1 PERFORMANCE EVALUATION OF MICROWAVE ASSISTED NANOFIBER FILTRATION FOR AEROSOLIZED BACTERIUM By QI ZHANG A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 200 8
2 2008 Qi Zhang
3 To my paren ts, Zhimin Zhang and Jinli Chen ; a nd my wife, Sijia Yu, for their endless love, understanding, and support.
4 ACKNOWLEDG MENTS First and foremost, I would sincerely thank Dr. Chang Yu Wu my supervisory committee chair Without his continuous and enthusiastic support, invaluable encouragement and inspiration, as well as patient help and advice, I would not have a chance to r each this achievement I would also like to thank m y supervisory committee members Dr. Wolfgang Sigmund and Dr. Ben Koopman for their insightful suggestions and constant guidance I really appreciate the help from Hyoungjun Park, who fabricated and delive red all the nanofiber mats for my research project. I would like to thank James Welch who has been assisting me for almost one year in my research project. He has given me a big hand in running the filtration tests and conducting the microbiological analy sis I would also like to thank Brian Damit, who helped me a lot in the microbiological experiments since this summer. Assist ance from the Major Analytical Instrumentation Center (MAIC) of the University of Florida for using the Wyko optical profilometer a nd Scanning Electron Microscope are also highly appreciated. I must acknowledge as well all other fellow s in our air resources group who kindly offer help to my research. Financial support from Defense Threat Reduction Agency, Joint Science and Technology Office, Chemical and Biological Program, Contract number HDTRA1 01 C 0064 for this research is also gratefully acknowledged Finally, I would like to express my appreciation to my parents in China and my wife with me here. Without their endless love and su pport, I can not fulfill this achievement.
5 TABLE OF CONTENTS Page ACKNOWLEDGMENTS ................................ ................................ ................................ .................... 4 LIST OF TABLES ................................ ................................ ................................ ................................ 6 LIST OF FIGURES ................................ ................................ ................................ .............................. 7 ABSTRACT ................................ ................................ ................................ ................................ .......... 8 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ ...................... 10 1.1 Background ................................ ................................ ................................ ......................... 10 1.2 Bioaerosol ................................ ................................ ................................ ........................... 10 1.3 Filtration ................................ ................................ ................................ .............................. 11 1.4 Nanofiber Filtration ................................ ................................ ................................ ............ 15 1.5 Microwave Sterilization ................................ ................................ ................................ ..... 16 2 OBJECTIVE ................................ ................................ ................................ ............................... 22 3 EXPERIMENTS ................................ ................................ ................................ ......................... 23 3.1 Testing Filters ................................ ................................ ................................ ..................... 23 3.2 Phase I: Physical Collection ................................ ................................ ............................... 24 3.3 Phase II: Biological Inactivation ................................ ................................ ....................... 26 4 RESULTS AND DISCUSSION ................................ ................................ ................................ 34 4.1 Physical Characterization of Nanofiber Filters ................................ ................................ 34 4.2 Static On Filter Inactivation ................................ ................................ ............................... 41 4.3 Dynamic In Flight Inactivation ................................ ................................ ......................... 42 5 SUMMARY AND RECOMMENDATIONS ................................ ................................ ........... 66 LIST OF REFERENCES ................................ ................................ ................................ ................... 68 BIOGRAPHICAL SKE T CH ................................ ................................ ................................ ............. 77
6 LIST OF TABLES Table page 1 1 Relative complex permittivity and loss tangent of select materials ................................ .... 21 4 1 Physical test results of ACF mats ................................ ................................ .......................... 45 4 2 Physical test results of single sandwiched PAN filters with comparison to two conventional HEPA filters and the military standard ................................ .......................... 47 4 3 P hysical test results of multi sandwiched PAN filters with comparison with two single sandwiched ones ................................ ................................ ................................ ..................... 49 4 4 Comparison of filter quality factors among our results and those of other researches ...... 51 4 5 Percentages of E scherichia coli concentration on each quadrant ................................ ....... 56 4 6 Survival fraction of E scherichia coli under microwave (50 0 W) irradiation Static on filter inactivation ................................ ................................ ................................ .................... 57 4 7 Comparison of the Escherichia coli inactivation by microwave irradiation among our static on filter experiments and those of other research es ................................ ................... 58 4 8 Escherichia coli in flight inactivation percentages under 500 W continuous microwave without a filter ................................ ................................ ................................ .................... 60 4 9 O n filter Esc herichia coli survival fraction of the dynamic in flight tests ......................... 61 4 10 Exponential regression of the Escherichia coli survival fraction on the filter versus the microwave power level of the dyna mic in flight experiment ................................ ............. 63 4 11 V iable Escherichia coli penetration fraction of the dynamic in flight tests ....................... 64
7 LIST OF FIGURES Figure page 1 1 C omplex permittivity in the complex plane ................................ ................................ ......... 20 3 1 Tested filters ................................ ................................ ................................ ........................... 31 3 2 Experimental set u p for physical tests and static on filter inactivation tests ..................... 32 3 3 Experimental set up for dynamic in flight inactivation tests ................................ .............. 33 4 1 Background particle concentration of nebulized water and nebulized ethanol solution ... 46 4 2 Thickness characterization of PAN nanofiber mats by Wyko optical profilometer. ......... 48 4 3 N anofiber mat stacking ................................ ................................ ................................ .......... 50 4 4 S canning electron microscope images of PAN nanofiber mat Millipore HEPA filter and LydAir HEPA filter ................................ ................................ ................................ ................ 52 4 5 Particle size distribution of the NaCl aerosol ................................ ................................ ...... 53 4 6 P enetration curve of PAN 30 nanofiber filter ................................ ................................ ....... 54 4 7 Comparison of the most penetrative particle size among our results and those of other researches ................................ ................................ ................................ ................................ 55 4 8 S canning electron microscope image of Escherichia coli on th e glass slide ..................... 59 4 9 Escherichia coli survival fraction on the filter after dynamic in flight microwave irradiation ................................ ................................ ................................ ................................ 62 4 10 Viable Es cherichia coli penetration fraction from the microwave/filtration system ......... 65
8 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Mast er of Science PERFORMANCE EVALUATION OF MICROWAVE ASSISTED NANOFIBER FILTRATION FOR AEROSOLIZED BACTERIUM By Qi Zhang December 2008 Chair: Chang Yu Wu Major: Environmental Engineering Sciences Aerosolization of biological agents poses one of the bigg est threat s of biological attacks. Besides, natural pandemics such as SARS and a vian flu transmitted by aerosolized pathogens have kept threatening the public health since the beginning of this century. Bioaerosol agents can be dispersed over a wide area i n a short period of time. In this study, a RHELP (Regenerative High Efficiency Low Pressure) purification system for building air ventilation system was developed to provide protection from exposure to pathogenic bioaerosols. The system utilizes a combinat ion of nanofiber filtration and microwave sterilization. The quality of the RHELP system was characterized by two phases of experiments. Phase I focused on the physical collection efficiency of the system. Polyacrylonitrile (PAN) nanofiber electrospun on activated carbon fiber (ACF) mats served as the filtration media. An ACF/PAN/ACF sandwich structure was used to protect the PAN layer from detachment. Filtration experiments were carried out to evaluate the sandwiched PAN nanofiber filters with different electrospinning deposition time s Results showed that three out of the four tested PAN filters achieved a higher filter quality factor than 0.0203 per Pa, the military standard (MIL F 51079D for HEPA filters. Interestingly, the filter quality of shorter te rm electrospun PAN filters
9 was found to be higher than those of longer term electrospun ones. Furthermore, multi sandwiched filters [ n layers of PAN nanofiber mats with ( n + 1) layers of ACF mats, n > 1] showed higher filter quality factors than those sing le sandwiched ones. The MPPS (most penetrative particle size) of the PAN nanofiber filters was also studied. Results showed that the MPPS had a shift to the smaller size compared to a normal 100 to 500 nm range due to the smaller fiber size. In P hase II, microwave irradiation was applied to the system. Results of static on filter inactivation tests showed that the survival fraction of E scherichia coli collected on the filter became below the detection limit when exposed to 500 W microwave irradiation for l ess than 90 seconds. Regarding dynamic in flight microwave irradiation, the Escherichia coli survival fraction was below 0.0 2 % no matter the microwave was applied to system continuously (10 min per 10 minute cycle) or periodically (5 or 2.5 min per 10 minu te cycle) when the microwave power was set at 500 W A n exponential correlation between the survival fraction and the microwave power applied was also observed. The experimental results demonstrated the advantage of the RHELP system for removal and inacti vation of biological aerosol agents. Future research may include characterizing of other filtration media and challenging the system with other types of biological agents such as viruses or spores.
10 CHAPTER 1 INTRODUCTION 1.1 Background Concerns about bio terrorism have been heightened after the anthrax attack in 2001. I ntentional use of infectious agents to hurt enemies has a long and disreputable history. Back to the 14 th century, the Mongolian troops flung corpses of bubonic plague victims over the city walls, and as a result of the pandemic that ensued, the defenders surrender ed (Schaechter et al., 2006). Aerosolized anthrax sent out during the 2001 attacks was easily inhaled and highly concentrated (Rega, 2001) so as to cause serious injury or death. L arge amounts of biological agents can be dispersed over a wide area in a very short period of time and one can easily contaminate an entire building by spray ing his/her bioaerosol into the ventilation system : a cheap medical nebulizer can be purchased by less than $2. Therefore, aerosolization is one of the most effective methods available for biological attacks (Henderson, 1998; Kortepeter and Parker, 1999). Owing to the outbreak of SARS (Severe Acute Respiratory Syndrome) in 2003, the long lasting threat of avian flu these years, and the increasing occurrence of asthma and respiratory tract disease, the need for controlling exposure from natural bioaerosol s has also increased rapidly (Agranovski et al., 2004; Douwas et al., 2003; Gustavsson, 1999; Pastusz ka et al., 2000; Ratnesar Shumate et al., 2008; Yu et al., 2008). 1.2 Bioaerosol Airborne particles of biological origins including viable bacteria, viruses, fungi and algae, as well as non viable pollen, endotoxin, mycotoxin and allergens are all consider ed as bioaerosols (Cox and Wathes, 1995 ; Hinds, 1999 ). They can vary in size from submicron to approximately a few hundred microns. More specifically, bacteria and fungal spores have sizes
11 ranging from 0.3 to 30 micrometers; viruses are smaller, ranging fr om 20 to 300 nanometers (Hinds, 1999). Since air is not a suitable growth medium, live bioaerosols must have originated somewhere else. Coughing, sneezing or even talking by human beings can generate a large amount of bioaerosols into the ambient air. Non human origins of bioaerosols include winds, waves, wastewater treatment plants, crop spraying and so on. Dusts and other forms of airborne debris act as routes of bioaerosol transmission. Bioaerosols can adhere to them and become re suspended during any di sturbances (Hinds, 1999; Prescott et al., 200 6 ). Exposure to bioaerosols has been a subject of concern over the recent decades due to the adverse health effects. Researchers have reported that exposure to concentrated airborne microorganisms is often assoc iated with allergic diseases as well as infections (Dales et al., 1991; Jo and Seo, 2005; Ren et al., 1999), such as aspergillosis in immun o compromised individuals (Millner et al., 1994; Millner, 1995), asthma and allergic rhinitis (Beaumont, 1988; Cockcr oft et al., 1977; Zuskin et al., 1994), extrinsic allergic alveolitis (Farmer s Lung) (Flannigan et al., 1991), chronic obstructive pulmonary disease (COPD) (Clapp et al., 1994; Lacey and Crook, 1988; Matheson et al., 2005), pneumonitis (Lacey and Crook, 1988; Siersted and Gravesen, 1993), upper airway irritation/mucous membrane irritation (Dutkiewicz, 1997; Herr et al., 2003; Wouters et al., 2006) and sick building syndrome (ACGIH, 1989; Dales et al., 1991). 1.3 Filtration F iltration is one of the most co mmonly applied methods for aerosol sampling and air purification. Its application spreads across a wide range of disciplines, including respiratory protection, air purification of smelter effluents, processing of nuclear and hazardous materials,
12 and clean rooms (Hinds, 1999). Due to their simple structure and low cost, fibrous air filters are among the most commonly used in the air purification processes (Yun et al., 2007). There are five basic mechanisms in aerosol filtration, including inertial impaction, interception, Brownian diffusion, enhanced collection of diffusing particles due to interception, and gravitational settling (Hinds, 1999). Particles larger than about 0.5 m are effectively collected due to impaction and interception, and those smaller t han about 0.1 m efficiently deposited due to diffusion. However, there is a size range of particles that are not collected as efficiently as other sized particles (Hinds, 1999; Podgorski et al., 2006). F ibrous filters achieve minimum collection efficiency at a so called most penetrative particle size (MPPS), which is usually between 0.1 and 0.5 m. Podgorski et al. (2006) analyzed the effect of the filter fiber diameter on the MPPS and found that the MPPS decreases as the fiber diameter decreases. Other th an particle size, many other aspects which affect the filter collection efficiency have been reported. T he filtration efficiency increases as the air face velocity decreases (Hinds, 1999). Boskovic et al. (2008) pointed out that the increased tendency of p article rebound from the filter because of the high kinetic energy as well as the reduced retention time of the particles in the filter caused the decrease in the filter collection efficiency. Boskovic et al. (2005; 2008) also studied the effect of particl e shape on filtration performance. They compared the filtration efficiency of spherical particles ( polystyrene latex, PSL) and cubic particles (MgO) and found the spherical particles are more effectively collected than cubic ones. They acknowledged the dif ference to the forms of particle motion on the collection surface. Spheres can slide or roll whereas cubes can slide or tumble, which is more vigorous and likely to rebound. Electrostatic force is another factor that affects the filtration efficiency. Kim et al. (2006) showed charged particles are easier to be collected than uncharged ones. Chen and Huang (1998) found that
13 charges on the filter fibers also improve the filtration efficiency. Wang (2001) summarized that application of electrostatic forces can improve the filtration efficiency, especially for the particles in the MPPS ranges. He also stated that the filtration efficiency improvement due to electrostatic forces depends on the chemical compositions of particles and fibers, the charges on particle s, the surface charge density of fibers, and the intensity of the externally applied electric field. Filter quality factor ( q F ), also known as the figure of merit (FOM) ( Wang et al., 2008a ), is a useful criterion for comparing different types and thickness es of filters (Dhaniyala and Liu, 1999; Hinds, 1999; Wang et al., 2008a ). The quality factor can be calculated using Equation 1 1 : (1 1) where P is the particle penetration (fraction of particles which get through the filter), and p is the pressure drop across the filter. Particle penetration is dependent on many factors, such as the filt er s packing density, the filter thickness, and the fiber diameter, etc. (Hinds 1999) The following equation shows the relationship between particle penetration and filter thickness as well as other parameters: (1 2) where is the packing density of the filter (the ratio of fiber volume to total volume in a particular filter ) also known as solidity E is the single fiber efficiency, t is the filter thickness, and d f is the fiber diameter. Pressure drop a rises due to the resistance to air flow through the filter media. If filters whose fiber size is smaller than 1 m are to be considered, slip effect s should not be neglected ( Wang et al., 2008a ). The expression of pressure drop based on Kuwabara flow with slip effect is (Brown, 1993): (1 3)
14 where is the air viscosity, U is the face velocity through the filter, and Kn (= 2 / d f ) is the Knudsen number with the mean gas free path (= 0.066 m at STP) (Hinds, 1999). From equations 1 1 to 1 3, it is obvious that q F is independent of t for a uniform filter, in which and d f can be considered as constants. It is generally recognized that the enhanced collection of diffusi n g particles due to interception and gravitational settling are the two weakest mechanisms among the five basic filtration mechanisms Thus, it is rational to credit only diffusion, interception and impaction for analyzing the single fiber efficiency. Pich (1965) gave the following expression for the single fiber efficiency due to diffusion ( E D ): (1 4) where Ku (= 0.5ln 0.75 + 0.25 2 ) is the Kuwabara hydrodynamic factor, and Pe (= Ud f / D ) is the Peclet number with the particle diffusion coefficient ( D ). The equa tion for the single fiber efficiency due to interception ( E R ) was given by Pich (1966): (1 5) where R is the ratio of the particle diameter to the fiber diameter ( d p / d f ). Though it is hard to obtain an analytical expression for the single fiber efficiency due to impaction ( E I ), an empirical equation (Landahl and Hermann, 1949; Thom, 1933) can be applied ( Wang et al., 2008a ): (1 6) where St is the Stokes number, which can be calculated as (Hinds, 1999): (1 7) where p is the particle density, and C c is the slip correction factor.
15 1.4 N anofiber Filtration Nanofiber media is a promising media which provides a great filtration efficiency and better performance than conventional fibers (Wang et al., 2008a). The nonwoven industry generally considers fibers with a diameter no larger than 0.5 m as nanofibers (George, 2007). Wang et al. (2008a) plotted the filter quality factor as a function of particle size according to equations 1 1 to 1 7 by inserting = 0.05, U = 10 cm/s, and four different d f values (0.15, 0.5, 5 and 20 m) According to the plot [Figure 1 in Wang et al. (2008a)] the y found that the filter quality factor increases as the fiber size decreases for particles larger than 0.3 m. However for particles smaller than 0.02 m, the quality factor decreases as the fiber size decreases. Compared to microfibers (fibers with d f > 0.5 m) nanofibers perform better for particles larger than about 0.1 m but worse for particles smaller than about 0 .1 m. Wang et al. (2008a) explained that the reason is when the fiber size decreases, the pressure drop increases much faster than the single fiber efficiency due to diffusion. They also plotted another graph of quality factor versus particle size by inse rting a constant fiber size (0.15 m) and multiple packing density and face velocity values [Figure 2 in Wang et al. (2008a)] Increased packing density gives a higher pressure drop, and the increase of pressure drop overweighs the increase of the single f iber efficiency due to diffusion of smaller particles. On the other hand, the increase of single fiber efficiency over come s the increase of pressure drop of larger particles ( Wang et al., 2008a ). Therefore, lower packing density leads to a slightly higher quality factor for particles smaller than about 0.1 m and vice versa. Through equation s 1 1 to 1 3, it is clear that the filter quality factor decreases as the face velocity increases. This trend is clearly shown in Figure 2 of Wang et al. (2008a) A simi lar trend was also found by Dhaniyala and Liu (1999).
16 Nanofibrous media have small fiber size so that they integrate the advantage of small pore size as well as large surface collection area. Besides, low basis weight, high permeability also make nanofiber filtration a promising application (Barhate and Ramakrishna, 2007). Several applications of nanofibrous filtration was summarized in the review of Barhate and Ramakrishna (2007), including engine air filtration, cabin air filtration and self cleaning air intake for gas turbines (Grafe et al., 2001), pulse clean cartridges in dust collection (Kosmider and Scott, 2002), penetrating aerosol particulate filtration (Podgorski et al., 2006), high efficiency particulate air filtration (Ahn et al., 2006), cigarett e filter for smoke filtration (Squires and Gardiner, 2005), catalytic gas filtration for respirators (Ramaseshan et al., 2006) and biocatalytic filtration (Kim et al., 2005; Salalha et al., 2006). Currently, there are three major techniques for producing n anofibers. They include electrospinning, multi component fiber spinning and improved modular melt blowing (Ward, 2005). Among the three methods, electrospinning (also called electrostatic spinning) is the most versatile one (Barhate and Ramakrishna, 2007; Huang et al., 2003; Reneker and Chun, 1996; Subbiah et al., 2005). Doshi and Reneker (1995) defined that the main principle in electrospinning is to generate a charged jet of polymeric solution by an electric field. W hen the jet travels in the air, the sol vent evaporates and a charged fiber is left and collected on a grounded plate. 1. 5 Microwave Sterilization Microwave is a form of energy that falls between the infra red and radio frequencies of the electromagnetic spectrum its frequency ranges from 300 M Hz to 300 GHz (Jones et al., 2002; Zhang and Hayward, 2006) Microwave was first developed for communications before the WWII. During the WWII, it began to be applied for radar detection. Microwave radiation as a
17 remote heating source was then developed su bsequently and widely applied in domestic microwave ovens and industrial systems (Jones et al., 2002; Tao, 2006; Zhang and Hayward, 2006) Domestic and industrial microwave ovens usually operate at a frequency of 2.45 GHz, which corresponds to a 12.2 cm wa velength or a 1.02 10 5 eV energy (Jacob et al., 2005). Volumetric heating, selective heating, and hot spot effect make microwave heating distinguishable from the conventional heating (Zhang and Hayward, 2006). Volumetric heating is defined as energy tha t is transferred from the electromagnetic field to thermal energy throughout the entire volume of the material penetrated by the microwave (Will et al., 2004). In other words, the entirety of the material exposed to the microwave irradi ation is heated simu ltaneously. Selective heating arises because different materials have different microwave energy absorbing abilities (Jones et al., 2002). A material s ability to absorb microwave is characterized by the complex dielectric permittivity ( F/m). This compl ex number can be expressed as (Will et al., 2004): (1 8) where is the dielectric constant which represents the energy stored in the material and is a characterization of the ability for electromagnetic energy to penetrate t he material is the dielectric loss which represents the ability of material to dissipate the electric field (Allan et al., 1980; Ma et al., 1997; Will et al., 2004), and i is the imaginary unit. Hence, the ability of a material with dielectric loss to convert electromagnetic energy into thermal energy can be characterized by the ratio of to (Bonnet et al., 2004), which represents the tangent of the angle between the complex permittivity vector and the positive real axis (tan ) in the complex pla n e Figure 1 1 shows the complex permittivity and relative concepts in the complex plane Among various materials, silicon carbide (SiC) and carbon are the most ideal microwave
18 absorbers, while materials such as pure quartz are transparent to microwave ener gy. Table 1 1 shows a list of the complex permittivity of s elect materials. Hot spots are some small sites inside the material having much higher temperature than the bulk of the material under the microwave irradiation. Certain reactions requiring high te mperature can occur on those sites with relatively lower bulk temperature so as to save energy (Zhang and Hayward, 2006). Since the mid 1980s, the microwave induced/assisted reactions have been studied and widely applied in the industry (Gedye et al., 1986 ; Giguere et al., 1986; Zhang and Hayward, 2006). Among these microwave assisted reactions, the use of microwave irradiation for killing pathogenic microorganisms is attracting significant attention (Apostolou et al., 2005; Awuah et al., 2005; Canumir et a l., 2002; Celandroni et al., 2004; Kiel et al., 2002; Kindle., 1996; Pellerin, 1994; Sasaki et al., 1998; Vaid and Bishop, 1998) because of its high efficiency and low cost. It should be noted that most of these studies about microwave sterilization were d one in liquid or aqueous phase. None of them has been carried out for air filtration. Some researchers, especially in earlier research studies believe that the microwave inactivates the microbes only by its thermal effect. Goldblith et al. (1967) studied the effect of microwave irradiation on Escherichia coli ( E. coli ) and Bacillus subtilis and did not detect any non thermal effects. Jeng et al. (1987) found no significant non thermal effects in a dry microwave sterilization process. Fujikawa et al. (1992) investigated the kinetics of E. coli destruction by microwave irradiation and found the difference between the microwave exposure and conventional heating was not remarkable. Zhang and Hayward (2006) pointed out many hypothesized microwave non thermal eff ects only supported by the difference in reaction/inactivation temperature between microwave and conventional heating (Yaghmaee and
19 Durance, 2005; Zielinski et al., 2007) were not solid due to possible hot spot effects. However, other studies, especially r ecent ones, did reveal the existence of the microwave non thermal effect during the inactivation of microorganisms. Some research studies distinguished the non thermal microwave effect by damaged cells (Campanha et al., 2007; Celandroni et al., 2004), dist orted membrane structure and function (Persson et al., 1992; Phelan et al., 1994), increased release of Ca 2+ K + DNA, dipicolinic acid and protein (Campanha et al., 2007; Celandroni et al., 2004; Vaid and Bishop, 1998; Woo et al., 2000) and altered enzyme activities (Dreyfuss and Chipley, 1980) after the microwave irradiation. Betti et al. (2004) acknowledged the disruption of weak bonds of protein active forms, enhanced production of reactive oxygen species, and cell signaling pathway interference to micr owave non thermal effects. Wantanabe et al. (2000) found the ionic strength can affect the microwave inactivation rate of microbes and linked the effect to the current in the cells generated by microwave.
20 Figure 1 1 C omplex permittivity in the complex plane
21 Table 1 1 R elative complex permittivity and loss tangent of select materials ( Baeraky, 1997; M a et al., 1997) Material / 0 a / 0 a tan Teflon 2.045 .111 .054 Cordierite 2.873 .138 .048 Al 2 O 3 3.006 .170 .057 SiO 2 3.066 .215 .070 TiO 2 7.020 .430 .061 ZrO 2 4.214 .186 .044 Fe 2 O 3 7.420 .219 .030 CuO 3.294 .170 .052 V 2 O 5 7.635 .504 .066 Carbon 10 3 .3 SiC b 6 .5 .08 a 0 is the dielectric constant of vacuum, therefore the ratios ( / 0 and / 0 ) here are the relative dielectric constant and relative dielectric loss of a certain material, respectively b Data of Baeraky ( 1997) other data are from Ma et al. (1997)
22 CHAPTER 2 OBJECTIVE The goal of this research is to develop a lab scale RHELP (Regenerative High Efficiency Low Pressure) building ventilation air purification system which integrates nanofiber filtrat ion and microwave sterilization in order to effectively collect and inactivate biological agents in the air. Specifically, the objective s are to evaluate if this lab scale RHELP system can do the following : Achieve > 99.97% physical collection efficiency f or 300 nm particles, as the military standard MIL F 51079D regulated for HEPA (High Efficiency Particulate Air) filters Have a pressure drop lower than 200 Pa, 50% of the military standard MIL F 51079D for removal of aerosols Effectively inactivate viabl e Escherichia coli ( E. coli ), the biological agents used in this study, collected in the system
23 CHAPTER 3 EXPERIMENTS Experiments were conducted in two phases. In phase I, filters were tested for physical collection efficiency and pressure drop; inert aerosols were tested to determine the filter quality. Phase II experiments demonstrated the ability of microwave to inactivate collected biological agents and to evaluate the time and energy needed for complete inactivation. 3.1 Testing Filters In this stu dy, electrosp un PAN (polyacrylonitrile) nanofiber mats of different thickness es served as the filtration media. However, the PAN nanofiber mat s were too thin to have the mechanical strength commonly found in conventional filters. Since activated carbon fib er (ACF) mats are cheap, easy to obtain and have very low pressure drop against air flow, these PAN nanofibers were deposited on ACF mats by electrospinning. Still, they could easily detach from the ACF mats and became stuck to the filter hold er structure. To avoid this, a sandwich structure (another layer of ACF mat covers the existing PAN nanofiber deposited ACF mat) was used in the experiments. Figure 3 1 shows the testing filters in this study. These PAN nanofiber mats were fabricated and provided by Dr Wolfgang Sigmund s lab in the Department of Materials Science and Engineering at the University of Florida. In order to have different fibermat thicknesses, different electrospinning time s were used. ACF/PAN/ACF single sandwich ed filters with different e lectrospinning deposition times as well as two conventional HEPA (High Efficiency Particulate Air) filters including a Millipore glass fiber HEPA filter (Cat. No.: AP1504700, Millipore, MA) and a LydAir MG High Alpha HEPA filter (Grade 4450H, Lydall Filtr ation, CT), were tested and compared for their filtration performance. Also, a structure of only two ACF mats was tested as a control. Moreover, multi sandwiched structure ( n layers of PAN nanofiber layers laid in between n+ 1 layers of ACF mats)
24 were teste d for an alternative way to improve the filtration quality of the system. A Wyko optical profilometer was used to evaluate the thickness of the PAN nanofiber layers. For biological tests, silicon carbide mats took the place of ACF mats due to their durabil ity under the microwave irradiation. 3.2 Phase I : Physical Collection In this phase, the physical collection efficiency and the pressure drop of different filters were tested. The experimental set up is shown in Figure 3 2. The cylinder air provided clean, dry air with stable flow to the system. The air flow was first split into two ways. The first one l ed to the six jet Collison nebulizer (Model CN25, BGI Inc., MA) with 5.5 LPM flow rate to generate the testing aerosols. NaCl solutions and polystyrene late x (PSL, monodisperse, 300 nm, Duke Scientific) suspensions were used in the nebulizer. The second flow went to the dilution dryer to dry the aerosols generated from the nebulizer; this flow rate was set at 11 LPM. After passing through the dilution dryer, the air flow was split into two different paths. The first flow went through the filter holder with a 4.1 LPM flow rate on the filter surface. This flow rate correspond ed to a face velocity of 5.3 cm/s (the minimum face velocity required to test HEPA filte rs according to the military HEPA standard MIL F 51079D) for 47 mm filters (effective diameter 40.5 mm due to the filter holder structure). Pressure drop was measured by a Magnehelic differential pressure gauge (Model 2010, Dwyer Instruments, IN). There we re two ports sampling inlet and outlet aerosols of the filters for particle size distribution analysis via a Scanning Mobility Particle Sizer (SMPS) (Model 3936, TSI Inc., MN). The second split flow from the dilution dryer l ed to the exhaust; there was a v alve in this path for controlling the flow distribution of the two paths after the dilution dryer and the total air resistance of the entire system.
25 According to the military HEPA standard MIL F 51079D, the filter s physical collection efficiency should be carried out at 300 nm particle size. Therefore, 300 nm monodisperse polystyrene latex (PSL) particles (Cat. No.: 5030, Duke Scientific, CA) were suspended in deionized water to produce a concentration of 2 g PSL/L. Moreover, same volume of pure ethanol wa s added into the suspension to ensure dispersion of the PSL particles. This PSL/DI/EtOH suspension had an aerosol of over 4 10 5 particle/cm 3 concentration by the aerosol generation system and flow rate settings mentioned above. This concentration would b e enough for filtration tests. Besides, 50% (volume ratio) ethanol solution was nebulized in the system. The particle concentration generated from this solution was measured for a background line in order to eliminate any particle concentration contributed by the cylinder air and the ethanol/water dispersant Then, several different PAN filters and two HEPA filters mentioned before were tested in the aforementioned experimental system. The PAN nanofiber filter with a particular electrospinning deposition ti me which will result in the highest filter quality factor ( q F ) would be then identified, and nanofiber multi sandwich structure of that certain time deposited PAN nanofiber filter could be applied to ensure enough physical collection efficiency at the lowe st pressure drop. For single sandwiched PAN filters and the HEPA filters, one experiment was conducted for each filter to make comparison and observe the trend. Triplicate tests were done for the background of ACF mats and multi sandwiched PAN filters in o rder to ensure the data reproducibility. Because the fiber diameter in the PAN nanofiber filter is much smaller than the conventional HEPA filter, prior studies (Podgorski et al., 2006) reported a shift of the most penetrative particle size to smaller size s To study this effect, 5% wt NaCl water solution was
26 atomized to produce polydisperse NaCl particles A spectrum of the particle size distribution of NaCl particles ranging from 10 to 400 nm was measured by the SMPS By comparing the aerosol concentratio n before and after the filter, a penetration curve was obtained for particle sizes from 10 to 400 nm. The most penetrative particle size was then determined from the penetration curve. In addition, particle concentration generated from pure deionized water was also measured for a background line to rule out the contribution from the cylinder air and the water solvent. 3.3 Phase II : Biological Inactivation The bioaerosols collected on the filter were not inactivated yet. They might detach from the filter and re enter the ambient air. In this phase, microwave irradiation was applied to inactivate the microorganisms collected on the filter. The microwave oven used in this study was manufactured by Panasonic, USA (Model NN T 945SF, 2.45 GHz, continuous irradiatio n ) and the system was fabricated by Cha Corp. at WY. Since two 1 holes were drilled on the backside of the microwave oven, the microwave leakage could not be completely avoided. A Federal standard sets the safe level of microwave leakage at 5 mW/cm 2 at ap proximately 2 inches from the oven surface, which is far below the level known to be harmful to human beings ( US FDA CDRH 2008) According to the fabricator s note, when the microwave power level was held under 625 W (50% of the highest power available) the maximum leakage recorded was about 3 mW/cm 2 which was below the maximum safe level. In order to have a safe operational condition, 500 W (40% of the highest power available) was used as the maximum power level in this study. E. coli was used as the challenging bioaerosol in this study. As a representative of vegetative cells, E. coli is a gram negative rod shaped bacterium whose typical size is 0.8 2
27 m ( Sundararaj et al., 200 4 ) and is often selected as a challenging aerosolized microorganism for g ermicidal tests (Keller et al., 2005; Lin and Li, 2003a; Lin and Li, 2003b; Maness et al., 1999; Vohra et al., 2005; Vohra et al., 2006; Yu et al., 2008) A non pathogenic strain of E. coli (ATCC number 15597) was used as the challenging microorganism in t his study. It was first inoculated on a Difco Nutrient A gar (Becton, Dickinson and Company, Lot No.: 8057703, MD) plate from the stock. This plate was then incubated under 37 C for 24 hours. E. coli colonies were then formed on the plate. Before each exper iment, one single colony was picked up an d inoculated on a Difco Nutrient Agar slant and incubated for another 24 hours. Then the E. coli population on the slant was transferred to Ringer s solution (Fisher, Cat. No.: S77939, NY) by vortex. Thus, this E. c oli suspension was ready for bioaerosol generation. When the E. coli culture on the plate was 3 days old, a colony was picked up and inoculated into another Difco Nutrient Agar plate to renew the culture. In this way, the E. coli culture was renewed every 3 days on the agar plate to keep the freshness of the culture and consistency of experimental conditions. There were two sets of experiments. The first set was a preliminary static on filter inactivation experiment, which gave important references to the s econd set. The experimental set up for physical testing was retained in this stage since the microwave irradiation was applied after the aerosol collection instead of during the aerosol flight. The two ports for taking aerosols to the SMPS were closed. The prepared E. coli suspension was aerosolized in the nebulizer. The flow rates of both the nebulizer and dilution dryer remained the same as the physical tests. The experiments were run for 30 minutes each time in order to collect enough E. coli for adequat e experimental resolution. After collection, the filter was taken out of the filter holder and cut into four equal quadrants. Since the ACF mats could not be used in the microwave oven due to its
28 undesirable ignition, the PAN nanofiber layer was peeled off the ACF mats. One of the quadrants was directly inoculated into the prepared Ringer s solution by vortex. This inoculated solution was diluted and plated onto different petri dishes labeled different dilution levels (dilution series from 1:10, 1:100, to 1 :10 5 ) and incubated at 37 C for 24 hours. The numbers of E. coli colonies formed on the plates were then counted. A plate with 30 to 300 colonies would be taken into account and the viable bioaerosol concentration would equal to the number of colonies on that plate multiplied by the dilution level. The other three quadrants were placed between two SiC discs and heated in the microwave oven for different periods of time to study how complete the microwave irradiation could inactivate the bioaerosols in a ce rtain period of time and how long it would take to completely disinfect the bioaerosols by microwave irradiation. SiC discs (Vesuvius, 47 mm OD 20 mm thickness, 45 PPI, OBSIC, Lot No.: AL7370101L) were used here because their high ability to absorb micro wave irradiation. After microwave irradiation, the quadrants were also inoculated into the prepared Ringer s solution and cultured as described above. In the case that there was no plate with more than 30 E. coli colonies for a certain quadrant, the plate of the lowest dilution level was then taken into account. The survival fraction ( S ) on the filter was then calculated as: (3 1) where c t is the viable bioaerosol concentration of the quadrant which is treated in microwave for a time t and c 0 is the viable bioaerosol concentration of the quadrant without microwave treatment. Additionally, i n order to demonstrate the rationality of the hypothesis that the 4 quadrants can collect similar amounts of viable E. coli a triplicate set of control tests (all of the 4 quadrants were not microwaved) were carried out. Also, in order to investigate whether the re is any significant microwave non thermal effect in killing the E. coli SEM images of E. coli
29 treated by three different methods (immobilized over flame, conventional oven heated, and microwave irradiated) on glass slides were taken and compared. The se cond set of experiment was a dynamic in flight inactivation. The microwave irradiation was applied during the bioaerosol flight and filtration. Since normal filter holders could not be directly exposed to microwave, a pure quartz filter holder was designed and placed inside the microwave oven in this set due to its transparency to the microwave. This microwave reactor can hold a standard 47 mm diameter round shaped filter. However, the effective diameter of the filter located in the reactor is only 40 mm du e to the filter holder structure. Since ACF mats was no longer able to act as the support layer because it would ignite inside the microwave oven, non combustible silicon carbide mats (Plain weave ceramic grade Nicalon cloth, COI Ceramics, UT) were used in this unit in order to pro vide the mechanical support for the PAN nanofiber filtration media. Figure 3 3 shows the modified experimental set up for the in flight tests. The air flow rates supplied to the nebulizer and the dilution dryer were both controlle d at 6 LPM in order to produce a suitable relative humidity (about 50%) for the E. coli testing ( Cox and Wathes, 1995; Wang and Brion, 2007 ) After the dilution dryer, the air flow was split equally into three paths (4 LPM each) The first one was directed to the pure quartz filter holder inside the microwave oven (the microwave/filtration system). The pressure drop across the filter was monitored by a differential pressure gauge. Flow rate across the microwave/filtration system was adjusted according to th e pressure drop. Since the set flow rate through the system was 4 LPM, the face velocity across the quartz filter holder was 5.3 cm/s, based on the effective filter diameter in the quartz filter holder (40 mm). At the downstream of the microwave/filtration system, the air stream was directed to a Biosampler (SKC, Cat. No.: 225 9595, PA ) for collecting the aerosol.
30 Biosampler was chosen rather than a conventional AGI 30 impinger because of its higher collection efficiency (Lin et al. 1999) The second 4 LPM split stream was directed to a filter holder outside the microwave oven with the same filter as the one in the microwave quartz reactor. Pressure drop was also measured to help adjusting the flow ra te. The third split stream was directed to another Biosam pler without any filtration to determine the feed concentration A vacuum pump (Cat. No.: G574, Marathon Electric, WI) was used to help drawing the air from the system in order to prevent air from accumulating at the glass joints and then popping them apar t. The filter inside the microwave oven was labeled A, and the one outside was labeled B. The Biosampler downstream of the microwave/filtration system was labeled C and the other one was labeled D. Therefore, the on filter E. coli survival fraction under microwave irradiation could be calculated by comparing the viable E. coli in the two filters (A/B). Meanwhile, the viable E. coli penetration from the microwave/filtration system could be obtained by comparing the viable E. coli concentration in the two Bi osamplers (C/D) The bioaerosol collection and inactivation was carried out for 30 minutes (three 10 minute cycles) for each single experiment. Three different microwave power level s (500 W, 250 W, and 125 W) and four different microwave application time s (continuously, last 5, 2.5, and 1.25 minutes of each 10 minute cycle) were combined and used as microwave irradiation conditions. Each combination was tested in triplicate.
31 Figure 3 1 Filters to be tested in this study
32 Figure 3 2 Experimental s et up for Phase I physical tests and Phase II static on filter inactivation tests
33 Figure 3 3 Experimental set up for Phase II dynamic in flight inactivation tests
34 CHAPTER 4 RESULTS AND DISCUSSION 4.1 Physical Characterization of Nanofiber Filte rs 4.1.1 Physical T ests of A ctivated C arbon F iber M ats ( B ackground) To evaluate the true performance of the PAN nanofiber only, both the pressure drop and particle capture of the ACF mats and other structures should be eliminated from the data of sandwich filters. Physical properties of different numbers of ACF mat layers were tested under 5.3 cm/s as background. Results are shown in Table 4 1. Because the total pressure drop is the sum of the pressure drop yielded from every element of the measured part, t he p vs. n relationship can be written as below: (4 1) The subscript t represents the total pressure drop, a refers to the pressure drop yielded by a single layer ACF mat, and s stands for the pressure drop produced by the filter holder and other structures (tubings, fittings, etc.). Making linear regression for the p versus n relationship according to the data listed in Table 4 1 gives Equation 4 2 : (4 2) By comparing E quation s 4 1 and 4 2, it can be concluded that a single layer ACF mat yields a 10.6 Pa pressure drop under the 5.3 cm/s air face velocity, and the filter holder and other structures produce 4.17 Pa. Thus, a 25 Pa pressure drop of any ACF/PAN/ACF sandwich filters is produced by 2 layers of ACF mats and other structures. T he final particle penetration fraction is the product of the penetration fractions through each element of the measured part (Wang et al., 2008b) Hence, the P versus n relationship can be written as below: (4 3)
35 The subscripts ( t a and s ) have the same meanings as the pressure drops described above. Making exponential regression for the P vs. n relationship according to the data in Table 4 1 gives Equation 4 4 : (4 4) By comparing E quations 4 3 and 4 4, it can be concluded that the particle penetration fraction across a single layer ACF mat is 0.931. The pre factor 1.02 ( 1) suggest s that the net effect of the other structures is negligible Accordingly, for 2 layers of ACF mats used in the sandwich structure the system has an effective particle penetration fraction of 0.88 4 4.1.2 Background Particle Concentrations T o eliminate the particle con centration contributed by the cylinder air and solvents of the nebulized solution, background tests were carried out. Figure 4 1 shows the background particle size distribution of the aerosols generated from pure deionized water and 50% (volume ratio) etha nol water solution. The results indicated that the background particle concentrations were at least 2 to 3 logs lower than the experimental particle concentrations ( 300 nm PSL and polydisperse NaCl aerosols, > 4 10 5 #/cm 3 ). In order words, these backgrou nd particle concentrations were negligible in this study. 4.1.3 Filtration P erformance of S ingle S andwiched F ilters Single sandwiched PAN nanofiber filters (one layer of PAN nanofiber mat between two ACF mats) were tested for their filtration efficiency. F our PAN nanofiber mats with different electrospinning deposition time s (15, 20, 25, and 30 minutes) were used. In addition, two conventional glass fiber HEPA filters (a Millipore HEPA filter and a LydAir HEPA filter) were included in the results for compar ison. The detailed filtration results are shown in Table 4 2. There were four filters having the filter quality factor higher than the military standard MIL F
36 51079D. However, there is no filter with a filter quality factor higher than the objective of thi s study (half pressure drop therefore double q F value). The conventional fibrous LydAir HEPA filter had the highest quality factor. Also, three out of the four PAN nanofiber filters had better quality factors than the military standard MIL F 51079D. Furthe rmore, it can be deduced that the PAN deposition time during the electrospinning affects the filter quality. There is a trend that shorter deposition time of the PAN nanofiber could yield a better quality factor. In order to identify the cause for this tre nd, the thickness uniformity of the PAN nanofiber mat was studied. A piece of glass slide deposited with a thin layer of PAN nanofiber was observed by a Wyko optical profilometer for the PAN thickness distribution. As shown in Figure 4 2, the thickness of the PAN nanofiber layer ha s the maximum at the center of the slide and decrease s toward the edge. This is likely due to the deposition pattern in an electrical field. However, the air flow tends to go through the thin part for the least resistance and cons equently lowers the efficiency. Podgorski et al. (2006) also pointed out that the existence of large pores in the thin parts of the nanofiber mats may lead to particle direct passage and diminish the filtration performance. Therefore, the less uniformity in the thickness as a result of longer electrospinning deposition time caused the lower filter quality factor. Though the thickness uniformity was improved after this finding, currently there is no effective technique to solve the problem completely (Barh ate and Ramakrishna, 2007). 4.1.4 Filtration P erformance of M ulti S andwiched F ilters The filter quality is not related to how many layers of filtration media were used according to its definition. Since a reverse trend of the electrospinning deposition ti me versus the filter quality was observed in the results of single sandwiched filters, it is worthwhile to examine overlapping two or more layers of the short term electrospun PAN nanofiber layers in order to
37 keep the high filter quality while achieving re latively better collection efficiency. Therefore, a new batch of filters w as tested for their filtration performance including two single sandwiched PAN nano filters (PAN 5 and PAN 15, i.e., 5 and 15 minutes deposited by electrospinning) and two multi sandw iched PAN filters (PAN 5 3 and PAN 15 2, n means n layers of PAN nanofiber mats supported and separated by n +1 ACF mats). Results are shown in Table 4 3. Since the electrospinning process had been modified for this batch of PAN nanofibers, the PAN 15 in this batch was not identical to the one mentioned in the previous section. The phenomenon that longer term deposited PAN fiber filters have lower quality factor was also observed in this batch of filters. Theoretically, PAN 5 and PAN 5 3 should have the s ame filter quality factor since the quality factor is designed to eliminate the effect of thickness for filters made by the same material. However, the filter quality factor of PAN 5 3 was much higher than PAN 5. The same situation also occurred at PAN 15 versus PAN 15 2. A possible reason is that stacking of fibermats with different orientation benefited the filtration performance. As described before, the PAN nanofiber mats were not uniform in their thickness. In other words, there were some parts thicke r while some other parts thinner that co existed in the fiber mat. Since those round shaped PAN were cut from larger square shaped fibermats, they did not necessarily all have the thickest part in the center. Thus, some thicker parts in a certain layer mi ght have compensated for those deficient thin ner parts in another single layer in the stacking and therefore improve d the filter quality. A schematic description of this explanation is shown in Figure 4 3. Another possible reason is relat ed to decreased pa rticle velocity. When a particle hit s the first stage but not collected due to rebound, part of its kinetic energy is absorbed by the fibers. T he decreased particle velocity improve s the filter quality. However, through the following theoretical calculatio n in the following text, it was found that
38 this hypothesis was not valid. Dahneke (1971) developed an equation for the kinetic energy required for particle rebound ( KE b ): (4 5) where d p is the particle siz e, A is the Hamaker constant, which is dependent on the materials and ranges form 6 10 20 J to 1.5 10 18 J for common materials (Hinds, 1999), e is the ratio of the rebound velocity to the approach velocity, usually ranges from 0.73 to 0.81 (Wall et al, 1990), x is the separation distance between the surface and the particle hitting on it, and is usually assumed to be 0.4 nm for smooth surfaces (Hinds, 1999). In order to estimate the minimum KE b 0.3 m was taken for the PSL particle size, and all the ot her variables in Equation 4 5 were chosen to yield a minimum KE b T he corresponding KE b for the PSL particles is 1.2 10 17 J. The kinetic energy ( KE ) of a certain spherical PSL particle can be calculated as: (4 6) where m p is the particle mass, p is the particle density, V p is the particle volume and v is the particle velocity. To estimate the KE of the PSL particles in this study, it is assumed that p = 1 .05 g/ c m 3 d p = 0. 3 m and v = 5.3 cm/s. T he calculated KE for PSL particles i s 2. 1 10 2 0 J, which is about 3 orders of magnitude lower than the KE b Therefore, the PSL particles a re not likely to rebound after hitting on the filter. It should also be noted that th e PAN 5 3 had a much higher filter quality factor than the PAN 15. This result supports the above mentioned strategy of making better filters. The fact that the quality factor of the PAN 5 3 was much higher than the one of the LydAir HEPA filter as well as the military standard MIL F 51079D suggested that the multi sandwiched filters with short term deposited nanofiber mats would be a good candidate in the RHELP system for collecting aerosol agents.
39 Many other research studies also examined the quality fact or of nanofibrous filters and multi layer fibrous filters. Yun et al. (2007) fabricated PAN nanofiber filters by electrospinning and challenged them with particles ranging from 10 to 80 nm. A clear trend that the quality factor decreased with increased par ticle size was observed. Therefore, all their PAN filters had a minimum quality factor at the particle size of 80 nm. The highest filter quality factor they got was about 0.045 Pa 1 Wang et al. (2008b) combined a microfiber mat with a nanofiber ( d f = 150 nm) mat to form a micro/nano composite fibrous filter. The nanofibers acted as collection layer facing the air stream containing challenging aerosols and the microfiber worked as the nanofiber support so as to be a back layer. Four nanofiber mats with diff erent solidities (packing densities) were used in this structure. One sample with an effective solidity close to conventional HEPA or HVAC filters was found to have a better quality factor than those conventional filters when collecting 300 nm particles. A mong the four composite micro/nano filters, the highest filter quality factor was obtained at about 0.03 Pa 1 Podgorski et al. (2006) investigated the filtration performance of some bilayer structures composed of a microfibrous support and a nanofibrous f acial collection layer. The nanofibrous layers were fabricated by the melt blown method. They found the quality factors of the micro/nano fiber filters were significantly higher than conventional HEPA filters for 300 nm particles. The highest filter qualit y fac tor was close to 0.008 Pa 1 They also studied the effect of adding more nanofiber layers to the bilayer structure. Results showed that the multi nano/single micro structure improved the filter quality a lot. Specifically, for 300 nm particles, the fi lter quality factor rose from 0.004 Pa 1 to 0.006 Pa 1 for one micro/nano fiber filter by adding one layer of the same nanofib er to the existing nanofiber mat. Our experimental results also agree with this finding.
40 Comparing with these three research studi es, our results showed a significant ly higher filter quality factor when PAN 5 3 was applied. Detailed comparison is shown in Table 4 4. A major difference between filters used in the present study and theirs is the usage of the support layer for the nanof iber mat. Research by Wang et al. (2008b) and Podgorski et al. (2006) applied a microfiber mat as the support. Although it was beneficial that the microfiber can contribute a small part of the collection efficiency, the cost of increasing pressure drop ove rweighed this benefit. In our study, nearly all particle collection was acquired by the nanofiber layer and the highly penetrative ACF mats produced a negligible pressure drop. Hence, a more optimal filter quality was achieved. 4.1.5 Most P enetrative P arti cle S ize According to the postulations of Podgorski et al. (2006), the actual most penetrative particle size for a filter will shift to smaller size s as the filter fiber diameter decreases. As shown in Figure 4 4, the PAN nanofiber filters used in this stu dy have their fiber diameter smaller than conventional HEPA filters; so, their actual most penetrative particle size is expected to be smaller than the conventional HEPA filters, which are usually between 100 and 500 nm (Podgorski et al., 2006). In order t o determine the actual most penetrative particle size, physical collection tests were conducted by aerosolizing polydisperse NaCl particles to obtain a penetration curve for particles ranging from 10 to 400 nm. Figure 4 5 shows a typical size distribution of NaCl particles nebulized from 5% wt NaCl solution in this study scanned by the SMPS. Figure 4 6 is the particle penetration curve through the sandwiched PAN 30 filter. As shown, the actual most penetrative particle size of that filter was less than 100 nm. Figure 4 7 is a plot of MPPS versus filter fiber diameter, including experimental data and theoretical prediction reported by other researches as well as the present study As shown, the MPPS of the
41 present study falls in the range that well agrees wit h the theoretical prediction. It should be noted that d f is not the only factor that affects the MPPS. The testing face velocity also has an effect on it. Since smaller particles are mainly collected due to diffusion, the MPPS will shift to smaller size s b ecause the increased face velocity will reduce the retention time for diffus ional collection (Hinds, 1999). 4.2 Static On Filter Inactivation The PAN 60 (1 hour deposited PAN nanofiber) filters, instead of a multi sandwiched thin nanofiber filter, were ch osen for the E. coli collection/filtration. The relatively easier fabrication and installation were its advantages. Also, the collection efficiency would be high enough considering the size of E. coli A triplicate set of control test (all of 4 quadrants w ere not microwaved) were carried out to demonstrate the rationality of the hypothesis that the 4 quadrants can collect similar amounts of viable E. coli R esults are shown in Table 4 5 T he E. coli concentration s in each quadrant are not perfectly the same But the deviations are not very significant considering th at microwave inactivation is usually measured exponentially Six sets of the static on filter E. coli inactivation tests were carried out under 500 W microwave power in order to have a relative ly high power and safe operation condition. As shown in Table 4 6, in less than 90 seconds the survival became below the detection limit. These results suggest that the microwave irradiation can effectively inactivate E. coli collected on the filtration medi a. The data had some deviation when microwave irradiation was applied for 30 seconds and 60 seconds. A possible reason is that the filter quadrants were not placed at the exactly same position in the microwave oven. Since this microwave oven has a working
42 frequency of 2.45 GHz, which corresponds to a wavelength of 12.2 cm, the position in the microwave oven is cr itical to the inactivation rate Several other research studies also reported the E. coli inactivation rate under microwave irradiation. Table 4 7 is a detailed summary of these results. As shown, 5 to 7 logs of viable E. coli are usually inactivated within 30 to 90 seconds when microwave power is higher than 800 W. Watanabe et al. (2000) used 500 W microwave, and their results were very close to the results of the present study It is also shown that the E. coli inactivation can be significantly accelerated if some assisting approach were added to the microwave irradiation, such as argon plasma and catalytic reactions. It should be noted that only th e present study was conducted in air while all the others were conducted in an aqueous system or a medium containing significant amount of water. SEM images of E. coli treated with different methods (immobilized over flame, conventional oven heated, and mi crowave irradiated) on glass slides are shown in Figure 4 8. As shown, no significant difference in the damaged cells were found. In order to further investigate any microwave non thermal effects, some other methods need to be applied to study the damages on the cell s inner structures. 4.3 Dynamic In Flight Inactivation I n order to determine if the microwave irradiation ha d an effect on the bioaerosols during their flight, the experiment without a filter was carried out. 500 W continuous microwave was appl ied. The results are listed in Table 4 8. As shown, about 95% (1.3 logs) of E. coli was inactivat ed during the flight across the microwave field. Since the bioaerosol total flight time in the microwave oven was less than 5 seconds, the results sug gest that microwave irradiation contributed a big part of the E. coli inactivation in the microwave/filtration system.
43 To study the integrated effect of microwave and filtration, all combinations of the three microwave power levels (500 W, 250 W, and 125 W) and fou r microwave application times (continuously, last 5, 2.5, and 1.25 minutes of each 10 minute cycle) were tested in triplicate. The dynamic on filter E. coli survival fractions (the A/B value) under different microwave power levels or microwave application times are listed in Table 4 9 and plotted in Figure 4 9 It is noticed that the E. coli survival fraction was more sensitive to the microwave power level than the application time. A statistical (ANOVA two way) test showed that the microwave power was a ma jor factor that affects the E. coli survival fraction (A/B). The p value was 0.003. On the other hand, the microwave application time did not have a significant influence. The p value was larger than 0.3. As shown in Figure 4 9 when 500 W microwave power was used, at least 3.7 logs of the viable E. coli w ere removed (survival fraction below 0.02%) no matter the microwave was applied to system continuously (10 min per 10 minute cycle) or periodically (5 or 2.5 min per 10 minute cycle). This suggests that 2.5 mins per cycle is efficient enough to inactivate the E. coli when 500 W microwave power is applied. For other combinations of the microwave power and application time, the microwave inactivation rates were lower. A general correlation between the survival fraction and the microwave power level can be determined from the data. It can be seen that the survival fraction decreases exponentially as the microwave power levels increases (linearly in the log scale ) except the series of 1.25 min /10 min cycle microw ave application time. By m aking the exponential regression for the A/B versus microwave power relationship of each microwave application time, four equations listed in Table 4 10 were obtained. Comparing the E. coli on filter survival fraction of the two s ets of biological tests (the 500 W microwave power level), it can be seen that the dynamic in flight microwave irradiation was slightly less efficient in killing the E. coli collected on the filter than the static method.
44 Though the inactivation fraction ( or log removal) was close, the microwave application time in the in flight experiments was longer. A possible reason is that the continuous air flow in the microwave and filtration system help ed the heat dissipation on the filter and reduced the temperatur e to which the E. coli expose. Therefore, the thermal effect of the microwave irradiation was less intensive than the static inactivation. Another possible reason is the location of the filter inside the microwave oven. The filter location was fixed for al l the dynamic in flight experiments. As described before, if the microwave power of this point was not as intensive as some of the static experiments, it was reasonable that the inactivation fraction was lower. The viable E. coli penetration fractions (the C/D value) under different microwave power levels or microwave application times are listed in Table 4 11 and plotted in Figure 4 1 0 However, as shown in the Figure 4 1 0 the data points are scattered and no obvious correlation can be seen. Furthermore, the variations of each data points (geometric mean of 3 experiments with same conditions) are large. Statistic analysis (t tests) showed that there are effectively no differences among t he data. The p values are larger than 0.05. Several reasons may cause the unstable data. The PAN 60 filters examined in this study were not identical to each other, which may be responsible for the deviation of filtration efficiency. Moreover, it has been observed that some of the PAN 60 filters were partially damaged during the microwave irradiation so that the particles could penetrate through more easily and resulted in scattered data. Recently, a cross linked PAN nanofiber mat was developed and tested. The strength of the cross linked PAN nanofibers ensures better durabil ity to the microwave irradiation than the current ones. If future studies can apply this cross linked PAN nanofiber filter, the damage of filter due to the irradiation may be reduced or even avoided.
45 Table 4 1 Physical test results of ACF mats Number of ACF layers ( n ) Penetration f raction a ( P dimensionless) Pressure d rop at 5.3 cm/s ( p Pa) 2 0.864 25 4 0.799 47.5 6 0.648 67.5 a T he penetration fraction is for 300 nm particles
46 Figure 4 1 Background particle concentration of (A) nebulized water; (B) nebulized 50% (volume ratio) ethanol water solution 0 200 400 600 800 1000 1200 9.82 14.1 20.2 28.9 41.4 59.4 85.1 121.9 174.7 250.3 358.7 d N /dlog( d p ) (#/cc) Partical Size (nm) 0 10000 20000 30000 40000 50000 60000 9.82 14.1 20.2 28.9 41.4 59.4 85.1 121.9 174.7 250.3 358.7 d N /dlog( d p ) (#/cc) Particle Size (nm) (A) (B)
47 Table 4 2 Physical test results of single sandwiched PAN filters with comparison to two conventional HEPA filters and the military standard Filter Particle p enetration a (dimensionless) Pre ssure d rop (Pa) Quality f actor (Pa 1 ) ACF 2 b 8.66 10 1 21 .0069 Millipore 1.5 10 4 1171 .0073 LydAir 1.8 10 4 284 .0304 PAN c 15 d 6.02 10 3 172 .0297 e PAN 20 7.12 10 3 215 .0230 PAN 25 1.03 10 3 272 .0253 PAN 30 4.35 10 3 340 .0160 M STD f 3 10 4 400 .0203 OBJ g 3 10 4 200 .0406 a P enetration of 300 nm particles b T wo layers of ACF mats only, background c PAN nanofiber layer filter (without two ACF mats) d D eposition time of the PAN nanofiber when made by electrospinning e The filtration test of PAN nanofiber filters were carried out under 7.14 cm/s due to the different experimental plan at the beginning of the research. Note that the result q F value will be a little higher if tested under 5.3 cm/s f M i litary HEPA standard MIL F 51079D g O bjective of this study
48 Figure 4 2 Thickness characterization of PAN nanofiber mats by Wyko optical profilometer A) The glass slide used for thickness measurement B) The nanofiber layer s thickness distr ibution along the center line of the glass slide. X is the distance from the left edge of the slide; Y is the distance from the center line of the slide; Z is the filter thickness. 0 10 20 30 40 50 60 70 80 0 0.5 1 1.5 2 2.5 3 Z (m) X (in.) X Y A) B )
49 Table 4 3 Physical test results of multi sandwiched PAN filters with c omparison with two single sandwiched ones Filter Penetration ( d imensionless) Pressure d rop (Pa) Quality f actor (Pa 1 ) PAN 5 5.23 10 1 18 0.0370 PAN 15 a 1.02 10 1 98 0.0233 PAN 5 3 b 7.95 10 2 41 0.0625 PAN 15 2 2.07 10 2 152 0.0255 a T his PAN 15 filter is not identical to the PAN 15 listed in Table 4 2 b PAN m n represents n layers of PAN m filters piled together to form a multi layer filter including n +1 layers of ACF mats for support
50 Figure 4 3 A schematic description of the nanofi ber mat stacking ( The stacking is not necessarily as perfect as in the picture. However, it does help to compensate some weak part of a certain layer.)
51 Table 4 4 Comparison of filter quality factors among our results and those of other researches Rese arch Filter Fiber d iameter (m) Material Quality f actor (Pa 1 ) Comments Podgorski et al. (2006) BL a 18 Polymer b ~ .003 Microfiber BL + NL c 4 1.18 d Polymer ~ .008 This filter had the highest quality factor in this study BL + NL1 .74 Polymer ~ 004 BL + 2NL1 .74 Polymer ~ .006 Double nano layer structure Yun et al. (2007) ES e 2 .27 PAN .045 f This filter had the highest quality factor in this study Wang et al. (2008b) Sample C ~ .15 N/A g .02 The one used for comparison w ith conventional filters Sample A ~ .15 N/A .03 This filter had the highest quality factor in this study Present study PAN 53 ~ .20 PAN .063 PAN 5 ~ .20 PAN .037 Other comparison Millipore ~ .50 G lass fiber .007 LydAir ~ .50 Glass fiber .03 M STD .02 OBJ .04 a M icrofiber back layer b A kind of polymer, not specified by the authors c Facial nano layer d Fiber diameter of the nanofiber layer e Electrospun filter f Based on results of 80 nm parti cles g Not mentioned by the authors
52 Figure 4 4 The SEM images of A) PAN nanofiber mat; B) Millipore HEPA filter; C) LydAir HEPA filter (Images are provided by Hyoungjun Park, Dr. Wolfgang Sigmund s lab in the Department of Materials Science and Engineering of the University of Florida )
53 Figure 4 5 Particle size distribution of the aerosol generated from nebulization of 5% wt NaCl solution Median size: 79 nm; Geometric mean size: 80 nm; Mode size: 75 nm; Geometric standard deviation: 1.91; Total number concentration: about 10 6 #/cm 3
54 Figure 4 6 The penetration curve of PAN 30 nanofiber filter T he MPPS appears around 50 nm 0.0E+00 5.0E 04 1.0E 03 1.5E 03 2.0E 03 10 100 Penetration ( P ) Particle Size ( d p nm) 20 400 200 50
55 Figure 4 7 Comparison of the most penetrative particle size among our results and those of other research es [the testing face velocity was 10 cm/s for both Podgorski et al. (2006) s and Wang et al. (2008a) s experimental results, and also for Podgorski et al. (2006) s calculation; the recorded filtration velocity in this study was 7.14 cm/s] 0 50 100 150 200 250 300 350 400 0.01 0.1 1 10 100 Most Penetrative Particle Size (nm) Filter Fiber Diameter (m) Podgorski et al. (2006) Theoretical Calculation Podgorski et al. (2006) Experimental result Wang et al. (2008a) Experimental Result PAN 30 Experimental Result (Present study)
56 Table 4 5 P ercentages of E. coli concentration on each quadrant Disk # Quad a A Quad B Quad C Quad D Mean SD b CV c 1 20% 24% 30% 26% 25% 4.16% 17% 2 35% 16% 30% 19% 25% 8.98% 36% 3 12% 24% 26% 38% 25% 10.65% 43% a Quadrant b Standard Deviation c Coefficient of Varia tion
57 Table 4 6 Survival fraction of E. coli under microwave (500 W) irradiation Static on filter inactivation Microwave irradiation time 30 s 60 s 90 s Test 1 8.3% < 0.6% < 0.6% Test 2 < 0.2% < 0.2% < 0.2% Test 3 2.8% < 0.1% < 0.1 % Test 4 < 0.05% < 0.05% < 0.05% Test 5 2.9% 1.7% < 0.05% Test 6 < 0.04% 0.08% < 0.04% Mean 2.38% 0.46% 0.17% Standard Deviation 3.20% 0.64% 0.22% Note: The < xx value means the survival fraction was below the detection limit. The xx value was calculated by putting a 1 in the nominator (assuming there is only one colony formed in the non diluted plate). The mean and standard deviation was also based on those xx data instead of a zero value.
58 Ta ble 4 7 Comparison of the E. coli inactivation by microwave irradiation among our static on filter experiments and those of other researches Research Inactivation r ate Microwave a pplication t ime (s) Microwave p ower (W) Assisting a pproaches E. coli s urvivin g m edia Goldblith et al. (1967) 6 logs 50 N/A a None PBS b Fujikawa et al. (1992) 6 logs 90 300 None PBS 4 logs 150 200 None PBS 5 logs 240 100 None PBS Watanabe et al. (2000) 3 logs 50 500 None PBS Apostolou et al. (2005) 6 logs 35 800 None Small chicken portion Awuah et al. (2005) 7 logs 55 1200 None Milk Park et al. (2006) 5 logs 30 1000 None Sponge 6 logs 60 1000 None Scrubbing pads Park et al. (2007) 4 logs 1 1000 Argon plasma Saline c Takashima et al. (2007) 8 lo gs 5 100 Catalytic reaction NB d Present study 3 4 logs 90 500 None Nanofiber filter a Not mentioned by the authors b Phosphate buffer solution c Normal saline, 0.9% NaCl water solution d Nutrient broth, Eiken Chemical Co. Ltd., Tokyo, Japan
59 Figure 4 8 The SEM image of E. coli on the glass slide A) Immobilized by the alcohol lamp B ) C onventional oven heated C ) M icrowave irradiated A 5000 nm B 5000 nm 5000 nm C
60 Table 4 8 E scherichia coli in flight inactivation percentages under 500 W continuous microwave without a filter Test 1 93.8 % Test 2 94.8 % Test 3 94.7 % Average 94.4 % Standard deviation 0.55 %
61 Table 4 9 (A) The on filter E. coli survival fraction (A/B values) of the dynamic in flight tests Microwave p ower (W) Microwave t ime (m in/10 mins) Test 1 Test 2 Test 3 Geo Mean GSD a ln GSD 500 10 1.87E 04 3.33E 04 b 9.43E 05 1.80E 04 1.88 0.63 5 3.05E 05 2.82E 04 8.40E 05 8.97E 05 3.05 1.11 2.5 1.55E 04 3.28E 04 6.85E 05 1.52E 04 2.19 0.78 1.25 4.26E 03 2.08E 03 5.77E 04 1.72E 03 2.75 1.01 250 10 1.38E 02 1.56E 03 7.04E 05 b 1.15E 03 14.19 2.65 5 2.54E 02 1.40E 02 2.37E 02 2.04E 02 1.39 0.33 2.5 1.79E 02 4.49E 02 9.76E 04 9.22E 03 7.38 2.00 1.25 1.39E 03 1.39E 03 3.30E 04 8.61E 04 2.29 0.83 125 10 1.02E 02 1.19E 01 1.03E 02 2.32E 02 4.12 1.42 5 4.49E 02 2.00E 02 5.21E 03 1.67E 02 2.97 1.09 2.5 1.50E 01 1.81E 02 7.35E 02 5.84E 02 2.93 1.08 1.25 4.09E 03 1.25E 02 3.92E 03 5.85E 03 1.93 0.66 a Geometric standard deviation b Below the detection limit. Th e values in the table are the largest possible values. (B) The results in the form of log removals of the on filter viable E. coli Microwave p ower (W) Microwave t ime (min/10 mins) Test 1 Test 2 Test 3 Mean SD 500 10 3.73 3.48 4.03 3.74 0.27 5 4.52 3.55 4.08 4.05 0.48 2.5 3.81 3.48 4.16 3.82 0.34 1.25 2.37 2.68 3.24 2.76 0.4 4 250 10 1.86 2.81 4.15 2.94 1.15 5 1.60 1.85 1.63 1.69 0.14 2.5 1.75 1.35 3.01 2.04 0.8 7 1.25 2.86 2.86 3.48 3.07 0.36 125 10 1.99 0.92 1.99 1.63 0.61 5 1.35 1.70 2.28 1 .78 0.47 2.5 0.82 1.74 1.13 1.23 0.4 7 1.25 2.39 1.90 2.41 2.23 0.2 9
62 Figure 4 9 E scherichia coli survival fraction on the filter after dynamic in flight microwave irradiation under different microwave power level and microwave application time 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 10 mins/cycle 5 mins/cycle 2.5 mins/cycle 1.25 mins/cycle Log removal of viable E. coli collected on the filter Microwave Application Time 125 W 250 W 500 W
63 Table 4 10 Exponential regression of the E. coli survival fraction on the filter versus the microwave power level of the dynamic in flight experiment Microwave application time (min/10 min cycle) Exponential equation a L inearize d equation R 2 value based on Geometric means R 2 value based on single values 10 S = .0586 exp ( .012 P M ) ln S = .012 P M + 2.84 .8971 .5888 5 S = .2519 exp ( .015 P M ) ln S = .015 P M + 1.38 .8723 .8010 2.5 S = .4558 exp ( .016 P M ) ln S = .016 P M + .79 .9993 .8280 1.25 S = .0041 exp ( .002 P M ) ln S = .002 P M + 5.50 .2227 .1265 a S stands for the E. coli survival fraction, dimensionless; P M is the microwave power level, in Watt.
64 Table 4 11 (A) The viable E. coli penetration fraction (C/D values) of the d ynamic in flight tests Microwave p ower (W) Microwave t ime (min/10 mins) Test 1 Test 2 Test 3 Geo Mean GSD ln GSD 500 W 10 2.58E 02 3.33E 04 a 2.09E 04 1.22E 03 14.24 2.66 5 2.00E 03 1.44E 01 7.17E 05 2.74E 03 45.26 3.81 2.5 8.30E 01 2.45E 04 3.33E 03 8 .78E 03 63.42 4.15 1.25 6.85E 02 5.11E 02 5.35E 03 2.66E 02 4.04 1.40 250 W 10 1.24E 03 3.09E 03 4.80E 04 1.23E 03 2.54 0.93 5 6.96E 02 1.39E 01 2.92E 01 1.41E 01 2.05 0.72 2.5 7.47E 03 1.22E 02 5.04E 02 1.66E 02 2.70 0.99 1.25 1.73E 01 3. 03E 01 3.03E 01 2.51E 01 1.38 0.32 125 W 10 6.78E 02 6.15E 02 3.86E 03 2.52E 02 5.09 1.63 5 2.65E 01 1.05E 02 1.04E 03 1.43E 02 16.17 2.78 2.5 6.91E 02 2.51E 02 4.33E 02 4.22E 02 1.66 0.51 1.25 5.90E 01 9.90E 03 1.50E 02 4.44E 02 9.48 2.25 a Below the detection limit. The values in the table are the largest possible values. (B) The results in the form of log removals of the viable E. coli in the air stream Microwave Power (W) Microwave Time (min/10 mins) Test 1 Test 2 Test 3 Mean SD 500 W 1 0 1.59 3.48 3.68 2.92 1.15 5 2.70 0.84 4.14 2.56 1.66 2.5 0.08 3.61 2.48 2.06 1.80 1.25 1.16 1.29 2.27 1.58 0.61 250 W 10 2.91 2.51 3.32 2.91 0.40 5 1.16 0.86 0.53 0.85 0.31 2.5 2.13 1.91 1.30 1.78 0.43 1.25 0.76 0.52 0.52 0.60 0.14 125 W 10 1.17 1.21 2.41 1.60 0.71 5 0.58 1.98 2.98 1.85 1.21 2.5 1.16 1.60 1.36 1.37 0.22 1.25 0.23 2.00 1.82 1.35 0.98
65 Figure 4 1 0 Viable E. coli penetration fraction from the microwave/filtration system under different microwave power level and micr owave application time 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 10 mins/cycle 5 mins/cycle 2.5 mins/cycle 1.25 mins/cycle Log removal of viable E. coli in the air stream Microwave Application Time 125 W 250 W 500 W
66 CHAPTER 5 SUMMARY AND RECOMMENDATIONS The research was focused on developing and characterizing a lab scale RHELP building ventilation air purification system for bioaerosols. Nanofibrous filtration media w ere used for collecting the bioaerosols in the air stream and microwave irradiation was applied to sterilize and regenerate the filters loaded with the biological agents. Two phases of experiments were carried out to evaluate the performance of this RHELP system. Phase I experim ents characterized the physical collection properties of the nanofibrous filtration media and Phase II experiments focused on sterilizing the biological agents collected in the system. Sandwiched electrospun PAN nanofiber filters were tested for their phys ical collection performance and compared with two conventional HEPA filters. Results showed that PAN nanofiber filters could achieve a quality factor higher than the military standard MIL F 51079D T he shorter term electrosp un PAN filters had a better qual ity factor than longer term one s. T he poor thickness uniformity of the nanofiber mats was a possible reason for this T he multi sandwiched filters achieved a better quality factor than single sandwiched ones of the same total electrospinning time (PAN 5 3 versus PAN 15) or unit electrospinning time (PAN 5 3 versus PAN 5) Stacking of the nanofiber mats might have compensated for the inferior thickness uniformity and improved the quality factor. Moreover, comparing with some other recent studies, the multi s andwiched PAN 5 3 filter in the present study had the highest filter quality factor. In addition, comparable to other investigations, the actual MPPS of the PAN nanofiber filters shifted towards smaller particle size s due to their smaller fiber size. In su mmary the electrospun PAN nanofiber filters achieved a higher quality factor than the military standards for HEPA filters Using a multi sandwiched structure of the PAN nanofiber filter, the research objective
67 filter quality of this study (half the pressu re drop and same high collection efficiency of the military standard MIL F 51079D) was achieved. E. coli was used as the challenging bioaerosol in this study. Static on filter inactivation tests revealed that the position in the microwave field was very cr itical to the inactivation rate of E. coli since the distribution of microwave energy is not uniform. The results also showed that the viable E. coli concentration became lower than the detection limit when exposed to 500 W microwave for less than 90 secon ds which agreed with most recent studies This demonstrated that the microwave irradiation is able to regenerate the filter s loaded with E. coli efficiently without consuming much energy. When dynamic in flight microwave irradiation was conducted, the E. coli survival fraction was below 0.0 2 % no matter the microwave was applied to system continuously (10 min per 10 minute cycle) or periodically (5 or 2.5 min per 10 minute cycle) when 500 W microwave power was applied Compared to the static on filter inact ivation, the dynamic in flight microwave irradiation killed the E. coli collected on the filter less efficiently. Possible reasons include d a temperature drop due to a continuous air flow and a less intensive irradiation at the point where filters were loc ated. Besides, a n exponential correlation between the survival fraction and the microwave power level applied was also observed. However, during the dynamic in flight microwave irradiation the filter in the microwave field might be partially damaged and d iminished the total filtration performance of the RHELP system It is recommended that the RHELP microwave/filtration system be challenged by various types of microorganism s such as viruses and spores to evaluate its range of application.
68 LIST OF REFERE NCES Agranovski, I.E., Safatov, A.S., Pyankov, O.V., Sergeev, A.N., Agafonov, A.P., Ignatiev, G.M., Ryabchikova, E.I., Borodulin, A.I., Sergeev, A.A., Doerr, H.W., Rabenau, H.F.,Agranovski, V., 2004. Monitoring of viable airborne SARS virus in ambient air. Atmospheric Environment 38, 3879 3884. Ahn, Y.C., Park, S.K., Kim, G.T., Hwang, Y.J., Lee, C.G., Shin, H.S., Lee, J.K., 2006. Development of high efficiency nanofilters made of nanofibers. Current Applied Physics 6, 1030. Allan, G.G., Krieger, B.B., Wor k, D.W., 1980. Dielectric loss microwave degradation of polymers: Cellulose Journal of Applied Polymer Science 25, 1839 1859. American Conference of Governmental Industrial Hygienists (ACGIH), 1989. Guidelines for the assessment of bioaerosols in the ind oor environment. Cincinnati, Ohio. In: Jo, W. K., Seo, Y. J., 2005. Indoor and outdoor bioaerosol levels at recreation facilities, elementary schools, and homes. Chemosphere 61, 1570 1579. Apostolou, I., Papadopoulou, C., Levidiotou, S., Ioannides, K., 20 05. The effect of short time microwave exposures on Escherichia coli O157:H7 inoculated onto chicken meat portions and whole chickens. International Journal of Food Microbiology 101, 105 110. Awuah, G.B., Ramaswamy, H.S., Economides, A., Mallikarjunan, K. 2005. Inactivation of E. coli K 12 and L. innocua in milk using radio frequency heating. Innovative Food Science and Emerging Technologies 6, 396 402. Baeraky, T.A., 2002. Microwave measurements of the dielectric properties of silicon carbide at high te mperature. Egyptian Journal of Solids 25, 263 273. Barhate, R.S., Ramakrishna, S., 2007. Nanofibrous filtering media: Filtration problems and solutions from tiny materials. Journal of Membrane Science 296, 1 8. Beaumont, F., 1988. Clinical manifestations of pulmonary Aspergillus infections. Mycoses 31, 15 20. Betti, L., Trebbi, G., Lazzarato, L., Brizzi, M., Calzoni, G.L., Marinelli, F., Nani, D., Borghini, F., 2004. Nonthermal microwave radiations affect the hypersensitive response of tobacco to tobacco mosaic virus. The Journal of Alternative and Complementary Medicine 10, 947 957. Bonnet L., Estel L., Ledoux A., Mazari B., Louis A, 2004. Study of the thermal repartition in a microwave reactor: application to the nitrobenzene hydrogenation. Chemical En gineering and Processing 43, 1435 1440.
69 Boskovic, L., Altman, I.S., Agranovski, I.E., Braddock, R.D., Myojo, T., Choi, M., 2005. Influence of particle shape on filtration processes. Aerosol Science and Technology 39, 1184 1190. Boskovic, L., Agranovski, I.E., Altman, I.S., Braddock, R.D., 2008. Filter efficiency as a function of nanoparticle velocity and shape. Journal of Aerosol Science 39, 635 644. Brown, R.C., 1993. Air Filtration. Pergamon Press, London. In: Wang, J., Kim, S.C., Pui, D.Y.H., 2008a. F igure of merit of composite filters with micrometer and nanometer fibers. Aerosol Science and Technology 42, 722 728. Campanha, N.H., Pavarina, A.C., Brunetti, I.L., Vergani, C.E., Machado, A.L., Spolidorio, D.M.P., 2007. Candida albicans inactivation and cell membrane integrity damage by microwave irradiation. Mycoses 50, 140 147. Canumir, J.A., Celis, J.E., de Bruijn, J., Vidal, L.V., 2002. Pasteurization of apple juice by using microwaves. LWT Food Science and Technology 35, 389 392. Celandroni, F., Longo, I., Tosoratti, N., Giannessi, F., Ghelardi, E., Salvetti, S., Baggiani, A., Senesi, S., 2004. Effect of microwave radiation on Bacillus subtilis spores. Journal of Applied Microbiology 97, 1220 1227. Chen, C. C., Huang, S. H., 1998. The effects of particle charge on the performance of a filtering facepiece. American Industrial Hygiene Association Journal 59, 227 233. Clapp, W.D., Becker, S., Quay, J., Watt, J.L., Throne, P.S., Frees, K.L., Zhang, X., Koren, H.S., Lux, C.R., Schwartz, D.A., 1994. G rain dust induced airflow obstruction and inflammation of the lower respiratory tract. A merican J ournal of R espiratory and C ritical C are M edicine 150, 611 617. Cockcroft, D.W., Ruffin, R.E., Dolovich, J., 1977. Allergen induced increase in non allergic br onchial reactivity. Clinical Allergy 7, 503 513. Cox, C.S., Wathes C.M. 1995. Bioaerosols Handboo k. CRC Press L td. Boca Raton, F lorida. pp. 3 14, 77 98. Dahneke, B., 1971. The capture of a erosol p articles by s urfaces. Journal of Colloid and Interface Science 37, 342 353. Dales, R.E., Zwanenburg, H., Burnett, R., Franklin, C.A., 1991. Respiratory health effects of home dampness and molds among children. American Journal of Epidemiology 134, 196 203. Dhaniyala, S., Liu, B.Y.H., 1999. Investigation of p article penetration in fibrous filters: Part II. Theoretical. Journal of the IEST 42, 40 46.
70 Doshi, J., Reneker, D.H., 1995. Electrospining process and applications of electrospun fibers. Journal of Electrostatics 35, 151 160. Douwes, J., Thorne, P., Pea rce, N., Heederik, D., 2003. Bioaerosol health effects and exposure assessment: Progress and prospects. Annals of Occupational Hygiene 47, 187 200. Dreyfuss, M.S., Chipley, J.R., 1980. Comparison of effects of sublethal microwave radiation and conventiona l heating on the metabolic activity of Staphylococcus aureus. Applied and Environmental Microbiology 39, 13 16. Dutkiewicz, J., 1997. Bacteria and fungi in organic dust as potential health hazard. Annals of Agricultural and Environmental Medicine 4, 11 16 Flannigan, B., McCabe, E.M., McGarry, F., 1991. Allergenic and toxigenic micro organisms in houses. Journal of Applied Bactriology 70, 61 73. Fujikawa, H., Ushioda, H., Kudo, Y., 1992. Kinetics of Escherichia coli destruction by microwave irradiation. Applied and Environmental Microbiology 58, 920 924. Gedye, R., Smith, F., Westaway, K., Ali, H., Baldisera, L., Laberge, L., Roussell, J., 1986. The use of microwave ovens for rapid organic synthesis Tetrahedron Letters 27, 279 282. George, J., 2007. Na nofiber manufacturing processes for filtration media (Abstract). American Filtration and Separation Society Annual Conference, Orlando, Florida. In: Wang, J., Kim, S.C., Pui, D.Y.H., 2008a. Figure of merit of composite filters with micrometer and nanometer fibers. Aerosol Science and Technology 42, 722 728. Giguere, R.J., Bray, T.L., Duncan, S.M., Majetich, G., 1986. Application of commercial microwave ovens to organic synthesis Tetrahedron Letters 27, 4945 4948. Goldblith, S.A., Wang, D.I.C., 1967. Effe ct of microwaves on Escherichia coli and Bacillus subtilis. Applied Microbiology 15, 1371 1375. Grafe, T., Gogins, M., Barris, M., Schafer, J., Canepa, R., 2001. Nanofibers in filtration applications in transportation. In: Filtration 2001 International Co nference and Expo of the Association of the Nowovens Fabric Industry, Chicago, Illinois, December 3 5, 2001. Gustavsson, J., 1999. How can air filters contribute to better IAQ? Filtration and Separation 36, 20 25. Heer, C.E.W., zur Nieden, A., Jankofsky M., Stilianakis, N.I, Boedeker, R. H., Eikmann, T.F., 2003. Effects of bioaerosol polluted outdoor air on airways of residents: a cross sectional study. Occupational and Environmental Medicine 60, 336 342. Henderson, D.A. 1998. Bioterrorism as a public h ealth threat. Emerging Infectious Diseases 4 488 492.
71 Hinds, C. W., 1999. Aerosol Technology. John Wiley & Sons, New York. pp 141 149, 182 205, 394 401. Huang, Z. M., Zhang, Y. Z., Kotaki, M., Ramakrishna, S., 2003. A review on polymer nanofibers by el ectrospinning and their applications in nanocomposites. Composites S cience and T echnology 63, 2223 2253. Jacob, J., Chia, L.H.L., Boey, F.Y.C., 1995. Review thermal and non thermal interaction of microwave radiation with materials Journal of Materials S cience 30, 5321 5327. Jeng, D.K.H., Kaczmarek, K.A., Woodworth, A.G., Balasky, G., 1987. Mechanism of microwave sterilization in the dry state. Applied and Environmental Microbiology 53, 2133 2137. Jo, W. K., Seo, Y. J., 2005. Indoor and outdoor bioaeros ol levels at recreation facilities, elementary schools, and homes. Chemosphere 61, 1570 1579. Jones, D.A. Lelyveld, T.P. Mavrofidis, S.D. Kingman, S.W. Miles, N.J 2002. Microwave h eating a pplications in e nvironmental e ngineering a r eview Resour ces Conserv ation and Recycl ing 34, 75 90 Keller, V., Keller, N., Ledous, M.J., Lett, M. C., 2005. Biological agent inactivation in a flowing air stream by photocatalysis. Chemical Communications 23, 2918 2920. Kiel, J.L., Sutter, R.E., Mason, P.A., Parker J.E., Morales, P.J., Stribling, L.J.V., Alls, J.L., Holwitt, E.A., Seaman, R.L., Mathur, S.P., 2002. Directed killing of anthrax spores by microwave induced cavitation. IEEE Transactions on Plasma Science 30, 1482 1488. Kim, B.C., Nair, S., Kim, J., Kwa k, J.H., Grate, J.W., Kim, S.H., Gu, M.B., 2005. Preparation of biocatalytic nanofibers with high activity and stability via enzyme aggregate coating on polymer nanofibers. Nanotechnology 16, S382 S388. Kim, C.S., Bao, L., Okuyama, K., Shimada, M., Niinum a, H., 2006. Filtration efficiency of a fibrous filter for nanoparticles. Journal of Nanoparticle Research 8, 215 221. Kindle, G., Busse, A., Kampa, D., Meyer Konig, U., Daschner, F.D., 1996. Killing activity of microwaves in milk. Journal of Hospital Inf ection 33, 273 278. Kortepeter, M.G. Parker, G.W. 1999. Potential b iological w eapons t hreats. Emerging Infectious Diseases 5 523 527. Kosmider, K., Scott, J., 2002. Polymeric nanofibers exhibit an enhanced air filtration performance. Filtration and Se paration 39, 20 22.
72 Lacey, J., Crook, B., 1988. Fungal and actinomycete spores as pollutants of the workplace and occupational allergens. Annals of Occupational Hygiene 32, 513 533. Landahl, H.D., Hermann, R.G., 1949. Sampling of liquid aerosols by wires cylinders, and slides, and the efficiency of impaction of droplets. In: Wang J., Kim, S.C., Pui, D.Y.H., 2008. Figure of merit of composite filters with micrometer and nanometer fibers. Aerosol Science and Technology 42, 722 728. Lin, C. Y., Li, C. S., 2003a. Effectiveness of titanium dioxide photocatalyst filters for controlling bioaerosols. Aerosol Science and Technology 37, 162 170. Lin, C. Y., Li, C. S., 2003b. Inactivation of microorganisms on the photocatalytic surfaces in air. Aerosol Science and Technology 37, 939 946. Lin, X., Reponen, T.A. Willeke, K. Grinshpun, S.A. Forade, K.K. Ensor D.S. 1999 Long term sampling of airborne bacteria and fungi into a non evaporating liquid. Atmospheric Environment 33 4291 4298. Ma, J., Fang, M., Li, P., Zhu, B., Lu, X., Lau, N.T., 1997. Microwave assisted catalytic combustion of diesel soot. Applied Catalysis A: General 159, 211 228. Maness, P. C., Smolinski, S., Blake, D.M., Huang, Z., Wolfrun, E.J., Jacoby, W.A., 1999. Bactericidal activity of phot ocatalytic TiO 2 reaction: Toward an understan ding of it killing mechanism. Applied and Environmental Microbiology 65, 4094 4098. Matheson, M.C., Benke, G., Raven, J., Sim, M.R., Krombout, H., Vermeulen, R., Johns, D.P., Walters, E.H. Abramson, M.J., 2005. Biological dust exposure in the workplace is a risk factor for chronic obstructive pulmonary disease. Thorax 60, 645 651. Millner, P.D., Olenchock, S.A., Epstein, E., Rylander, R., Haines, J., Walker, J., Ooi, B.L., Horne, E., Maritato, M., 1994. Bioaero sols associated with composting facilities. Compost Science and Utilitization 2, 6 57. Millner, P., 1995. Bioaerosols and composting. Biocycle 36, 48 54. Pastuszka, J.S., Paw, U.K.T., Lis, D.O., Wlazlo, A., Ulfig, K., 2000. Bacterial and fungal aerosol i n indoor environment in Upper Silesia, Poland. Atmospheric Environment 34, 3833 3842. Park, B.J., Takatori, K., Lee, M.H., Han D. W., Woo, Y.I., Son, H.J., Kim, J.K., Chung, K. H., Hyun, S.O., Park, J. C., 2007. Escherichia coli sterilization and lipopoly saccharide inactivation using microwave induced argon plasma at atmospheric pressure. Surface and Coatings Technology 201, 5738 5741. Park, D. K., Bitton, G., Melker, R., 2006. Microbial inactivation by microwave radiation in the home environment. Journa l of Environmental Health 69, 17 24.
73 Pellerin, C., 1994. Alternatives to incineration: there s more than one way to remediate. Environmental Health Perspectives 102, 840 845. Persson, B.R.R., Salford, L.G., Brun, A., Eberhard, J.L., Malmgren, I., 1992. I ncreased Permeability of the Blood Brain Barrier Induced by Magnetic and Electromagnetic Fields Annals of New York Academy of Sciences 649, 356 358. Phelan, A.M., Neubauer, C.F., Timm, R., Neirenberg, J., Lange, D.G., 1994. Athermal a lterations in the s t ructure of the c analicular m embrane and ATPase a ctivity i nduced by t hermal l evels of m icrowave r adiatio n. Radiation Research 137, 52 58. Pich, J., 1965. The filtration theory of highly dispersed aerosols. In: Wang J., Kim, S.C., Pui, D.Y.H., 2008. Figure of merit of composite filters with micrometer and nanometer fibers. Aerosol Science and Technology 42, 722 728. Pich, J., 1966. The effectiveness of the barrier effect in fibre filter at small Knudsen numbers. In: Wang J., Kim, S.C., Pui, D.Y.H., 2008. Fi gure of merit of composite filters with micrometer and nanometer fibers. Aerosol Science and Technology 42, 722 728. Podgorski, A., Balazy, A., Gradon, L., 2006. Application of nanofibers to improve the filtration efficiency of the most penetrating aeroso l particles in fibrous filters. Chemical Engineering Science 61, 6804 6815. Prescott, M. L., Harley, P.J., Klein, A.D. 200 6 Microbiology 6 th Edition McGraw Hill, New York. pp 829 830. Ramaseshan, R., Sundarrajan, S., Liu, Y., Barhate, R.S., Lala, N.L., Ramakrishna, S., 2006. Functionalized polymer nanofiber membranes for protection from chemical warfare simulants. Nanotechnology 17, 2947 2953. Ratnesar Shumate, S., Wu, C. Y., Wander, J ., Lundgren, D., Farrah, S., Lee, J. H., Wanakule, P., Blackburn, M., Lan, M. F., 2008. Evaluation of physical capture efficiency and disinfection capability of an iodinated biocidal filter medium. Aerosol and Air Quality Research 8, 1 18. Rega, P.P., 200 1. The b iological t errorism r esponse m anual. MASCAP, Inc., Maumee, OH. p p 66 Ren, P., Jankun, T.M., Leaderer, B.P., 1999. Comparisons of seasonal fungal prevalence in indoor and outdoor air and in house dusts of dwellings in on Northeast American county. Journal of Exposure Analysis and Environmental Epidemiology 9, 560 568. Reneker, D.H., Chun, I., 1996. Nanometre diameter fibres of polymer, produced by electrospinning. Nanotechnology 7, 216 223.
74 Salalha, W., Kuhn, J., Dror, Y., Zussman, E., 2006. Enca psulation of bacteria and viruses in electrospun nanofibers Nanotechnology 17, 4675 4681. Sasaki, K., Honda, W., Miyake, Y., 1998. Evaluation of high temperature and short time sterilization of injection ampoules by microwave heating. Journal of Pharmace utical Science and Technology 52, 5 12. Schaechter, M., Ingraham, J.L., Neidhardt, F.C., 2006. Microbe. ASM Press, Washington, D.C. pp. 452. Siersted, H.C., Gravesen, S., 1993. Extrinsic allergic alveolitis after exposure to the yeast Ehodotorula rubra. Allergy 48, 298 299. Squires, S.B., Gardiner, M.J., 2005. Cigarette filter incorporating nanofibers. US Patent Application No. 2005/0139223. Subbiah, T., Bhat, G.S., Tock, R.W., Parameswaran, S., Ramkumar, S.S., 2005. Electrospinning of nanofibers. Jo urn al of Applied Polymer S cience 96, 557 569. Takashima, H., Miyakawa, Y., Kanno, Y., 2007. Microwave sterilization with metal thin film coated catalyst in liquid phase Materials Science and Engineering C 27, 898 903. Sundararaj, S., Guo, A., Habibi Nazhad B., Rouani, M., Stothard, P., Ellison, M., Wishart D.S., 2004. The CyberCell Database (CCDB): a comprehensive, self updating, relational database to coordinate and facilitate in silico modeling of Escherichia coli Nucleic Acids Res earch 32 (Database is sue) D293 D295 Thom, A., 1933. The flow past circular cylinders at low speeds. In: Wang J., Kim, S.C., Pui, D.Y.H., 2008. Figure of merit of composite filters with micrometer and nanometer fibers. Aerosol Science and Technology 42, 722 728. US FDA RH (Center for Devices and Radiological Health) 2008. CDRH consumer information Microwave oven radiation. FDA/CDRH, Rockville, MD URL address: http://www.fda.gov/cdrh/consumer/microwave.html ( last accessed : Nov 1 9 2008). Vaid, A., Bishop, A.H., 1998. The destruction by microwave radiation of bacterial endospores and amplication of the released DNA. Journal of Applied Microbiology 85, 115 122. Vohra, A., Goswami, D.Y., Deshpande, D.A., Bl ock, S.S., 2005. Enhanced photocatalytic inactivation of bacterial spores on surfaces in air. Journal of Industrial Microbiology and Biotechnology 32, 364 370. Vohra, A., Goswami, D.Y., Deshpande, D.A., Block, S.S., 2006. Enhanced photocatalytic disinfect ion of indoor air. Applied Catalysis B: Environmental 64, 57 65.
75 Wall, S., John, W., Wang, H. S., Goren, S., 1990. Measurements of kinetic energy loss for particles impacting surfaces. Aerosol Science and Technology 12, 926 946. Wang, C. S., 2001. Electr ostatic forces in fibrous filters a review. Powder Technology 118, 166 170. Wang, J., Kim, S.C., Pui, D.Y.H., 2008 a Figure of merit of composite filters with micrometer and nanometer fibers. Aerosol Science and Technology 42, 722 728. Wang, J., Kim, S .C., Pui, D.Y.H., 2008b. Investigation of the figure of merit for filters with a single nanofiber layer on a substrate. Journal of Aerosol Science 39, 323 334. Wang, M., Brion, G., 2007. Effects of RH on glass microfiber filtration efficiency for airborne bacteria and bacteriophage over time. Aerosol Science and Technology 41, 775 785. Ward, G., 2005. Nanofibres: media at the nanoscale. Filtration and Separation 42, 22 24. Watanabe, K., Kakita, Y., Kashige, N., Miake, F., Tsukiji, T., 2000. Effect of ion ic strength on the inactivation of micro organisms by microwave irradiation. Letters in Applied Microbiology 31, 52 56. Will, H., Scholz, P. Ondruschka B. 2004 Heterogeneous gas phase catalysis under microwave irradiation a new multi mode microwave applicat or. Topics in Catalysis 29 175 182. Woo, I. S., Rhee, I. K., Park, H. D., 2000. Differential damage in bacterial cells by microwave radiation on the basis of cell wall structure. Applied and Environmental Microbiology 66, 2243 2247. Wouters, I.M ., Spaan, S., Douwes, J., Doekes, G., Heederick, D., 2006. Overview of personal occupational exposure levels to inhalable dust, endotoxin, (1 3) glucan and fungal extracellular polysaccharises in the waste management chain. Annals of Occupational Hygiene 50, 39 53. Yaghmaee, P., Durance, T.D., 2005. Destruction and injury of Escherichia coli during microwave heating under vacuum. Journal o f Applied Microbiology 98, 498 506. Yu, K. P., Lee, G.W. M., Lin, S.Y., Huang, C.P., 2008. Removal of bioaerosols by the combination of a photocatalytic filter and negative air ions. Journal of Aerosol Science 39, 377 392. Yun, K.M., Hogan C.J. Jr., Mats ubayashi, Y., Kawabe, M., Iskandar, F., Okuyama, K., 2007. Nanoparticle filtration by electrospun polymer fibers. Chemical Engineering Science 62, 4751 4759.
76 Zhang, X., Hayward, D.O., 2006. Applications of microwave dielectric heating in environment relat ed heterogeneous gas phase catalytic systems. Inorganica Chimica Acta 359, 3421 3433. Zielinski, M., Ciesielski, S., Cydzik Kwiatkowska, A., Turek, J., Debowski, M., 2007. Influence of microwave radiation on bacterial community structure in biofilm. Proce ss Biochemistry 42, 1250 1253. Zuskin, E., Schachter, E.N., Kanceljak, B., Mustajbegovic, J., Wiltek, T., 1994. Immunological and respiratory reactions in workers exposed to organic dusts. International A rchives of O ccupational and E nvironmental H ealth 66 317 324.
77 BIOGRAPHICAL SKE T CH Qi Zhang was born in Shanghai, China in 1984 He graduated from the No. 2 Middle School a ttached to the East China Normal University in June 2002. He passed the Chinese college entrance examination and was admitted to Fudan University that summer. He started working in the Aerosol Laboratory of the Department of Environmental Science and Engineering, Fudan University in 2004. In summer 2006, Qi Zhang finished his bachelor s degree and came to the University of Florida to pur sue his master s degree. He joined the A erosol and Particulate Research Lab under Dr. Chang Yu Wu in the Department of Environmental Engineering Sciences in January 2007 and started working on his research project. His current research project is microwa ve assisted nanofiber filtration for bioaerosols.