; de Jong, J. C. Interacti on of Some Factors in the Mechanism of Inactivation of Bacteriophage MS2 in Aerosols. Appl. Microbiol. 1973, 26, 252 257. 94. EPA. US EPA Manual of Methods for Virology ; Cincinnati, Ohio, 2001. 95. Claus, D.; Berkeley, R. C. W. Genus Bacillus In Bergey's Manual of Systematic Bacteriology; Sneath, P. H. A.; Ma ir, N. S.; Sharpe, M. E.; Holt, J. G., Eds.; Williams & Wilkins: Baltimore, MD, 1986; Vol. 2, pp 1105. 96. Burton, N. C.; Adhikari, A.; Grinshpun, S. A.; Hornung, R.; Reponen, T. The Effect of Filter Material on Bioaerosol Collection of Bacillus subtilis Spores Used as a Bacillus anthracis Simulant. J. Environ. Monit. 2005, 7 475. 97. Nakamura, L. K. Taxonomic Relationship of Black-Pigmented Bacillus subtilis Strains and a Proposal for Bacillus atrophaeus sp. nov. Int. J. Syst. Bacteriol. 1989, 39, 295. 98. Nicholson, W. L.; Setlow, P. Sporul ation, Germination and Outgrowth. In Molecular Biology Methods for Bacillus; Harwood, C. R.; Cutti ng, S. M., Eds.; Wiley: New York, 1990; pp 391. 99. Taylor, J. R. Propagat ion of Uncertainties. An Introduction to Error Analysis: The Study of Uncertainties in Physical Measurements 2 ed.; University Science: Sausalito, CA, 1997; p 61. 87
88 100. Heimbuch, B. K.; Stone, B. R.; McDona ld, R. S.; Kinney, K. R.; Lumley, A. E. Assaying Benefits of Poly[styrene-4-(trimethylammonium)methyl triiodide] in Respiratory Protection Devices ; Technical report AFRL-RX-TY-TR-2009-4593; Air Force Research Laboratory: Tyndall AFB, FL, 2009; pp 16.
BIOGRAPHICAL SKETCH Brenton Ross Stone was born in 1984 to par ents Suzanne and Michael Stone. He is a native of the central New York town of Dolgeville, and graduated from Dolgevilles James A. Green High School in June of 2002. He began attending the University of Buffalo State University of New York in 2002 as a member of the University Honors Program and graduated summa cum laude in 2006 with the degree of Bachelor of Science in mathematics, concent rating in applied mathematics. After receiving his undergraduate degree, Brenton moved to Panama City, FL to work with Applied Research Associates, Inc ., a contractor that supplies laboratory and technical support to the Air Force Research Laboratory. In 2008 he began distance coursework through the University of Flor ida Department of Environmental Engineering Sciences, with tuition support from his empl oyer. Brenton received the degree of Master of Science in environmental engi neering sciences in May of 2010. 89