1 DIELECTRIC BARRIER DISCHARGE (DBD) SURFACE PLASMA STERILIZATION: AN IN DEPTH STUDY OF THE FACTORS CONTRIBUTING TO AND ENHANCING THE STERILIZATION PROCESS By NAVYA MASTANAIAH A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL O F THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2013
2 2013 N avya Mastanaiah
3 To the one person who believed in me with her whole heart. Mom Ad Astra Per Aspera
4 ACKNOWLEDGMENTS This thesis has seen the light of day, because of the direct support and encouragement of a number of people as well as the indirect inspiration and guidance provided by a vast number of people, whose names if I listed here, would make for a separate dissertation sized document! First and foremost, heartfelt gratitude is due to my advisor, Dr. Subrata Roy. I have learnt a tremendous amount from his thoughtful input, his patience and encouragement and hi An equally heartfelt note of thanks is due to Dr.Judith Johnson, who has provided me, a student with an engineering background, a functional and tremendously helpful understanding in the bi ological aspects of my research. Her enthusiasm and readiness in helping me plan my experiments truly inspired me to give my best. Thanks are also due to my committee members, Dr.D avid Hahn, Dr. Renwei Mei and Dr.Tommy Angelini. Their valuable suggestion s and expert guidance helped me better understand many facets of the work that has gone into this dissertation. I have been blessed with a fun group of lab mates and would be highly remiss if I did not thank James, Mark, Tomas, Ryan, Jignesh, Ariel, Moses Arnob, Pengfei and Ankush for the high spirited discussions, the thoughtful suggestions and for the memorable lab get togethers. Especially James for cheerfully guiding me in my initial days at APRG. Of course, my work in plasma sterilization would be in complete if I did not thank Raul A. Chinga for his invaluable work in optimizing the electrical circuit used for plasma generation. It is truly a pleasure to work with and learn from someone so patient and meticulous
5 Words are not enough to express gratit ude to all my wonderful friends, throughout my time here, for all the late night discussions, the potlucks, the gator nights through the roughest of times and Sidd h esh and Ajet a, who have a wonderful way of cheering me up every time. Last, but definitely not the least, this dissertation and every other success I have been and will be fortunate enough to see in the future is dedicated to and solely because of the constant love, s upport and strength provided by the three most important people in my life: Mom, Dad and Kutty. I am nothing if not for you three.
6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIGURES ................................ ................................ ................................ .......... 9 ABSTRACT ................................ ................................ ................................ ................... 14 CHA PTER 1 INTRODUCTION ................................ ................................ ................................ .... 16 1.1 Sterilization Current State of the Art ................................ ................................ 16 1.2 What is Plasma? ................................ ................................ ............................... 21 1.3 DBD Plasma Sterilization ................................ ................................ .................. 27 1.3.1 Parametric Studies of the Factors Affecting Plasma Sterilization ........ 29 1.3.2 Mechanism of Plasma Sterilization ................................ ..................... 36 1.3.3 Plasma Interaction with Biofilms ................................ .......................... 44 1.3 .4 Numerical Modeling of Plasma Sterilization ................................ ........ 45 2 RESEARCH MOTIVATION ................................ ................................ ..................... 48 3 METHODS AND MATERIALS ................................ ................................ ................ 50 3.1 Experimental Setup used for Plasma Sterilization Experiments ....................... 50 3.2 Design of the Plasma Device ................................ ................................ ............ 52 3.3 The Portable Sterilization Setup ................................ ................................ ....... 56 3.4 Description of Biological Pathogens Tested ................................ ...................... 57 3.5 Experimental Protocols Followed ................................ ................................ ...... 60 3.5.1 Preparation of Bacterial Biological Test Samples for Experiments: .... 61 3.5.2 Post Processing of Bacterial Sampl es after Experiments ................... 63 3.5.3 Ozone Safety Protocol ................................ ................................ ........ 63 4 PARAMETRIC STUDIES IN DBD SURFACE PLASMA STERILIZATION .............. 66 4.1 Type of Microorganism ................................ ................................ ..................... 66 4.2 Inoculation Volume ................................ ................................ ........................... 78 4.3 Nature of Dielectr ic Material ................................ ................................ .............. 80 4.4 Input Power and Frequency ................................ ................................ .............. 90 4.5 Operating Pressure ................................ ................................ ........................... 95 4.6 Discussion ................................ ................................ ................................ ........ 99 5 UNDERSTANDING THE MECHANISM OF DBD SURFACE PLASMA STERILIZATION ................................ ................................ ................................ ... 105
7 5.1 Spectroscopic Studies ................................ ................................ .................... 105 5.2 Ozone Studies ................................ ................................ ................................ 113 5.2.1 Characterization of Ozone Production and Decay During DBD Plasma Generation ................................ ................................ ..................... 114 5.2.2 The Effect of Ozone Produced During DBD Plasma Generation on E. Coli ................................ ................................ ................................ ......... 124 5.3 Temperature Studies ................................ ................................ ...................... 133 5.4 Microbiological Analysis ................................ ................................ .................. 141 5.4.1 Evaluation of Membrane Damage by Live/Dead BacLight TM Assay 142 5 .4.2 Mutation Studies. ................................ ................................ .............. 147 5.4.3 Microscopic Analysis of Plasma Interaction with B. subtilis Biofilms 148 5.5 Discussion ................................ ................................ ................................ ...... 151 6 CONCLUDING REMARKS AND RECOMMENDATIONS FOR FUTURE WORK 159 6.1 Scope of Technology ................................ ................................ ...................... 163 6.2 Further analysis of the Mechanism of Plasma Sterilization ............................. 164 6.3 Numerical Modeling in Plasma Sterilization ................................ .................... 165 REFERENCES ................................ ................................ ................................ ............ 168 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 178
8 LIST OF TABLES Table p age 1 1 A c omparison of various factors of an ideal sterilant for different sterilization methods ................................ ................................ ................................ .............. 20 1 2 Micro discharge properties in air at atmospheric pressure ................................ 25 3 1 Listing of all the microorganisms tested ................................ .............................. 58 3 2 Description of organelles and essential proteins in different types of microorganisms ................................ ................................ ................................ .. 60 4 1 Results obtained from Plasma Sterilization Experiments with BSL II pathogens ................................ ................................ ................................ ........... 77 4 2 Description of protocols that the different devices were subject ed to, prior to SEM Testing. ................................ ................................ ................................ ...... 84 5 1 Values of the constants used in Equation (5 5) (5 6) ................................ ....... 119
9 LIST OF FIGURES Figure page 1 1 Schematic of Sterrad and Plazlyte Plasma Sterilization Systems. ...................... 19 1 2 Plasma temperatures and number densities ................................ ...................... 21 1 3 Dependence of voltage on current for various DC discharges ........................... 23 1 4 Schematic of Volume Plasma DBD configuration and DBD surface plasma configuration ................................ ................................ ................................ ....... 24 1 5 End on view of micro discharges in atmospheric pressure air ........................... 26 1 6 Schematic of an apparatus used earlier on for plasma steriliza tion .................... 28 1 7 Various Factors involved in DBD Plasma sterilization ................................ ........ 30 1 8 Comparison of spore mortality in MW and RF O2/CF4 plasma s, after 5 minutes of exposure ................................ ................................ ........................... 32 1 9 Results of inactivation of E. coli on the agar surface by direct and indirec t plasma treatment ................................ ................................ ................................ 34 1 10 Wavelength and energy of radiation in the UV and visible portion of the spectrum ................................ ................................ ................................ ............. 36 1 11 UV spectrum of a DBD in air in the 200 300 nm wavelength range .................... 38 1 12 I ncrease of sample temperature vs. plasma dissipated power for a DBD volume plasma in air at atmospheric pressure ................................ ................... 41 1 13 Schematic illustration of the different phases in a plasma sterilization survival curve ................................ ................................ ................................ ................... 43 1 14 Numerical modeling of the plasma interaction with s kin ................................ ..... 46 3 1 Schematic of the experimental setup used ................................ ......................... 50 3 2 dered All dimensions are listed ................................ ................................ ................................ .... 53 3 3 T he three electrode configurations, shown in Figure 3 2, when powered ........... 54 3 4 Stamp Test (at 60s) for t he Sawtooth electrode design, electrode design and comb like ele ctrode design ................................ ................................ .......... 55
10 3 5 Portable experimental setup, using devices made of Rogers 3003C semi ceramic dielectric ................................ ................................ ................................ 57 3 6 The bacteri al cell wall for t he Gram positive envelope and t he Gram negative envelope ................................ ................................ ................................ ............. 58 3 7 Structure of the yeast cell wall. Ergosterol is the major lipid component of the underlying plasma structure. ................................ ................................ ............... 59 3 8 Typical growth curve for a bacterial population in a batch culture ...................... 61 4 1 S urvival curves obtained using FR4 plasma devic es using S. cerevisiae (Yeast) and E. coli as test micro organisms. ................................ ...................... 70 4 2 Survival curves obtained using FR4 devices for B. subtilis cells grown in LB and MSgg medium ................................ ................................ ............................. 73 4 3 Survival curves, using FR4 devices for G stearothermophilus spores ............... 74 4 4 Comparison of D values for the different test microorganisms, using FR 4 devices ................................ ................................ ................................ ............... 76 4 5 Survival curves for inoculation volume= 40 l of E. coli ................................ ...... 79 4 6 Survival curves comparing FR4 and SC plasma dev ices for S. cerevisiae (yeast) and E. coli ................................ ................................ ............................... 82 4 7 Comparison of D values for the different (dielectric, test pathogen) ................... 82 4 8 S EM images of FR4 devices at 2000x Images A E and F J correspond to the SEM images of the dielectric and electrode surface of the same devices .... 85 4 9 SEM images of SC devices at 2000x Images A E and F J correspond to the SEM images of the dielectric and electrode surface of the same devices. ........ 86 4 10 EDS analysis of the dielectric surface and electrode surface for a FR4 plasma dev ice ................................ ................................ ................................ ..... 87 4 11 EDS analysis of the dielectric surface and electrode surface for a SC plasma device ................................ ................................ ................................ ................. 88 4 12 Comparison between a FR4 device and a PMMA device. ................................ .. 89 4 13 Comparison of the temporal variation of input power for clean and inoculated devices in the case of FR4 and SC dielectric. ................................ .................... 91 4 14 Comparison of the average measured input power (W) for each input voltage (kV pp) ................................ ................................ ................................ ................ 92
11 4 15 Temporal variation of input power for different input voltages u sing inoculated devices for FR4 and SC dielectric material ................................ ......................... 93 4 16 Dependence of sterilization effectiveness on input voltage (V) .......................... 93 4 17 S terilization results at frequency f= 60 kHz showing t emporal variation of input power and d ependence of sterilization on input voltage at f= 60 kHz ........ 94 4 18 I mages of the devices at operating pressures of 760, 500 and 400 Torr ............ 96 4 19 Sterilization behavior at reduced operating pressures for FR4 and SC plasma devices. ................................ ................................ ................................ .............. 98 4 20 Dependence of bacterial inactivation on the applied plasma dose ................... 102 5 1 S pectral signature of a clean and inoculated FR4 device. Y axis lists emission intensity in arbitrary unit ................................ ................................ .... 106 5 2 Spectral signature of a clean and inoculated SC device. Y axis lists emission intensity in arbitrary unit ................................ ................................ ................... 107 5 3 E xpanded version of the s pectral signature of a clean and an inoculated FR4 device. ................................ ................................ ................................ .............. 109 5 4 Expanded version of the s pectral signature of a clean and an inoculated SC device. ................................ ................................ ................................ .............. 109 5 5 Spectroscopic comparison of FR4 and SC devices at reduced operating pressure ................................ ................................ ................................ ........... 110 5 6 Schematic of Chamber#4. The grey square in the midd le represents the plasma device. ................................ ................................ ................................ .. 115 5 7 Spatial v ariation of ozone distribution along the X axis and the Y axis in the sterilization chamber ................................ ................................ ......................... 116 5 8 The two different configurations in which the device is placed. ........................ 117 5 9 Comparison of ozone concentrations in all four chambers. The plasma device is powered at 0 min and turned off at 2 minutes ................................ .... 119 5 10 Correlation of the ozone levels with the corresponding chamber volumes. ...... 120 5 11 Correlation o f the total number of ozone molecules present in each chamber at t= 60,120,240,360,420s to the respective chamber volumes. ..................... 122 5 12 Comparison of ozone concentrations during 7 minutes for FR4 versus SC. The plasma device is powered at 0 min and turned off at 2 minutes ................ 123
12 5 13 Experimental schematic for the ozone exposure tests. ................................ ... 124 5 14 Inactivation plots due to ozone exposure using a FR4 and a SC plasma generator ................................ ................................ ................................ .......... 126 5 15 I nactivation plots due to ozone exposure in the different chambers using a FR4 pla sma generator and an inoculated FR4 substrate ................................ 128 5 16 Comparison of ozone production with and without charcoal for Chamber#1. Plasma is turned off at 2 min. ................................ ................................ ........... 129 5 17 Inactivation plots due to ozone exposure in Chamber#1 with and without charcoal.. ................................ ................................ ................................ .......... 130 5 18 DBD Surface Plasma Sterilization, comparing air and nitrogen a s the discharge gas. Two sterilization times (60s and 120s) are tested .................... 132 5 19 Variation of temperatures at t= 30s, 60s,90s, 120s for Clean and Inoculated FR4 device ................................ ................................ ................................ ....... 136 5 20 Variation of temperatures at t= 30s, 60s,90s, 120s for Clean and Inoculated SC device ................................ ................................ ................................ ......... 137 5 21 Comparison of average surface temperatures during plasma generation for clean and inoculated FR4 and SC devices ................................ ....................... 138 5 22 Sterilization plots analyzing the effect of temperature I noculated devices were heated up to the av g. temperature measur ed during plasma generation. 139 5 23 Sterilization plots analyzing the effect of temperature I noculated devices were heated up to the max temp measured during plasma generation. ......... 140 5 24 Fluorescence Images obtained of the different cell suspensions after exposure to plasma for t= 0,30,60,90 and 120s respectively. .......................... 143 5 25 An example of the calibration curve calculated by measuring Ratio G/R for cell suspensions with different proportions of live/dead cells. ................................ 144 5 26 Plot of the Ratio G/R calculated afte r different plasma exposure times ............... 146 5 27 Images of bio films before, during and after plasma exposure. CFP indicates motile cells and YFP indicates matrix producing cells. ................................ ... 150 5 28 Mean intensity variation in CFP and YFP modes for imaging order A E. On the Y axis is plotted the % reduction in intensity ................................ ............... 151 5 29 Plo t of pH values, obtained by rinsing devices with Millipore water after plasma generation and measuring the pH val ue of this water in each case ..... 156
13 6 1 Scheme of reactions u sed for Lipid Peroxi dation ................................ .............. 166
14 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy DIELECTRIC BARRIER DISCHAR GE (DBD) SURFACE PLASMA STERILIZATION: AN IN DEPTH STUDY OF THE FACTORS CONTRIBUTING TO AND ENHANCING THE STERILIZATION PROCESS By Navya Mastanaiah August 2013 Chair: Subrata Roy Major: Aerospace Engineering Plasma sterilization is a method in which ion ized gas is used to treat contaminated objects or surfaces. It can be divided into two regimes: volume and surface plasma sterilization. While a lot of research has been done in understanding the former, the latter is yet to be fully explored. The purpose of this study is to identify and understand the key contributors controlling the process of surface plasma sterilization under atmospheric conditions. This is accomplished by a two pronged approach of parametric studies and mechanistic studies. The plasma used for the purpose of this study is known as dielectric barrier discharge (DBD) plasma. Parametric studies help characterize the dependence of sterilization on DBD surface plasma generation parameters such as input power operating pressure etc. In doin g so, the energy flux introduced by plasma (J/ cm 2 ) has been identifi ed as an intrinsic parameter for sterilization. A threshold plasma dose of approximately 285 J/cm 2 is required for complete bacterial inactivation in the case of E. coli.
15 The mechanism of the sterilization process is studied using diagnostic information measured in terms of spectroscopic signature, ozone concentrations and surface temperatures of the dielectric surface during plasma generation. Results from these studies imply that ozone pl ays a major role in the process of surface plasma sterilization. Isolating the role of temperature during plasma generation leads to the hypothesis that ozone and heat may play a synergistic role in the process of surface plasma sterilization. Additionall y, the interaction of plasma with the microorganisms is visualized using both high resolution microscopy and fluorescence microscopy. In studying plasma interaction with biofilms, it is determined that cell types expressing a matrix forming phenotype are m ore susceptible to plasma treatment as compared to cell types expressing a motile phenotype. This provides a valuable insight into the effect of plasma treatment on the cell biology. This dissertation demonstrates DBD surface plasma as a very effective st erilization method and identifies key control parameters and mechanisms. The understanding of such a technology will allow access to many life critical medical technologies including self sterilizing operating tables and food counters, and even portable st erilization kits for triage situations.
16 CHAPTER 1 INTRODUCTION Plasma s terilization is quickly evolving into a sought after method of sterilization in multiple industries: food preparation, healthcare, medicine etc. Specifically, DBD p lasma sterilizatio n is an interesting topic to pursue due to its ability to operate at atmospheric conditions as well as its operational simplicity It finds application in a wide variety of real world scenarios. However a deeper understanding of the fundamentals of plasm a sterilization is required before this technology can be transplanted from research to reality This study is an effort in that direction. 1.1 Sterilization Current State of the Art Sterilization destroys all micro organisms [ 1 ] It is the certaint y that everything is killed. The term micro organisms covers a broad spectrum of pathogens, including bacteria, fungi, viruses, endospores and prions. Of these, endospores and prions deserve a special mention, because of the challenge they pose to existi ng methods of sterilization. Endospores are tough, dormant, reproductive cells produced by some bacteria as a survival mechanism when threatened by harsh conditions. In times of stress, th i s bacterium replicate s its DNA and develop s a double membrane and thick cell wall around it, forming what is called an endospore. Once the harmful conditions tide over, th e endospore re germinates. Endospores are particularly resistant to most of the current sterilization methods. In fact, spore s are often used as biolog ical indicators to test the sterilization efficiency of techniques such as autoclaving. One factor for this resistance is attributed to the lo wer water content of the spore [ 2 3 ] Prions are infectious agents com posed of proteins in a mis folded form. They differ from other infectious agents which contain nucleic acids. When a prion enters a
17 healthy organism, it acts as a template to guide the misfolding of more protein s into prion form s Sterilizing prions there a state that it is no longer capable of inducing protein mis folding in molecules containing normal proteins. Prions are usually highly resistant to proteases (enzymes s function ), heat, radiation and chemical treatments, although their infectivity can be reduced by such treatments. All know n prion diseases affect structure s of the brain or other neural tissue s, and are all currently unt reatable and universally fatal [ 4 ] Conventional sterilization methods consist of moist and dry heat sterilization [ 5 6 ] ; chemical sterilization using ethylene oxide and glutaraldehyde [ 7 8 ] ; irradiation by high energetic rays irradiation and UV irradiation and more recently gas plasma sterilization [ 9 11 ] A brief description of each of these methods is given below. I t is to be noted that for all these methods, effective sterilization can be achieved only when the object to be sterilized is initially wiped clean, so that any organic matter remaining on the instrument is removed. Heat sterilization : This p rimarily consists o f two techniques: moist heat sterilization and dry heat sterilization. Moist heat sterilization (autoclaving) uses high pressure and temperature to achieve complete sterilization. Unwrapped objects are exposed to steam at 121 o C for 20 minutes or at 137 o C f or 15 minutes. Dry heat sterilization comprises exposing unwrapped objects to intense heat (170 o C) for 1 hour and then cooling them down for 2 2.5 hours. In both moist and dry heat sterilization, cycles begin only when the objects being sterilized reach th e specified sterilization temperature and significant amounts of time are required for cooling before use In
18 addition to the time required, the major drawback to heat sterilization methods is that they damage many heat sensitive materials. Chemical ster ilization : This technique u tilize s common disinfectants such as Ethylene oxide (EtO), glutaraldehyde or formaldehyde. O bjects need to be soaked at least for 10 hours in 2 4% glutaraldehyde solution or for 24 hours in 8% formaldehyde solution. EtO steriliza tion is a much more complex procedure, requiring a chamber in which the contaminated objects are exposed to EtO vapor for at least 2 hours and thereafter the chamber is aerated for a long period to dispel any toxic vapor. Irradiation : This process uses hi ghly energetic gamma ( ) rays Gamma irradiation is the irradiation of contaminated matter using photons in the gamma part of the electromagnetic spectrum (wavelength<10 12 m) Radiation is obtained through radio isotopes such as cobalt 60 or cesium 137. I t is used to sterilize medical devices used in operations and other healthcare treatmenets. Gamma irradiation is also used for sterilizatio n in food, pharmaceutical, cosmetic, horticultural and automotive industries However owing to the inherent hazards o f such a technique the main radioactive source has to be shielded for the safety of the operators [ 12 ] UV irradiation is another disinfection method that uses at short wavelengths (~254 nm) to kill micr o organisms. At this wavelength, it is effective in destroying the nu cleic acids in micro organisms by a process known as dimerization [ 13 ] which br eaks down the DNA bonds in a micro organism Gas plasma sterilization : Commercial p lasma s terilization s ystems, called Sterrad systems are shown in Fig ure 1 1.
19 Figure 1 1. Sc hematic of A ) Sterrad and B [ 14 ] The S terrrad, shown in Figure 1 1( A ), consist s of a vacuum chamber in wh ich wrapped items we re placed on trays and exposed to hydrogen peroxide (H 2 O 2 ) gas for 45 minutes. Following this, 300 W of RF power at 13.56 MHz wa s applied at very low pressure to create plasma inside the chamber. T to be mainly due to the H 2 O 2 vapor with the plasma primarily removing the toxic residue Similarly, the Plazlyte TM shown in Figure 1 1(B) consist ed of a vacuum chamber in which wrapped items were placed. Peracetic acid (PAA) vapor was pumped into the chamber upstream of where the wrapped items were placed. Plasma treatment consist ed of excitation of a mixture of oxygen, hydrogen and argon at low pressure, using microwave (MW) plasma at 2.45 GHz. The Plazlyte treatment consist ed of alternate cycles of the vapo r treatment and plasma treatment. Note that both of these sterilization systems are not plasma sterilizers, in the strict sense of the term, as the objects being sterilized do not come into contact with the plasma [ 15 ] Eventually in 1998, the FDA issued an alert on using the Plazlyte TM sy s tem for ophthalmic
20 instruments The problem appeared to be the deposition of copper and zinc salts on devic es sterilized with this system, which caused serious eye injuries in some patients (FDA talk paper T98 17, 1998). An ideal sterilant a s defined by Moisan et al. [ 16 ] should provide (a) short sterilization times (b) low processing temperatures (c) versatility of operat ion and (d) harmless for patients, operators and materials. Table 1 1 provides a comparison of all these factors for the different sterilization methods discussed. Table 1 1. A comparison of various factors of an ideal sterilant for different sterilization methods [ 16 ] Sterilization Method COMPARISON Sterilization time Processing Temperatures Toxicity/Handling Hazard Versatility Autoclaving 50 60 min 121 o C 137 o C Cooling down time required Cannot be used for heat sensitive polymers Dry Heat Ov ens 3 4 hours 160 o C 170 o C Cooling down time required Same limitations as autoclaving EtO Sterilization 1 2 hours 55 o C Toxic vapors, possibility of chemical residue on objects Can damage certain sensitive polymers UV Irradiation Dependent on UV do sage Low temperatures Exposure to UV rays can burn the skin and eyes Can effectively irradiate the topmost layer of cells only. Dependent on object geometry Irradiation Short times, but requires a longer standing time for radiation levels to reduce to safe levels. High temperature Operation requires thick concrete shields and safety measures for handlers. Used only in some facilities d ue to low availability of radioactive isotopes
21 DBD p lasma at atmospheric conditions meets all of the above requiremen ts for an ideal sterilant The time scales of plasma sterilization range from 2 20 minutes. Since it is generated from air, once the plasma is turned off, ionized species recombine into components of air, which are non toxic. Finally, because it is low tem perature and non toxic, it can be used with a wide variety of materials, even heat sensitive polymers. There is an abundant cache of literature on experimental methods of plasma sterilization, wherein different plasma sources, using different plasma param eters, have been used to sterilize different standard bacterial samples [ 17 20 ] Before this topic can be discussed further, a brief introduction to p lasma, espe cially DBD p lasma must be provided. 1.2 What is Plasma? Plasma is known as the fourth state of matter. It makes up the majority of the universe. The b est known natural plasma phenomenon in e lightning [ 21 ] Figure 1 2. Plasma temperatures and number densities [ 21 ]
22 The seeds for this t heory were first sown in about 1750, when Benjamin Franklin suspected that lightning is an electrical current and conducted experiments with a kite. Another natural phenomenon is the aurora borealis. Natural and man made plasmas occur over a wide range of pressures, temperatures and electron number densities. Figure 1 2 above shows this diversity. Apart from naturally occurring plasmas, plasmas can also be generated for industrial purposes. Plasma finds applications in electronics, lasers, fluorescent lamps television screens, computer and cell phone hardware and more recently, in medical applications. Plasma is an ionized gas, made up of ions, electrons and neutrals. It is most commonly categorized either on the basis of temperature or electron number den sity. Temperature: L aboratory plasmas can be distinguished into two categories: high temperature plasmas and low temperature plasmas [ 22 ] These plasmas can also be divided into l ocal t hermal e quilibrium (LTE) p lasmas and n on LTE p lasmas. The heavy particles in plasma s (ions, atoms) are at a much lower temperature than the electrons, which are typically highly energetic particles. In a high temperature plasma, the high temperature serves to equilibrate the high temperature of the electrons with the ion temperature thus establishing LTE I n non LTE plasma, this thermal equilibrium is n ot established. LTE discharges are typically used for high temperature applications such as welding. Non LTE plasmas are typically used for low temperature applications such as etching or plasma deposition. Degree of ionization of plasma s : A common condit ion in plasma chemistry is for the gases to be only partially ionized [ 21 ] The ionization degree (ratio of density of
23 major charged species to neutral s pecies) for most conventional plasma chemical systems is in the range of 10 7 10 4 Plasmas, with an ionization degree close to one are called fully ionized plasmas Most thermo nuclear and space plasma systems belong to this category. Weakly ionized plasm a s with a low degree of ionization, encompass most laboratory plasmas Figure 1 3. Dependence of voltage on current for various DC discharges [ 23 ] In any volume of gas, there exist free el ectrons. If the electric field is high enough, these will accelerate and collide with molecules of the gas, releasing more electrons, which in turn will do the same, creating an electron avalanche. As long as net charge is not sufficient enough to distort the electric field, this electron avalanche moves with the electron drift velocity, applied to the electric field. If during this avalanche, secondary electrons are generated, then they create newer avalanches. This
24 mechanism, by which current grows expone ntially, is known as the Townsend Breakdown Mechanism. Different types of plasma discharges can be obtained, depending on the applied voltage and discharge current. Fig ure 1 3 shows this dependence. The Townsend discharge is a self sustained dark discharg e. The transition from the Townsend discharge to the sub normal/normal dark discharge regime is accompanied by a decrease in voltage and a simultaneous increase in discharge current. A further increase in discharge current leads to an irreversible transiti oning of the glow discharge into the arc regime. The DBD discharge occurs in the transition between the corona and normal glow discharge, which will be described in further detail in the next section. D ielectric barrier discharge (DBD) plasma was known as early as 1857, when Werner Von Siemens reported experimental investigations wherein a flow of oxygen or air was subjected to the influence of a DBD maintained in a narrow annular gap between two coaxial electrodes, to which an alternating electric field wa s applied [ 24 ] Figure 1 4 Two types of plasma configuration A) Volume Plasma DBD configuration used in most experimental setups Schematic b ased on a similar schematic shown in [ 25 ] B) DBD surface plasma configuration used in this study In its simplest configuration, DBD is the gas discharge between two electrodes, separated by one or more dielectric layers and a gas filled gap. The most common
25 configurations are shown above in Figure 1 4. Figure 1 4 (B) shows the DBD configuration used in this thesis. It consists of two electrodes separated by a dielectric barrier. Plasma is seen on top of the electrode T his type of plasma is also known as d via the configuration shown in Figure 1 4 (A). In (A), a small scalpel that needs to be sterilized would be placed in the discharge gap between the dielectric barrier and grounded electrode since this is where plasma is generated. In (B), the same small scalpel would be placed on top of the dashed black surface, since this is where the plasma would be generated. When high alternating current (AC) voltage is applied to one of the electrodes, resulting electric field is adequate to produce ionization of the gas in/above the gap. The radicals, ions and electrons produced are attracted towards the electrodes of opposite polarity and form a charge layer on the surface of the dielectric. This accumulated charge cancels the charge on the electrodes, so that the e lectric field in the gap falls to zero and the discharge stops. Hence a low current, low power discharge is obtained. Table 1 2 Micro discharge properties in air at atmospheric pressure [ 24 ] Property Range Duration Filament radius Peak Current Current density Total Charge Electron density Mean Electron Energy Filament Temp 10 9 10 8 s ~10 4 m 0 .1 A 10 6 10 7 Am 2 10 10 10 9 C 1 0 20 10 21 m 3 1 10 eV Close to average gas temperature in the gap When electric field is sufficiently high to cause breakdown of the discharge gas, a large number of micro discharges can be observed emanating from the electrodes.
26 Micro discharges are thin, conductive c hannels that are formed w hen a voltage difference is applied to the discharge gap, thus causing a critical stage in the electron avalanche, wherein extremely fast streamer formation is possible [ 24 ] These micro discharges spread uniformly along the surface of a dielectric and are shown in Figure 1 5. Figure 1 5. End on view of micro discharges in atmospheric pressure air [ 24 ] The gap width is of the order of a few millimeters. Since the dielectric layer in between cannot pass DC current, these devices requires AC voltage. The dielectri c acts as ballast it imposes an upper limit on the current density in the gap. Typically DBDs are operated at 1 100 kV and frequencies of 50 Hz 1 MHz. At higher frequencies, it becomes tougher to impose the dielectric limitation on the current density [ 24 ] For a long time, DBDs were primarily utilized in industrial ozone generators [ 26 ] Apart from this, DBDs are also implemented in surface modification [ 27 ] plasma chemical vapor deposition, pollution control [ 28 ] excitation of CO 2 laser and plasma display panels. More recently, DBD plasma at atmospheric pressure has found newer
27 applications in the medical and sterilization industries Kalghatgi et al. [ 29 ] reported experiments testing the effect of DBD plasma on endothelial cells, reporting that low power non thermal plasma is relatively non toxic to endothelial cells at short exposure times. This enables the application of plasmas to therapeutic applications such as wound healing [ 30 ] and blood coagulation [ 31 ] However prolonged plasma exposure ca n also have an adverse effect on malignant cells [ 32 ] Hence plasma also finds application in the sterilization industry [ 33 34 ] Dermatology [ 35 ] and dentistry [ 36 37 ] are two other areas wherein DBD plasma has been identified as an easily accessible, effective met hod of sterilization. Plasma sterilization meet s all criteria of a n ideal sterilant, listed by Moisan et al. [ 16 ] A s a good alternative to conventional sterilization techniques, it has also sparked off a remarkable volume of research aimed at u nderstanding its underlying mechanism. Understanding the mechanism of sterilization is vital to the successful implementation of this technology. Before discussing the possible mechanism, a brief summary of research in plasma sterilization is necessary. 1.3 DBD Plasma Sterilization The origins of plasma sterilization can be traced all the way back to 196 8 when Menashi [ 38 ] filed a patent for a high temperature/ high pressure p lasma sterilization process. A separate patent was filed in 1972 for a low temperature plasma sterilization process. Further patents by Boucher and Bithell [ 39 40 ] cemented a growing in terest in plasma sterilization. Experiments consisted of placing a contaminated object inside a chamber as shown in Figure 1 6 below and generating plasma inside the chamber to sterilize the object.
28 Figure 1 6. Schematic of an apparatus used earlier on for plasma sterilization [ 39 ] Early experiments, during this decade, were conducted at low pressures of 0.1 10 Torr and mostly used Helium or Argon as discharge gases. Further experiments consisted of refining the experimental setup used and varying experimental parameters such as discharge gas used (Ratner et al. [ 41 ] showed that plasma sterilization is efficient with common discharge gases such as N 2 O 2 air etc. ), input power density (Boucher also reported that sterilization efficacy increased with RF power density absorbed in the discharge) and type of micro organism (difference between plasma sterilization times take for Gram negative and Gram positive pathogens was identified early on by Baier et al. [ 42 ] ). Researchers also speculated on the role of UV radiation and oxygen atoms in the sterilization process. However while many were divided about whether UV radiation played a role [ 43 45 ] some experiments conclusively proved tha t sterilization via an O 2 plasma was much more effective [ 46 ] The experiments during the 70s and 80s were mostly in the volume plasma range. The later part of the 20 th century and early 21 st cent ury sparked off research in surface plasma sterilization at atmospheric pressure. Initial research was published by Roth et al. [ 47 ] This was the time when sterilization using glow discharge at atmospheric
29 pressure and more, specifically, DBD sterilization was recognized as a much more accessible alternative to earlier low pressure plasma sterilization systems. A flurry of publications followed from numerous research groups pursuing dielectric barrier discharge sterilization [ 48 52 ] Research in plasma sterilization can be classified in to two categories: (i) Parametric studies of the factors involved (ii) Studies aimed at determining the underlying mechanism. The former involved researchers embarking upon an empirical path, varying each and every factor and noting its significance on th e plasma sterilization time as well as the efficiency and efficacy of the process while the latter involved in depth studies aimed at understanding the role of several plasma agents in the plasma sterilization process. Both categories of research were purs ued using experimental methods primarily Another area of research, that could potentially provide a more fundamental insight into the plasma sterilization process, is the numerical modeling of plasma sterilization. However numerically modeling the interac tion of plasma with the biological cell is a highly complex process and is perhaps the reason why very little research exists in this area. More recently, plasma interaction with biofilms has also been investigated as a potential sterilization method espe cially in dental applications The next few sections briefly describe a summary of literature available in each of these four research areas. 1.3.1 Parametric S tudies of the F actors A ffecting P lasma S terilization Lerouge et al. [ 14 ] outlined a schematic of the numerous factors involved in plasma sterilization. This schematic, modified for DBD plasma sterilization is given below in Figure 1 7.
30 Figure 1 7. Va rious Factors involved in DBD Plasma sterilization Given below is a brief summary of the research conducted by various authors in evaluating the various factors affecting plasma sterilization. Gas Composition: Plasma can be generated using various dischar ge gases: Air, N 2 O 2 He, Ar, O 2 +CF 4 etc. Over the decades, extensive research has been conducted in determining the best possible gas mixture for optimum plasma sterilization. Hury et al. [ 53 ] studied the destruction of B acillus subtilis spores in oxygen(O 2 ) based plasmas sustained in the mTorr pressure range. They c onfirmed 2 plasma achieved more killing than Ar plasma. Similarly Lerouge et al. [ 54 ] conducted experiments in the mTorr range, wherein different gas compositions were compare d in terms of destruction efficiency. They found that the O 2 /CF 4 plasma was most effective, due to the combined etching action of both oxygen and fluorine atoms. Most of the O 2 tests were conducted in a low pressure regime. However in 2006, Ying et al. [ 55 ] compared yeast inactivation in He, Air and N 2 DBD (volume) plasma at atmospheric pressure W orking at a frequency of 0 20 kHz and a n
31 input voltage of 40 kV pp for a treatment time of 5 minutes, they reported a 10 5 10 6 and 10 7 reduction in bacterial concentration using plasma generated in N 2 air and He respectively Th ey concluded that for their case, the electrostatic tension caused the rupture of the cell membrane, leading to cell death and that the effect of this electrostatic tension was exacerbated in the case of He plasma. Gas Pressure : Moisan et al. [ 16 ] state the three pressure regimes in which most plasma sterilization experiments have been conducted: low pressure (1 10 mTorr), medium pressure (0.1 10 Torr) and atmospheric pressure. Until a couple of years ago, most reported experiments pertained to t he medium pressure range. Using a n RF discharge reactor operating at 13.56 MHz and a discharge gas mixture of O 2 /CF 4 Wrobel et al. [ 56 ] concluded that rising pressure produce d competing effects in plasma; upto a certain limit it increase d residence time of the gas molecules, thus increasing the concentration of active species, possibly promoting sterilization effectiveness. Howev er, beyond a certain point, increasing pressure decreased the plasma volume and increase d the gas temperature. Kao et al. [ 57 ] noted that when exposed to a microwave (MW) plasma, comple te inactivation of E scherichia coli was noted at lower pressures like 43 200 mTorr. However a higher pressure of 400 mTorr showed incomplete inactivation. The effect of low pressure versus atmospheric pressure is still contested It has been argued that l ow pressure allows the emission of Vacuum UV (VUV) radiation ( This will be discussed in detail in a later section. On the other hand, it is commonly agreed that at atmospheric pressure the lethal UV photons/oxidizing atoms produc ed
32 are easily reabsorbed, thus eliminating the possibility of UV irradiation or etching by oxidation. Input Power and Frequency : The role of power has always been coupled with the role of surface temperature. Rising power leads to an increased concentrati on of active species in the plasma, which could lead to increased microbiocidal and sporicidal activity. Bol'shakov et al. [ 58 ] conducted experiments using a n RF O 2 plasma source and D.radiodurans as the test pathogen, wherein they concluded that sterilization time depended inversely on plasma density and fluxes of active species (O and O 2 *), whereas these fluxes increased linearly with input power but changed weakly with pressure under their conditions. However increased power also implies increased heating of the substrate, which could be detrimental to the materials be ing sterilized. In discussing the effect of input frequency on plasma sterilization, the debate has always been about r adio f requency (RF) versus microwave (MW) plasma. Figure 1 8. Comparison of spore mortality [log10(No/N)] in MW and RF O2/CF4 plasmas after 5 minutes of exposure (P=200W, p= 80 mTorr, F= 80sccm, [CF4]= 12%) [ 59 ]
33 RF frequencies belon g to the range of 3 kHz 300 GHz, while MW frequencies occupy the upper range of RF frequencies (0.3 300 GHz). Lerouge et al. [ 59 ] have reported experiments wherein they noted high sporicidal activity of MW plasma (2.45 GHz) as compared to its RF counterpart (13.56 MHz). Figure 1 8 above shows this comparison of RF and MW plasmas using O 2 / CF 4 Authors c oncluded that the frequency determine d the Electron Energy Density Function (EEDF), which determine d the concentration of high energy electrons in MW plasma, as compared to its RF counterpart This was hypothesized as the reason for the higher sporicidal a ctivity observed with MW plasma. Effect of Afterglow : Direct p lasma sterilization refers to experiments in which the bacterial sample is directly exposed to the generated plasma. A fterglow based sterilization on the other hand, indicates experiments in w hich plasma is generated and the bacterial sample is exposed downstream of the reactive chemical species produced during plasma generation (afterglow). Microwave experiments using the afterglow of N 2 O 2 plasma to achieve complete inactivation of B. subtili s spores within 40 min wit h an absorbed power of 100 W were also reported [ 60 ] However, authors noted that the efficacy of such a system was highly dependent on the gas flow reaching all parts of the object to be sterilized and on short lived active species being transported with sufficient rapidity. E xperiments using the afterglow of non thermal plasma at atmospheric pressure to inactivate E. coli and B. cereus (within 15 minutes) and P. aeruginosa (within 10 minutes) have also been reported [ 61 ] In the case of afterglow, the bacterial concentrations are in essence being exposed to the reactive chemical species produced during plasma generation i.e.
34 neutrals and charged species. E xperiments using the reduced pressure afterglow stemming from d ischarge in a N 2 O 2 mixture were reported [ 62 ] wherein it was concluded that sterilization time was the shortest when the O 2 percentage in the mixture was set to maximize UV emission intensity. This was an example of sterilization being influenced by the UV photons in the plasma afterglow. Figure 1 9. Results of inactivation of E. coli on the agar surface by direct (a,b) and indirect (c,d) plasma treatment [ 63 ] As is evident, direct plasma leads to a clean spot in the cent er with no bacterial growth while indire ct plasma leads to incomplete inactivation. Dobrynin et al. [ 63 ] also report ed experiments wher ein they compare d bacterial samples exposed to direct plasma (discharge is ignited on the treated surface) to those exposed to indirect plasma (a grounded metal mesh is used as a second electrode, thus cutting of f charged species. The mesh wa s corrected to produce the same amount of UV radiation). They conclude d that direct plasma wa s more effective than indirect plasma due to the combined action of charged species, neutrals and UV photons. This comparison is shown above in Figure 1 9. Thus, with the use of a plasma afterglow for
35 sterilization, the agent responsible for sterilization in such a case remains highly debated. In general, a review of literature commonly indicates greater sterilization times using the afterglow, as compared to when using direct pl asma. Miscellaneous Factors : Various other factors have also been debated: nature and surface density of test pathogens, packaging, geometrical factors involved in the design of the sterilization reactor, gas flow rate etc. For instance, the type and surfa ce density of test pathogens is a huge factor in determining efficiency of plasma sterilization. Vegetative pathogens ( E. coli y east) are usually less resistant than bacterial or fungal spores ( G stea r othermophilus, B. subtilis ) Hury et al. [ 53 ] also concluded that the surface density of spores used was an importan t factor in determin i ng plasma sterilization efficiency i.e. higher the surface density of spores, longer the sterilization time. Presence of absence of packaging material was another factor considered to determine the effectiveness of a plasma sterilizati on process. In order to ensure that an object sterilized using plasma remains sterile until use, an alternative method could be to enclose the object in a polymeric package and expose it to plasma. However Lerouge et al. [ 14 ] demonstrated that this led to negligible reduction in spore population, due to the reduced number of active species passing through the package material. To date, it is suggested tha t any object be first plasma sterilized and then immediately inserted into sterile sealed packages until further use. It is evident from the volume of research reviewed above that a plethora of factors influence plasma sterilization. While the role of som e factors (nature and surface spore density of test pathogen, role of packaging) has been clearly identified, the role of others (discharge gas, power, and pressure ) still continues to be debated. However one
36 thing is clear: All the debatable factors in pl asma sterilization are coupled, meaning that more often than not, for efficient plasma sterilization, a trade off between the different factors will have to be considered. 1.3.2 Mechanism of Plasma Sterilization DBD p lasma is a soup of UV photons and reactive c hemical species at a slightly elevated temperature In understanding the mechanism of plasma sterilization, it is necessary to evaluate the roles of each of these components. This section reviews the majority of the literature published in evaluating these roles. UV photons: The UV ra diation spectrum is gi ven below in Figure 1 10 below. Figure 1 10. Wavelength and energy of radiation in the UV and visible portion of the spectrum [ 59 ] The vacuum ultra violet (VUV) ranges from 10 200 nm while the UV portion of the spectrum ranges from 200 380 nm. At atmospheric pressure, emitted VUV radiation is usuall y absorbed by atmospheric oxygen. Of the VUV range, Far UV radiation (100 s of micro organisms. C, UV B and UV A. There are two main mechanisms for UV damage (a) direct effects of
37 UV radiation that are based on UV energy absorp tion by cellular macromolecules. Typically, UV C and UV B are capable of induc ing a reaction between two pyrimidine molecules (thymine and c ytosine, adjacent to each other on the same strand of DNA) causing them to form a dimer. The presence of this dimer affects base pairing and cause s mu tations during DNA replication [ 64 ] (b) DNA, protein and lipid alterations caused by the UV A induced disturbance in the cellular redox state. UV A can induce the release of intr acellular reactive species (RS), causing oxidative degradation of lipids and DNA [ 65 ] The role of UV radiation in plasma sterilization has often been closely con nected nm), which has been argued to be one of the key factors in influencing plasma sterilization. Lerouge et al. [ 59 ] hypothesize d that very energetic VUV photons emitted in a glow discharge plasma may have a greater effect on spores by attacking not only DNA but a lso the spore membranes. However in conducting experiments at low pressure using both RF and MW plasmas, they note d that VUV radiation in the range 115 170 nm required more than a 5 minute exposure to kill 90% of the micro organisms. From their observatio ns, they conclude d that VUV radiation did not appear very promising, since it was not more efficient in killing micro organisms than UV C (254 nm). On the contrary, Halfmann et al. [ 66 ] conducted experiments wherein they determined that the wavelength range of 235 300 nm played a maj or role in sterilizing spores of B acillus atrophaeus. They also concluded that active species played a minor role but we re not negligible. Most of the earlier work reported in plasma sterilization has been reported in the low and medium pressure regime s
38 The more recent body of work in plasma sterilization concentrates on sterilization due to glow discharge and dielectric barrier discharge volume plasma at atmospheric pressure. It has often been speculated that in higher pressure regimes (such as atmospher ic pressure) the more reactive VUV photons and oxidizing atoms produced during plasma generation often recombine instantly or are re absorbed instantly [ 67 ] thus making them unavailable for sterilization. Authors have reported experim ents confirming that the role of UV radiation in plasma sterilization at atmospheric pressure is minimal. Experiments were conducted in which the spectroscopic signature of DBD volume plasma in air at atmospheric pressure was measured [ 68 ] Their results (Figure 1 11 below) show ed that no significant UV emission occurs bel ow 285 nm, which led them to conclude that UV might not play a significant direct role in the sterilization process by low temperature air plasmas. Figure 1 11. UV spectrum of a DBD in air in the 200 300 nm wavelength range [ 68 ]
39 Dobrynin [ 63 ] reach ed a similar conclusion after conducting experiments examining the effect of plasma on inoculated slides, protected from direct discharge by a MgF 2 slide (which is tran sparent to VUV photons >140 nm). Thus far, based on all the literature published on this subject, it seems that the only clear consensus on the role of UV radiation in plasma sterilization is that different pressure regimes lead to the emission of differe nt ranges of UV radiation, which in turn might influence the process of plasma sterilization. Reactive Chemical Species: Among the primary products of electron collision are atomic and metastable oxygen and nitrogen, with subsequent reactive collisions p roducing a cocktail of neutral and ionic species. Atmospheric pressure discharges differ from low pressure plasma in that their chemistry is dominated by reactive neutral species such as oxygen atoms, singlet oxygen and ozone rather than ions. In corona an d DBD, ozone is the main reaction product. In other plasma sources, oxygen atoms represent a larger proportion of the reactive species. Ozone was found to be the dominant species with single oxygen and atomic oxygen being at a concentration five to six ord er lower than that of ozone [ 25 ] Other reactive species produc ed by plasma down hydrocarbons, chlorocarbons and CFCs [ 24 ] As disc ussed in Section 1.3 .1, Dobrynin et al. [ 63 ] evaluate d the difference between direct plasma (ba cterial sample directly exposed to discharge) and indirect plasma (grounded mesh inserted in between the discharge and the bacterial sample, thus blocking the flow of charged species). In doing so, they conclude d that the blocking of charged species led to a reduction in inactivation effect of the discharge. C harged
40 particles wer e hypothesized to play a role in plasma sterilization by way of creating electrostatic tension on the outer surface of the bacterial cell membrane causing the rupture of the cell me mbrane and killing the bacteria [ 69 ] However a detailed study of the interaction of charged pa rticles with cells and tissues concluded that the experimental difficulties in detecting and characteri zing charged species form an obstacl e to obtaining direct evidence [ 70 ] On the other hand, ne utral species (O 3 NO 2 OH) have also been known to play a role in plasma sterilization. The efficacy of the plasma afterglow in sterilizing bacteria is mainly due to role of neutral species. Often times, charged particles are too short lived to be able to reach the bacterial sample in the afterglow region Laroussi and Leipold [ 68 ] state d that cell membranes, made up of lipid bilayers, whose important component is unsaturated fatty acids, may also be attacked by OH radicals, causing them to break down. It was also speculated that o xygen based and nitrogen based reactive spe cies may have strong oxidative effects on the outer structures of cells [ 71 ] The germicidal action of ozone has been well documented [ 72 73 ] However the role of ozone in plasma sterilization has not been discus sed in great detail. Efremov et al. [ 74 ] discuss ed reaction r ate constants for production and destruction of ozone formed during plasma generation. They conclude d that under a small specific energy input, in dry air, there are insufficient fast processes for O 3 molecule destruction with rates comparable to the rate of its formation. This led them to also conjecture that the antiseptic property of the excited dry air flowing out of a discharge chamber is determined by its ozone concentration. They follow ed up this conclusion by demonstrating in their paper that expos ing micro organism concentrations to discharge
41 excited air, even for a short while, substantially reduced their amount. However in doing so, the composition o f the discharge excited air is not discussed. Dobrynin et.al. [ 63 ] pursue d a different approach in isolating the role of ozone in plasma sterilization. They measure d the ozone concentration pr oduced by a DBD discharge in room air at ~60% relative humidity as 28 ppm. Consequently they use d a commercial ozone generator (~500 ppm max output, Quinta Inc.) to produce the same concentration of ozone and examine d the inactivation effect of this ozone concentration on E. coli and skin flora. They note d that no inactivation effect occur ed However in this case, it is to be noted that the levels of ozone noted were ~28 ppm, which might be too l ow a concentration to have an inactivation effect on bacterial concentrations. Temperature: Although glow discharge and DBD plasma at atmospheric pressure are primarily low temperature plasmas, the role of temperature or heat in plasma sterilization still needs to be discussed. Figure 1 12. Increase of sample te mperature vs. plasma dissipated power for a DBD volume plasma in air at atmospheric pressure [ 68 ]
42 As already mentioned in Section 1.3.1, increased power can cause increased heating of the substrate, which in turn could contribute to killing. The effect of temperature was evaluated by exposing B. subtilis spores to CO 2 pla sma and evaluating the effect of temperature in this case [ 53 ] It was concluded that the lower temperature of 15 o C led to a lower destruction efficiency while a higher temperature of 60 o C led to the highest destruction efficiency. However, Laroussi et al. [ 68 ] measured the gas temperature as well as the temperature in a sample placed 2 cm away from an atmospheric pressure DBD discharg e in air. They observed that gas temperature remained close to room temperature and that a variation in power from 1 7 W showed negligible change in temperature (Figure 1 12). At typical running power levels, a maximum increase of 21 o C was observed, which led authors to conclude that no substantial thermal e ffects on bacterial cells occur. This led them to conclude that heat does not play a major role in killing bacterial cells. Ohkawa et al. [ 75 ] report ed that using a pulse modulated high frequency plasma sterilization source ca used a decrease in sterilization time with inc rease in neutral gas temperature. Thus,once again, contrary results we re presented. Thus, in conclusion, the main agents involved in the process of plasma sterilization are UV photons, reactive chemical specie s (charged and neutral) and temperature. Literature presents several cases that argue for and against each of these agents. The common hypothesis about plasma sterilization that was initially propounded, came as a result of the phasic behaviour seen in pl asma sterilization. A schematic illustrationof this phasic behavior is shown below in Figure 1 13.
43 Figure 1 13. Schematic illustration of the different phases in a plasma sterilization survival curve [ 67 ] A survival curve is one that plots the reduction in number of micro organisms over time. While all other co nventional sterilization methods seemed to produce linear, mono phasi c survival curves, plasma sterilization produced bi phasic and tri phasic survival curves [ 67 76 ] Tri phasic and bi phasic survival curves led authors to initally believe that plasma sterilization followed a phasic behaviour in which UV photons inactivated the top layer of spores r apidly and any remaining debris was then etched away by a combination of UV photons and ROS [ 16 ] This hypothesis might still hold true, though it has not been proved conclusively What needs to be determined is a clear mechanism of interaction o f these killing agents with biologi cal pathogen, and the order of events that causes cell destruction.
44 1.3.3 Plasma Interaction with B iofilms A major volume of work done in plasma sterilization focuses on the effect of plasma on individual microbe concentratio ns. However, recent research has also focused on the fact that most microbes prefer to live as part of communities where interactions take place [ 77 ] Biofilms are microbial communities attached to a surface and embedded in a matrix composed of exopolysaccharides together with proteins and excreted nucleic acids. Work involving the use of plasma in eradi cating biofilms increases in frequency from 2007 onwards. Sladek et al. [ 78 ] reported experiments with S treptococcus mutans biofilms using a plasma needle (13.56 MHz, 100 mW, t=60s) and reported incomplete inactivation. This could be because of the low exposure time. The interaction of plasma (atmospheric pressure, He/N 2 gas mixture, with an input power of 4.8W) with bacterial biofilms w as visualized through AFM [ 79 ] AFM images show minor morphological changes to cells in 5 minutes, but major cell damage in 60 min utes. Lee et al. [ 80 ] used a 2.45 GHz, 1 kW MW induced Ar gon plasma source to completely inactivate different bacterial biofilms ( E. coli methicillin resistant Staphylococcus aureus ( MRSA ) ) in 20s. SEM micrographs show damaged morphologies of cells, as compared to untreated cells. The use of plasma in biofilm inactivation holds major potential in denta l health and food processing industries. For the purpose of this study the interaction of plasma with biofilms has been studie d because (a) Biofilms are more robust in structure, which enables their exposure to plasma and consequently, microscopic analysi s (b) A single species biofilm encourages the organization of bacteria into colonies, with each colony being capable of one specific function. Some might be responsible for the exo polysaccharide matrix formation while others might be more motile. Studying the plasma
45 interaction with biofilms microscopically helps determine which cellular function is affected the most by plasma. 1.3.4 Numerical M odeling of Plasma Sterilization O ne aspect of research that has been minimally researched is the possibility of constructing a numerical model to simulate plasma sterilization. Some researchers have modeled the destruction of cells via innovative methods Kumar et al. [ 81 ] report ed an experimental as well as numerical study performed by exposing spores to elevated temperatures. They simulate d the flow field inside the thermal exposure system using a turbulence model and buil t another model to simulate the thermal response of the spores in a high temperature gas environment. The two models we re clubbed t o investigate further experimental parameters such as the dependence of this thermal response on water content in spores and thermal property uncertainties. Gallagher et al. [ 82 ] provide d a numerical characterization to help predict and understand the inactivation mechanism of DBD plasma. Their simple exponential model use d rate constants (chemical kinetics) to solv e for species concentration. Akishev et al. [ 83 ] use d an empirical mathematical approach to predict bacterial inactivation, using not only cell inactivation data, but also cell reparation data. A new venue of numerical modeling in plasma sterilization was opened up by solving for different species concentrations using a hydrodynamic model of equations [ 84 85 ] The most exciting research in this area was reported by Babaeva and Kushner [ 86 ] They conducted computational studies on the interaction of plasma streamers in atmospheric pressur e DBDs with human skin tissue. In Figure 1 14 (A) below, t he skin wa s assumed as consisting of four layers: outer membrane, epidermis, inner membrane
46 and dermis. Each layer wa s assigned a different such that there were four different conducting layers. Figure 1 14. Numerical modeling of the plasma interaction with skin A ) Computational Model used, depicting the breakup of the skin into different layers as well a s the plasma source placement B) Electric field inside the epidermal layer, at different simulation time points [ 86 ] The plasma source wa s modeled along the lines of the floating electrode DBD [ 31 ] wherein plasma (source of electrons) wa s positioned above the surface of the skin. The propagation of a plasma filament towards the surface of the skin penetrating the surface of the skin and its propagations inside wa s modeled. Plasma wa s modeled by solving for the complete set of plasma air wa s modeled throughout the domain. The plasma filament propagate d from the source and hit the surface of the skin at 1.1 ns. In Figure 1 14(B), it wa s observed that at 0.7 and 0.9 ns, before the filament hit the skin surface, the electric field wa s fairly low. It skyrocket ed at 1.1. ns (when the plasma filament hit the skin surface) with the topmost layer of the skin showing the
47 highest electric field distribution. This study wa s especially useful in that it laid out a graphical visualization of what happen ed when a plasma filament interfere d with the skin and promote d the hy pothesis that the induced electrical field inside the cells caused cell electroporation. The objective of Section 1. 3 was to provide a summary of the research in plasma sterilization. From 2000 onwards, a huge amount of research has been done in understan ding plasma sterilization. However the fundamental questions still remain the same (a) What is the best combination of experimental factors to achieve efficient and effective plasma sterilization (b) What is the mechanism of plasma sterilization? Which pla sma agents are responsible for killing? What happens to the cells when they are exposed to plasma? The goal of th is study is to provide reasonable answers to both questions.
48 CHAPTER 2 RESEARCH MOTIVATION Chapter 1 gave a detailed overview of conventiona l sterilization methods, the characteristics of DBD p lasma and more importantly, a voluminous introduction to plasma sterilization. The pros and cons of different methods of plasma sterilization have been abundantly outlined in Section 1. 3 The fundamenta l questions highlighted at the end of Chapter 1 were : What is the best combination of experimental factors to achieve fast and effective plasma sterilization (i.e. complete bacterial inactivation on plasma exposure in the shortest time) ? What is the mechan ism of plasma sterilization? Which plasma agents are responsible for killing of microorganisms? The key to answering the first question is to design a set of experiments aimed at isolating each parameter and studying its significance on the time taken for plasma sterilization. As has been mentioned in Chapter 1, there is no one set of optimal parameters for safe, efficient and effective plasma sterilization. The best conclusion to be drawn is that every experimental parameter requires a trade off. For inst ance, too high a temperature is good for effective sterilization, but bad in terms of dielectric surface heating and damage to the substrate materials. The goal in such a scenario would be to find a suitable dielectric material that can stand high temperat ures, but at the same time be usable for effective sterilization. Similarly other experimental parameters to be investigated for determining an optimum set of experimental parameters for the case of DBD Plasma Sterilization would be a) Type of Micro or gani sm b) Input Power c) Input Frequency d) Nature of Dielectric Material.
49 The key to answering the second question above is to use diagnostic equipment or chemical reagents to analyze the role of a single plasma component in the process of plasma sterilizati on process. As regards the mechanism of plasma sterilization, three responsible plasma components can be identified: UV photons, reactive chemical species and temperature ( to a lesser extent ) In order to figure out whether these three work synergistically to kill bacteria or whether one component plays a dominant role in sterilization, diagnostic equipment such as spectrometers (to study the emission patterns of the UV photons), fluorescent and electron microscopic analysis (to study the damage to the biol ogical cells after plasma exposure) or chemical reagents (reacting specifically with different plasma chemical species or by products ) can be used. The goal is to figure out a mechanism that explains the systematic breakdown of the microorganism. Thus pla sma sterilization provides for an intriguing field of research, with its fair share of challenges. These challenges need to be overcome so that plasma sterilization may be implemented as a safe, efficient and effective alternative to conventional methods o f sterilization. Chapter 3 discusses the experimental setup and methodologies used for the DBD surface plasma experiments. Chapter 4 focuses on outlining the parametric studies conducted in an effort to characterize the sterilization efficiency of the DBD surface plasma setup used in this thesis. Chapter 5 outlines the experiments conducted in understanding the mechanism of DBD surface plasma sterilization. Chapter 6 summarizes all the results obtained in Chapter 4 and 5 and outlines areas of future work.
50 CHAPTER 3 METHODS AND MATERIALS 3 .1 Experimental Setup u sed f or Plasma Sterilization Experiments Figure 3 1 shows the schematic of the experimental setup used in plasma generation. It is to be noted that this experimental setup is very rudimentary and w as used in the initial stages of this project. This experimental setup has subsequently been built into a more compact form, that uses the same electrical components, but in a much more power efficient way. Figure 3 1 Schematic of the experimental setu p used In Figure 3 1 a function generator (HP 33120A) is used to generate a n RF sine wave of frequency 14 kHz. The power of this signal is then amplified by using an amplifier (model Crown CDi4000). This amplified signal is then passed through a step up t ransformer (Corona Magnetics, Inc.) which steps up the voltage. The input power from the transformer is fed to the powered electrode (red) of the device via a metal connector. The final signal being fed into the plasma device has an input voltage of 12 kV p eak p eak (pp) The other electrode (blue) of the device is grounded via a grounded
51 electrical bench, atop which the device sits. The powered and grounded electrodes are separated by a sheet of dielectric material, about 1.6 mm thick. Th e design of the pla sma device is further described in Section 3 .2. The spectroscopic signature of the generated DBD plasma is determined using the Ocean Optics USB 2000+ spectrometer. This spectrometer has a detector range of 200 1100 nm, an optical resolution of ~0.3 10 n m (FWHM), a dynamic range of 1300:1 for a single scan and is fitted with a custom made grating designed to be sensitive to wavelengths between 200 650 nm. An uncoated UV Fused Silica Plano Thor Labs, Inc. ) is used to collect and focus the incident plasma glow from the plasma device, which is then detected by the spectrometer via a fiber optic probe. Baseline spectroscopic data for each device was collected with the device powered for 2 minutes at a sampling rate of 10s. Readings were also taken during sterilization experiments. A 2B Tech Ozone meter is used to measure the emitted ozone at fixed time intervals. This ozone meter operates on the principle that the maximum absorption of ozone takes place at 254 nm. Air is drawn in to the ozone meter at a flow rate of about 1L/min and passed to an absorption cell via two methods: (i) directly and (ii) after passing through an ozone scrubber. The intensity of light passing through the absorption chamber in case (i) and (ii) is measure d and used to determine the level of ozone in absorbed air. Air is sampled every 10s and the ozone meter has an accuracy of about 2%. The sampled ozone levels are saved to a computer via a LabView Interface.
52 An FLIR A320 Infrared camera is used to reco rd a thermo graphic mapping of the electrode surface area, while it is being powered. This helps provide a visualization of the temperature fields du ring plasma generation which would further assist in understanding the role of temperature in plasma steri lization. The A320 operates at a spectral range of 7.5 between the plasma device and infrared camera, ambient temperature and humidity and the emissivity of the FR4, SC dielectric were measured to be 0.2667 + 0.0127 m, 24.4+2.3o C, 59 + 3% RH and 0.9097+0.03, 0.929+0.03 respectively. 3.2 Design of the Plasma Device The devices themselves have two important components: the dielectric surface and the electrodes. The two dielectrics used are Flame Retardant 4 (FR4) and semi ceramic (SC). FR4 is a co mposite material composed of woven fiberglass cloth with an epoxy resin binder that is flame resistant. FR4 is used as the primary insulating backbone in a vast majority of printed circuit boards (PCBs). For the purpose of this thesis, commercial copper cl ad FR4 sheets (Advanced Circuits ) were used to manufacture the FR4 devices. 2 layer, 1.6 mm thick FR4 sheets, overlaid 1 Oz copper (Cu) are milled in the requisite electrode pattern. The Cu layer was coated with a tin (Sn) finish. The FR4 used for the ma nufacture of the boards has a dielectric constant ( of 4. 7 The other dielectric material, that has been used for testing is Rogers 3 003C semi 00 + 0.04 For the manufacture of the SC devices, SC boards (of the same thickness as FR4), copper cla d with 1 oz thick copper foil, wer e etched (via immersion into ferric chloride (FeCl3 solution) into the requisite electrode pattern.
53 For experiments in this study, bacterial samples we re deposited on the plasma device and subjected to the effect of plasma. The device has a surf ace plasma configuration, as shown in Figure 1 4(B) in Chapter 1. It consists of a sheet of dielectric, both sides of which are embedded with electrodes. The grounded electrode is a square sheet of metal Various designs were tested out for the top electro de, as shown in Figure 3 2 before deciding on the comb like design in Figure 3 2 ( C ). The top electrode is typically powered (i.e. input voltage is supplied to this electrode and plasma is visible on this electrode surface) during sterilization experiments. Figure 3 2 ( on the opposite side of the dielectric surface ) Figure 3 2 ( A ) shows a sawtooth like electrode design for th e powered electrode. Figure 3 2 ( B shaped electrode design. Figure 3 2 (C) shows a comb like electrode, which is the current electrode design used. In all three design configurations shown in Figure 3 2 the light grey colored square outlining the electrode represents the sheet of copper (grounded electrode) embedded on the opposite side of the dielectric surface. Figure 3 3 below illustrates the powered devices. In all three images, a red arrow denotes the point at which a metal connector is at tached to the electrode surface.
54 This is the point through which input voltage from the transformer is supplied to the electrode. Figure 3 3 The three electrode configurations, shown in Figure 3 2 when powered Initial feasibility tests with S accharom yces cerevisiae ( b y east) were 3 3 ( A ). However, over time, it was observed that experimental results with this design were beginning to show inconsistency. The reason became apparent a fter a couple of stamp tests. A stamp test comprises a device being inoculated with a bacterial sample (i.e. the sample is deposited on top of the powered electrode and spread uniformly over the electrode surface), powered for a required time interval and then stamped face down onto an agar plate. This agar plate is then incubated for 24 48 hours. For a device that has been completely sterilized, there should be no visible CFU (colony forming units) on the agar plate for 20 minutes All three electrode desi gns were subjected to the stamp test. The results are shown in F igure 3 4, given below. In Figure 3 3(A), it is evident that when the device is powered, plasma completely covers the dielectric surface in between the electrodes, but the actual electrode su rface
55 area is not enveloped by plasma. This observation is supported by the results from the stamp tests, as shown in Figure 3 4(A) below. In this figure, it i s observed that the surface area on the agar plate, covered by the sawtooth electrode itself, is dotted with a number of colonies while the rest of the plate is clean. Thus the wide electrode provides a haven where the test organism survive s This was not seen with narrower electrode configurations. Figure 3 4 Stamp Test ( at 60s) for three diffe rent electrode designs A ) Sawtooth B electrode C ) comb like electrode design electrode design, shown in Figure 3 3 ( B ) was initially devised to combat the problem with the sawtooth electrode. However, as the stamp test from Figure 3 4 ( B ) indicates while the overall vulnerability to incomplete sterilization is reduced due to a simpler design, the thickness of the two main electrodes is still too large for plasma sterilization to be effective. This conclusion is corroborated by the image of the powe red electrode device, shown in Figure 3 3 ( B ), wherein it is evident that plasma encompasses the entire inoculated area, except the thick electrode surface. These observations led to the current comb like electrode design, shown in Figure 3 2 ( C ) and 3 3 ( C ). The entire electrode surface area is covered with uniform plasma, as is evident from Figure 3 3 ( C ). Although the connector electrode for this design is 1 2 mm thicker than the individual electrodes, the plasma coverage seems to
56 compensate for this vul nerability, as is evident from the stamp test in Figure 3 4 ( C ). Therefore, this electrode design was chosen for all subsequent experiments. This set of experiments also emphasized the significance of electrode surface area as a factor in plasma steriliz ati on: Lesser the width of the electrodes, more efficient the sterilization. 3 .3 The Portable Sterilization S etup The APRG lab, wherein most of the experimental work has been done has been set up as a BSL I (Bio safety level I) facility. For experiments wit h BSL II pathogens (i.e. experiments had to be conducted at the Emerging Pathogens Institu t e (EPI), UF, which is a BSL II Facility. However this posed a problem, since the experimental setup shown in Figure 3 1 is quite cumbersom e and non was desirable Such a portable experimental setup has important real world applications with battery operated, portable sterilizers desirable for scenarios such as triage situations in third wo rld countries or disaster relief situations. A portable experimental setup, shown in Figure 3 5 below (shown in Figure 3 1) was developed by Raul Chinga 1 using appropriate electrical components. The power supply for this setup achieves an output voltage of 1 0 kV pp at 47kHz. The whole setup measures about 10.16 x 6.35 cm 2 The previous setup utilized a crown CDI4000 audio amplifier, which is great for audio amplification purposes, but very inefficient for amplification of single sinusoidal signal coming f rom the function generator. Compared with the previous set up used, this design operates at a single frequency band, which 1 The author profusely thanks Raul A.Chinga, of the Department of Electrical Engineering ,UF for his invaluable work in building the portable experimental setup. Tests with BSL II pathogens woul d have been infinitely more difficult to setup and organize, if not for the portable setup.
57 greatly reduces the size of the system due to the need for fewer and simpler components. Figure 3 5 Portable experimental setup, using devices made of Rogers 3 003C semi ceramic dielectric The system is operated at the frequency band at which the transformer resonates with the electrode. This frequency band can be shifted depending on the physical characteristics of the transformer, which is set by the user. However, t he intense heat produced at this higher operational frequency was too much for the FR4 to handle, which is why this setup was operated with the SC devices only 3 .4 Description of Biological Pathogens Tested A wide var iety of microorganisms were tested for sterilization. These are summarized in the Table 3 1 given below. Owing to the BSL I nature of the testing facility, most of the parametric studies as well as other diagnostic tests were conducted using the BSL I orga nisms listed in Table 3 1. The sterilization tests with the BSL II
58 pathogens were conducted to demonstrate the sterilization efficiency of the DBD plasma. Table 3 1 Listing of all the microorganisms tested MICROORGANISM STRAIN TYPE TYPE OF PATHOGEN BSL G RAM / G RAM + Saccharomyces cerevisiae N/A Fung us I G+ Escherichia coli (non pathogenic) C600 Bacteri um I G Mycobacterium smegmatis(non pathogenic) ATCC 19420 Bacteri um I G + Pseudomonas aeruginosa 6003 7 Bacteri um II G Yersinia enterocolitica SSUD 4037 Bacteri um II G Salmonella.enterica EPI 6031 Bacteri um II G Listeria monocytogenes EPI 1132 Bacteri um II G+ Vancomycin Resistant Enterococc us (VRE) faecium VRE 82 Bacteri um II G+ Escherichia coli(pathogenic) EPI 562 Bacteri um II G Vibrio chol era N16961 Bacteri um II G Acinetobacter baumannii MD112 Bacteri um II G Methicillin Resistant Staphylococcus Aureus (MRSA) WCH132 Bacteri um II G+ Geobacillus stearothermophilus N D Bacterial Spore I G+ Bacillus subtilis NCIB3610 Bacterial Spore I G+ Most bacteria are either gram positive (G+) or gram negative (G ) and stain purple or red respectively when subjected to a Gram stain test The differences in staining are due to fundamental differences in the structure of their cell walls, as shown in Fig ure 3 6 b elow. Figure 3 6. The bacterial cell wall (a) The Gram positive envelope (b) The Gram negative envelope
59 In G+ bacteria, the lipidic plasma membrane with embedded proteins is covered by a multi layered peptidoglycan shell decorated with polysacc harides, teichoic acids and proteins. In G bacteria, a thin peptidoglycan layer surrounds the plasma membrane and is covered by an asymmetrical outer membrane containing lipopolysaccharides, which lies on the peptidoglycan layer. Thus G bacteria are more easily killed by species that damage membranes. [ 87 ] Yeasts which stain G+, are eukaryotic organisms with a polysaccharide cell wall consisting of a moderately branched 1,3 glucan backbone cross linked with 1,6 glucan, chitin and proteins. The structure of the yeast cell wall is sho wn below in Figure 3 7. Figure 3 7. Structure of the yeast cell wall. The wall is primarily composed of mannoproteins and glucan that is linked (1 >3) and (1 >6). Ergosterol is the major lipid component of the underlying plasma structure. [ 88 ] In order to understand the interaction of plasma with the structure of a microorganism, it is necessary to understand the structure of the micro organism i.e. the primary cell structures protecting a micro organism (cell envelope), the proteins essential to its survival and the organelles needed for it to breathe, grow and reproduce.
60 Given below is a brief description of the cell structure for each of the types of pathogens discussed. Table 3 2 D es cription of organelles and essential proteins in different types of microorganisms Type of Pathogen Cell Envelope Important cell organelles Necessary proteins/compounds Yeast ( Fungi) Cell wall, Periplasm, Plasma Membrane Mitochondria (respiration), Nucleu s (DNA replication and repair), Golgi apparatus and vacuoles (protein breakdown) Proteins, Glycoproteins, Polysaccharides, Polyphosphates, lipids, nucleic acids G+ bacteria Cytoplasmic lipid membrane, thick peptidoglycan layer Capsule polysacchari des, flagella (only in some species), ribosomes, nucleus Teichoic acid, peptidoglycan, polysaccharides, lipoproteins G bacteria Outer membrane containing lipopolysaccharide, Cytoplasmic membrane, thin peptidoglycan layer, Flagella (only in some spe cies), ribosomes, nucleus Peptidoglycan, polysaccharides, lipoproteins. Spores Encased in a protein rich coat, sometime surrounded by an exosporium Depends on the host organism, forming the endospore. Usually consists of the DNA and a portion of the cytoplasm of the host. 3 .5 Experimental Protocols Followed Before any sterilization experiment, both the bench and the metal connector are swabbed with alcohol to disinfect the experimental setup. Additionally, before each experiment, the optical dens ity (OD) of the microbial sample is measured using a spectrophotometer. The optical density is an important indicator of the viability of a bacterial sample.
61 Figure 3 8 Typical growth curve for a bacterial population in a batch culture [ 89 ] Figure 3 8 above gives a plot of the OD versus time, during the lifetime of a bacterial cell concentratio n. As is seen in Figure 3 8 any bacterial culture starts off with a lag phase (wherein a few bacterial colonies have grown), an exponential or logarithmic phase (wherein growth is linear and cells are actively reproducing ), a stationary phase (when maximu m growth has been reached ing effect is seen) and a death phase wherein cells begin breaking down. With the test organisms used in this study, an OD of 0.5 1 corresponds to the late logarithmic phase (which is the actively growing phase) an d a high concentration of cells (~106 8 CFU /ml ) in the culture. 3 .5.1 Preparation of Bacterial Biological T est S amples for E xperiments: Saccharomyces cerevisiae ( b aker's yeast) was obtained from the grocery store, mixed in hot water and held at room tempe rature for 1 hour. A sterile loop wa s dipped into this solution and streaked onto a Sabouroud's (SAB) agar plate, which wa s then incubated at 30C overnight. A single colony was re streaked on SAB agar and
62 incubated at 30 overnight Mycobacterium smegmati s ATCC 19420 was grown on Middlebrook 7H10 agar plates or in Middlebrook 7H9 liquid medium at 37C. G stearothermophilus was grown on trypticase soy agar or broth at 50C. E coli C600 was grown on Luria Bertani (LB) agar or broth at 37 o C. All cultures w ere frozen at 80 o C in the appropriate broth with 25% glycerol and inoculated onto fresh plates before use. For each sterilization experiment, one to three colonies were inoculated into the appropriate broth and incubated at the appropriate temperature, with shaking, until the optical density (OD) of the microbial sample was between 0.5 1. Samples of B. subtilis were grown via two different methods (a) In LB medium (commonly used to culture E. coli and other related species) (b) In minimal salts glutamate glycerol (MSgg) medium (a bio film promoting medium). A sterile inoculating loop was swabbed with frozen cultures of B. subtilis an d streaked onto an LB plate. This plate is incubated at 37 o C for 12 hours. Another inoculating loop was then used to pick up a single colony from the incubated plate. This loop wa s then swirled in a glass test tube filled with 3 ml of LB broth. This test tube wa s then vortexe d in a shaker, maintained at 37 o C for 3 hours. After 3 hours, the optical density (OD) of the sample wa s checked to ensure viability. This ma de up the LB broth culture. Subsequently, 0.3 ml of this culture wa s mixed with 2.7 ml of MSGG broth and vortexed for an additional half hour. After half an hour, this sample wa s also checked for viability by checking OD. For each plasma sterilization experiment, the plasma device wa s inoculated with the requisite volume of bacterial sample. This requisite volume is known as the inoculation volume
63 ment ioned. This bacterial sample is then spread uniformly over the entire electrode surface area, using a sterile inoculating loop. 3 .5.2 Post Processing of Bacterial Samples a fter Experiments Once the experiment is completed, this device is taken and deposi ted in a sterile bag filled with 5 ml of culture broth (relevant to the bacterial sample being tested). The bag is sealed and agitated thoroughly using a Fisher Scientific Mini Vortexer Lab Mixer to wash off any micro organisms clinging to the device. 0.1 ml of this broth is pipetted out into a clean dilution blank filled with 0.9 ml of Phosphate Buffered Saline (PBS) Solution and the dilution blank is vortexed for 10s. This dilutes the number of colony forming units (CFU)/ml in the bag by a tenth. Th i s process is repeated, using a new dilution blank each time, until the fourth dilution is reached. Thus, a dilution series is made for each device used in the experiment. 0.1 ml from each dilution blank in the dilution series is then pipetted out onto a f resh agar plate (relevant to the type of inoculating pathogen) and spread uniformly. These plates are then incubated at the required temperature for 24 48 hours. C olony count methods are used to estimate the number of CFU /ml on the 0 th dilution. This proc ess is repeated for each device being tested in the experiment. A survival curve (plot of logarithm of number of CFU (colony forming units)/ml versus testing parameter ) is then plotted. Testing parameter can be sterilization time, input voltage etc. Exp eriments are performed in triplicate (unless otherwise mentioned) to ensure repeatability 3 .5.3 Ozone Safety Protocol 2 The DBD plasma devices also produce ozone as a byproduct. This concentration of ozone is greatest at the locations nearest to the devic e. To ensure that laboratory personnel are not exposed to unsafe levels of ozone, an ozone monitor is used to 2 The author thanks Poulomi Banerjee for her rigorous work in setting up an ozone safety protocol
64 determine safe operating conditions. O ccupational Health and Safety Hazard (OSHA) standards regulate employee exposure to ozone gas through its A ir Contaminants Standard, 29 CFR 1910.1000. The permissible exposure limit (PEL) is listed as an 8 hour, time weighted average value of 0.1 part of ozone per million parts of air (ppm) and the short term exposure limit (15 minutes) is 0.3 ppm All experim ents must be done under conditions that stay below these exposure limits. This protocol is an internal safety document recommended for the plasma generation device with chamber door open or closed. Based on the test results it was established that for th is particular set up at any instant ozone is within allowable levels according to OSHA regulations. Step 1 : Switch on the 2B Tech 202 Ozone m onitor (1ppb resolution) 10 15 minutes before the experiment. It should read room ozone concentration as ~ 0.02pp m Step 2 : Connect circuit according to Figure 3 1 Double check all connections Make sure the L ab VIEW interface is able to read the ozone meter. Step 3 : Set up the experiment. Usually, e xperiments are conducted within an acrylic chamber, which is alwa ys kept closed while the device is running. Additionally, as a precaution, the chamber is kept closed for another 10 minutes after switching off the device. Cover the device with a pre designed charcoal mesh (to be explained, in Chapter 4) with particle si ze between 1.4 mm to 4.75 mm. The charcoal adsorbs the ozone Step 4 : User should always be in the safe zone from the device, which is at least 36 inches away from the outer acrylic wall of the chamber. It was observed that ozone
65 does not exceed maximum all owable limit at this distance even when the chamber door is open. Step 5 : During an experiment, do not keep the device running for more than 20 minutes at a time. No more than 15 experiments should be run in one day The charcoal mesh should not be remove d at any point of the experiment. Step 6 : When removing the concerned device from the acrylic chamber (for further post processing), check the ozone levels inside the chamber. A 3M 8514 respirator mask is available for ozone protection up to 10 times OSHA PELs (Permissible exposure limits) and may be worn while shutting down the device or for all subsequent protocols. If there is excess ozone at any stage of the experiment, power down the setup immediately. Open the doors and windows of the room. Step awa y from the device. Thus far, Chapter 3 described the experimental setup used, the diagnostic equipment used to measure different plasma parameters, preparation protocols for the different types of micro organisms tested and the experimental protocols emp loyed before and after all plasma sterilization experiments completed during the course of this study. Chapter 4 and 5 describe the bulk of the research completed in understanding DBD plasma sterilization.
66 CHAPTER 4 PARAMETRIC STUDIES IN DBD SURFACE PLA SMA STERILIZATION In trying to understand DBD surface plasma sterilization, my research followed two paths (1) The bulk of research in plasma sterilization uses mostly volume plasma configurations. The DBD surface plasma setup used in this paper required a different set of experimental protocols to be developed, in order to facilitate the testing of microorganisms exposed to plasma. Before understanding the mechanism of plasma sterilization, it was necessary to conduct a parametric study in order to unders tand the sterilization capabilities of such a setup. The variation of the different input parameters involved in plasma generation help understand what enhances DBD surface plasma sterilization (2) Studying the mechanism of surface plasma sterilization inv olves understanding how each component of plasma (UV photons, reactive chemical species and temperature) affects the process of plasma sterilization. This has been further explained in Chapter 5. This chapter describes the parametric studies conducted in understanding DBD surface plasma sterilization. The plasma source used in this study is an AC, RF plasma operating at an input frequency of 14 kHz and an input voltage of 12 kV pp (unless otherwise mentioned) The different parameters tested were 1) Type o f Pathogen 2) Inoculation Volume 3) Nature of dielectric substrate 4 ) Input power/frequency 5 ) Operating Pressure 4.1 Type of Microorganism The protocol for each of the sterilization experiments is the same: take a clean plasma device, inoculate it with th e requisite volume of microorganism sample (this volume is defined as inoculation volume), power the device for a fixed time interval (
67 carefully from the electrical bench and then subject it to the post processing protocol described in Section 3.5.2 Plates recovered from each device tested are incubated as previously described and colony counts are obtained in order to recover the number of microbial survivors (N) in each case ( CFU ). A survival curve is the plot of log10N versus the plasma exposure time ( In this study, survival curves we re obtained using S. cerevisiae (Yeast), E coli, B. subtilis G. stearothermophilus spores and a wide range of BSL II pathogens further described below. In the case of plasma sterilization, the norm is to triplicate each experiment in order to ensure repeatability. Hence each sterilization experiment described in this study has been triplicated whenever possible However it is prudent to analyze the source of error in these experiments. Error analysis for the purpos e of experiments discussed herein can be classified into two types : Analysis of the error associated with the microbiological technique used during post processing : The post processing protocol that each device is subject to after plasma exposure is expla ined in detail in Section 3.5.2 Most of the error analysis methods described below are obtained from Niemela et al. [ 90 ] Assuming that there is no significant change in the volume of bacterial sample deposited on the device, due to the inoculating loop, the other experimental uncertainties introduced are given below:
68 (a) Variation of particle numbers due to uncertainties in counting : This is expressed by a term known as Poisson scatter ( of colonies observed). (b) Uncertainty of the test portion volume (w v ): This itself is the result of three main influences: a) repeatability of filling and emptying the measuring device (pipette) b) specification of glassware manufacturer c) temperature effect when calibration and measurement takes place at different temperatures. For our pu rpose, (a) and (c) are assumed to remain insignificant (i.e. no systematic errors). (c) Uncertainty of the dilution factor : The uncertainty variance of a dilution step is obtained from the below formula: (4 1) w here a= suspension transfer volume (0.1 ml), b= dilution blank volume (0.9 ml), u a = l, from manufacturer), u b (1.5 l, from man ufacturer), w a If the total 2 of the dilution factor = For instance, if a sample has been diluted to 10 4 k= 4. Hence total uncerta inty of the result = w y = where and are already defined and is the uncertainty associated with the 0.1 ml that is transferred from the dilution blank to the agar plate (Since the same pipette is u sed for transferring w v = w a ) Each sterilization experiment typically has 4 5 data points at which to test sterilization. For each of these points, a different plasma device is used. Although each
69 device is manufactured using the exact same spec i fications, there is always a n element of uncertainty involved. Moreover, once the inoculated device is placed on the electrical bench and the device powered, there is a short lag time, during which the input voltage is manually increased to 12 k V pp This also introduces a factor of uncertainty. Since DBD plasma sterilization occurs which in turn affects the total uncertainty calculated. Thus, for the purpose of the experiments discussed herein, experimental error is difficult to estimate via this method Analysis of the variation of plate counts obtained in each sterilization experiment: Most commonly, the Analysis of Variance (ANOVA) is used to compare groups of measurement data. For ins tance, if a group of five random students in four different laboratories were to conduct the same experiment, the ANOVA could be used to compare the variance of the mean of the observations within each group of students as well as the average variance with in each laboratory. In a one way ANOVA there is one measurement variable (values that are measured or recorded) and one nominal variable (categorical variables which are not measured, but rather defined at the beginning of the experiment). However for ea ch sterilization experiment, there are usually two nominal variables (the type of device used and the sterilization parameter being tested). Experimental protocols designed for this setup do not permit the usage of the same device for each data point in a single sterilization experiment. The application of the one way ANOVA assumes homoscedasticity (data obtained in all three trials have the same standard deviation), which does not hold true for the plasma sterilization experiments. For instance, a plasma e xposure of 60s, using a SC device, might lead to a 4 log10 reduction in one
70 trial, but a 3 log10 reduction in another and a single log10 reduction in the 3rd trial. Hence it is unsuitable to use ANOVA for the sterilization tests described in this study T hus, for the purpose of this study sterilization data in different cases is presented in one of two ways. For experiments with three trials or more, the mean of data, over all the trials, at each data point has been calculated and plotted on a log10 scale (For instance, Figure 4 1 below). The variation in plate counts between trials at each time point is shown in terms of the standard deviation. However, for some other experiments with two trials or more, data for all the trials has been presented in one plot, simply because the variation in data is better presented this way (For instance, Figure 4 2, 4 3 ). BSL I Microorganisms Microorganisms tested in this case we re S. cerevisiae, E. coli C600 and M. smegmatis (Table 3 1). Even though it is a fung us S. cerevisiae stains similar to gram positive ( G+ ) bacteria. The latter two are classified as gram negative ( G ) and gram positive ( G+ ) bacteria respectively. A detailed description of G+ and G bacteria is given in Chapter 3. Figure 4 1 Survival curves obtained using FR4 plasma devices and S. cerevisiae (Yeast) and E. coli as test micro organisms Complete sterilization is obtained within 90 120s.
71 All BSL I microorganisms have been tested on both FR4 and SC devices. Survival curves obtained using SC de vices are described in Section 4.3. Figure 4 1 above shows the survival curves for two test pathogens, S. cerevisiae and E. coli respectively. Plasma treatment times of t= 30,60,90,120s were tested using FR4 devices. The sterilization plots in Figure 4 1 have been plotted by obtaining the average of log 10 (N) over a number of trials (usually 3) versus sterilization time (t). As explained before the error bars, plotted as the standard deviation of this data, do not indicate error associated with the exper iments, but rather the variation observed in the number of survivors at that particular time point. Figure 4 1 demonstrates that, using FR4 devices, complete inactivation of Yeast and E. coli is obtained in 90s and 120s respectively Complete inactivation (or sterilization) implies the reduction of a pathogen concentration from N o CFU (at t= 0s) to zero CFU For yeast, an additional sterilization time point of 180s was also tested, wherein also complete inactivation i s observed From Figure 4 1, it is also evident that a greater variation in plate counts is seen in the case of yeast as compared to the case of E. coli This variation is especially noted at t=60s and 90s. Due to the noted variation, sterilization experiments using yeast were repeated over 6 t rials instead of the standard 3. During these trials, it was observed that in 4 out of 6 trials, complete inactivation of yeast occurred after t= 60 and 90s. The incomplete inactivation in the remaining 2 trials at t= 60s and 90s is what contributes to th e large variation ob served for yeast in Figure 4 1. Ying et al. [ 55 ] use d a 10 kV, 6.5 kHz volume plasma configuration to report a 5 log 10 reduction in yeast concentration after 5 minutes of plasma exposure time. 100%
72 reduction was not obtained within this time. Sohbatzadeh et al. [ 61 ] report ed a 100% reduction in E coli concentration after exposing bacterial samples for 15 minutes to a 50 Hz, 5.4 kV DBD plasma (volume discharge configuration). On the other hand, Lee et al. [ 45 ] report ed complete sterilization of E. coli on expos ure to a 2.45 GHz, microwave plasma. Compared with these results, our reported sterilization time is considerably reduced O nly a single sterilization test for M. smegmatis was done, using both FR4 and SC devices T he data is not reported here. Incomplete inactivation was noted for both FR4 and SC devices in the case of M. smegmatis Starting from an intial bacterial concentration of 10 6 CFU a plasma exposure time of 2 minutes resulted only in a 2 3 log 10 reduction. B. subtilis cells were also exposed to plasma at the same input parameters of 14 kHZ frequency and 12 kV pp B. subtilis is a Gram positive, rod shaped spore forming bacterium, commonly present in soil and the human gut. They have become widely adopted as a model organism for laboratory studie s. As described in Section 3.5.1. the B. subtilis samples were prepared in two different media: LB and MSgg media. For the sake of discussion, B. subtilis cells grown in LB and MSgg medium will be referred to as B. subtilis I and B. subtilis II respective ly While LB medium enables the B. subtilis cells to express the wild type single cell phenotype, MSgg medium enables them to express a phenotype that promotes biofilm growth [ 91 ] A biofilm is an assemblage of one or more bacterial species that forms on a surface. The bacteria are embedded in an exo polysachharide matrix. Biofilms tends to be m ore resistant to biocides than the wild type. Figure 4 2 shows the survival curves for both types of B. subtilis cells. While B.
73 subtilis II were completely inactivated within 4 minutes, it is interesting to note that B. subtilis I concentration platea u s at about 10 4 CFU after 6 minutes of plasma treatment. We do not know the number of spores formed under these two growth conditions and the plateau may represent spore survival. Figure 4 2 Survival curves obtained using FR4 devices for B. subtilis cell s grown in LB and MSgg medium Sterilization tests with B. subtilis were conducted in duplicate. Hence data for both trials are shown in Figure 4 2 above. C omplete inactivation is obtained in 4 minutes in the case of B. subtilis II while even after 6 minut es, incomplete inactivation is observed in the case of B. subtilis I. As already stated, B. subtilis I and B. subtilis II differ in the ty pe of phenotypes expressed. Simply put, the cells in B. subtilis I tend to have a higher degree of locomotion [ 92 ] while the cells in B. subtilis II are more likely to form exo polysaccharide matrices [ 93 94 ] Thus Figure 4 2 seems to indicate that matrix forming B. subtilis cells are more susceptible to plasma than motile B. subtilis cells. The percentage of spores formed under each condition is not known at this time. Akishev et al. [ 83 ] report ed similar experiments wherein they use a plasma jet (operating at a power of 60 W) to inactivat e both vegetative cells and spores of B.
74 subtilis Their experiments reported incomplete inactivation of both types (~4 log 10 reduction) in CFU even after 10 minutes of plasma exposure. However they reported complete inactivation in one particular type of vegetative cells ( B. subtilis ) cultured on a less nourishing medium. Hence they conclude d that the type of medium in which cells are cultured also affects their inactivation time. This observed susceptibility of a particular cell type to plasma will be d iscussed in greater detail in Chapter 5. The last kind of BSL I microorganism to be tested was a purified suspension of G eobacillus stearothermophilus spores. G. stearothermophilus spores are rod shaped, Gram positive bacteria, widely distributed in soil and are usually a cause of spoilage in food products [ 95 ] G. stearothermophilus spores are commonly used as biological indicators for periodic checks of sterilization cycles as they are highly heat resistant G. stearothermophi lus purified spores were used for sterilization tests to examine the resistance of spores themselves without any contaminating vegetative cells. Figure 4 3. Survival curves, using FR4 devices for G stearothermophilus spores Figure 4 3 above shows the survival curves for G. stearothermophilus spores
75 sterilization te sts were conducted in duplicate. D ata for both trials is shown in Figure 4 3. C omplete inactivation of G. stearothermophilus spores is obtained within 2 0 minutes Figure 4 3 shows a triphasic behavior, with an initial drop in spore concentration followed by a lag phase and finally, a rapid tail phase. One suggested mechanism of inactivation due to plasma treatment is the generation of hydroxyl radicals t hrough absorption of reactive oxygen species (ROS) such as ozone during plasma generation and the subsequent reaction of these absorbed ROS with liquid inside the cells to form hydroxyl radicals through a chain of reactions called the Fenton Mechanism [ 63 ] However G. stearothermophilus being a spore producing bacterial species forms resistant spore s when treated with plasma. These spores are typically devoid of any kind of liquid, which in turn, disables one of the main mechanisms of cell damage due to plasma treatment. This may be one reason as to why G. stearothermophilus is so resistant to plasma treatment. In order to compare sterilization efficiency, the D value is often used as a reduction of 90% in the CFU i.e. the time taken for a single log 10 reduction The D value can be calculated using the formula [ 71 ] (4 2) From Figure 4 1, a phasic behavio r is noted (bi phasic for E. coli and tri phasic for y east). Typically a D value can be calculated for each of these phases.However owing to the experimental protocols designed for each of these tests, select time points (30s 120s) can only be tested. From survival curves shown in Figure 4 1 a slow inactivation phase followed by a steep drop in pathogen concentration is observed.
76 Since experimental protocols allow only a limited number of time points, D value for each phase cannot be calculated. However, the D value for the initial linear portion of the survival curve has been calculated and shown below in Figure 4 4. Figure 4 4 Comparison of D values for the different test microorganisms using FR4 devices In Figure 4 4 above, D value is calculated as an average of the D values from individual sterilization trials for each test organisms. Thus, for E. coli y east and G. stearothermophilus, the D values are calculated as 50 s 35 s and 15 5s (2.5 minutes) respectively. The larger variation in D value in the case of yeast is mirrored in Figure 4 1. For the purpose of sterilization experiments, yeast seemed a little less dependable as a test organism Hence for all successive sterilization tests, E. coli was used as the test organism BSL II Pathogens: St erilization experiments were also carried out with a host of BSL II pathogens, using the portable experimental setup (Section 3.3) As previously described, due to the limitations imposed by the experimental setup, only SC plasma devices could be used for sterilization tests with this setup. Plasma generation
77 parameters used were an input voltage and frequency of 1 0 kV pp and 47 kHz respectively. Table 4 1 below shows the results of these tests. Table 4 1. Results obtained from Plasma Sterilization Experi ments with BSL II pathogens Type of pathogen Sterilization time (min) Observed reduction in bacterial concentration Complete inactivation G ve or G+ve P.aeruginosa 6003 7 2 8 log 10 YES G Y.enterocolitica SSUD 4037 2 8 log 10 YES G S.enterica EPI 6031 3 7 log 10 YES G Listeria monocytogenes 3 8 log 10 YES G+ Vancomycin Resistant Enterococci (VRE) 3 8 log 10 YES G+ Escherichia coli 3 8 log 10 YES G Vibrio cholera 3 8 log 10 YES G Acinetobacter baumannii 3 4 log 10 NO G MRSA WCH132 2 3 log 10 NO G+ While a detailed survival curve was plotted for BSL I micro organisms using equally spaced plasma exposure time intervals (30s, 60s, 90s and 120s) the same was not done for BSL II pathogens. Owing to the caution needed in a BSL II environment and the availability of a single portable experimental setup during this time, only a single sterilization time point was tested for most of the BSL II pathogens. However, a complete survival curve was plotted for MRSA WCH132 ( initial concentration= 10 8 CFU ) wherein different inoculated devices were exposed to plasma for 30s, 60s, 90s, 120s. Only a 3 log 10 reduction was observed in 2 minutes. Even when the plasma treatment t ime was extended to 3 minutes, only a 4 log 10 reduction in bacterial concentration was observed, but complete inactivation proved elusive suggesting that this strain of MRSA is more plasma resistant than other vegetative cells.
78 The main purpose of the BSL II tests was to characterize the effectiveness of the generated DBD plasma in steril izing BSL II pathogens. In doing so, it is obvious that the DBD plasma used in this study is capable of completely inactivating a wide variety of resistant pathogens within 2 3 minutes. Table 4 1 also points to the fact that inactivation due to the generat ed plasma exposure seems to be independent of whether the pathogen is gram negative (G ve) or gram positive (G+ve), since similar sterilization times have been noted for both types of pathogens. Thus Section 4.1 provided a summary of the sterilization effe ctiveness of DBD surface plasma against a wide range of vegetative and spore producing micro organisms Another factor that also determines sterilization effectiveness is the inoculation volume used for sterilization tests i.e. the amount of pathogen sampl e deposited on the surface of the plasma device. This is further discussed in Section 4.2. 4.2 Inoculation Volume For the purpose of this study, t he volume of pathogen sample deposited on the electrode surface of the plasma device is defined as the inocul ation volume This sample is then spread uniformly over the entire surface area of the electrode, using a sterile inoculation loop. This experimental parameter is important, owing to the relationship between pathogen density and sterilization time and due to the presence of water, proteins and salts in the inoculum The purpose of testing a higher inoculation volume was to understand the dependence of sterilization time on the volume of liquid bacterial sample deposited on the dielectric surface i.e. whethe r a higher volume of liquid sample led to a longer sterilization time. To that end, survival curves were obtained using a higher inoculation volume (40
79 Figure 4 5 Survival curves for inoculation volume= 40 l of E. coli Figure 4 5 above shows the survival curve for all three trials, using 40 l of E. coli The sample of E. coli used for these experiments had an OD in the range of 0.5 1, which co rrelates to 10 8 CFU /ml. Hence 20 l and 40 l of this sample should correlate to 2x10 6 CFU and 4x10 6 CFU respectively. However, this difference in CFU is not significant enough i.e. essentially, the same number of E. coli cells are being deposited in diffe rent sample volumes. As is evident, while 20 l of E. coli requires a complete sterilization time of 90 120s, 40 l requires a complete sterilization time of 150 180s. Furthermore, was noted in Figure 4 1 is extended by about 30s here i.e. the rapid drop in E. coli concentration occurs after t> 90 s, as opposed to after 6 0s in the case of the lower inoculation volume (20 l). Thus a higher inoculation volumes leads to a longer steril ization time. However this extension in sterilization time seems to be more dependent on the volume of sample deposited than the number of CFU in the sample.
80 Thus far, in the sterilization tests discussed in Section 4.1 and 4.2, only FR4 devices have been used. However a different dielectric material was also used for sterilization tests, the results of which are discussed in Section 4.3. 4. 3 Nature of Dielectric Material The nature of the dielectric material/substrate is a very significant factor in plas ma sterilization. The ability of the dielectric material to withstand larger number of plasma cycles determines the lifetime of a plasma sterilization device. This ability could be dependent on sterilization time, type of organic residue/material usually r emnant on the substrate material and input plasma power density. Kelly Wintenberg et al. [ 19 ] comm ented on the nature of the substrate material value. They used a volume DBD plasma setup, using E. coli as the test organism and conducted sterilization tests using polypropylene, glass and agar as possible substrates. They fou nd that sterilization on polypropylene surface took d that the cells drying out on glass slides require a larger concentration of active species for effective sterilization, as opposed to c ells which do not penetrate the fibers of poly propylene surfaces and hence require lesser concentration of active species. Lerouge et al. [ 14 ] specula te d on the relationship between substrate material and dielectric heating of the substrate material according to relationship f 2 (4 3 ) f is Different pathogens cling to different surfaces differently, depending on degree of adhesion.
81 Substrate interference with the process: This can happen, if during the course of plasma exposure, substrate itself gets pitted or etched, thus increasing its adsorption capacity and hence enabling it to adsorb a larger concentration of chemica l species available, thus hampering sterilization. the ionization process. Sterilization experiments with FR4 devices have been discussed in great detail in Section 4 .1 and 4.2. Figure 4 6 below compares sterilization effectiveness between FR4 and SC devices using (A) y east and (B) E. coli as test organisms The testing protocol with SC devices is the same as with FR4 devices. Inoculation volume used was As is e vident from Figure 4 6, the time taken for complete sterilization in the case of SC devices is 180s and 120s in the case of Yeast and E. coli respectively Even for complete bacterial inactivation, FR4 devices take ~30s lesser time for complete sterilizati on, as compared to SC devices. The D value (time required to sterilize 90% of the bacterial concentration) for the FR4 dielectric is lesser than the SC dielectric, both for E. coli and Yeast. This is shown in Figure 4 7 below. rs to be for FR4 in the case of yeast. For E. coli the D value in the case of FR4 is ~20s shorter than in the case of SC. The D value in the case of yeast, for SC, is the highest and shows the highest standard deviation. The high standard deviation in thi s case implies that the sterilization of yeast using the SC dielectric was highly unreliable. Scanning Electron Microscopy ( SEM ) and Energy Dispersive Spectroscopy ( EDS ) analysis was also used to observe the modification of the dielectric surface of the p lasma devices after prolonged plasma exposure, both in clean and inoculated cases.
82 Figure 4 6 Survival curves comparing FR4 and SC plasma devices for A) S. cerevisiae and B) E. coli Figure 4 7 Comparison of D values for the different (dielectric, test pathogen)
83 SEM imaging and EDS analysis were done in the Major Analytical Instrumentation Center (MAIC ) 3 at the University of Florida SEM imaging of a device consists of scanning an electron beam across the surface of a sample, line by line, much li ke reading a book. This is a called a raster pattern. At each location where the electron beam strikes the sample, an electron signal is used to produce contrast in the image displayed on a cathode ray tube (CRT) viewing screen. In SEM terminology, magnifi cation is the difference between the size of the scanned area on the sample surface and the size of the display showing the resulting image. Simply put, SEM magnification is akin to imaging a small 1 mm x1 mm square on a surface and displaying it in a 100 mm x 100 mm square, which corresponds to a 100x magnification. This is fundamental in understanding, simply because the magnification in the case of SEM is significantly different from light microscopy. For the purposes of the SEM studies discussed herein, a JEOL SEM 6400scanning electron microscope, with an accelerating voltage of 15 kV and a magnification of 10x 300000x was used. The SEM images shown herein were obtained at a magnification of 2000x. The scale bar, at the top left corner of the image, prov ides an idea of the length scale of structures in the image. In addition to analyzing the appearance of the dielectric substrate, an idea of the elemental makeup of the individual dielectric substrate was also obtained using EDS analysis. The elemental m akeup can determine the difference between the FR4 and SC dielectric and detect the deposition of additional salts or oxides during plasma sterilization. Owing to the nature of the EDS setup, a qualitative, rather than quantitative 3 For the SEM imaging, Thanks are due to the folks over at Major Analytical Instrumentation Center (MAIC), UF especially Wayne.A.Acree and Dr.Mike Kesler.
84 idea of the elemental ma keup of the surface can be achieved i.e. for a given dielectric surface, its elemental makeup can be determined, but not the percentage of each element. When an electron beam hits the surface, atoms in the dielectric surface interact with this beam and und ergo energy shell transitions, resulting in the emission of an X ray. Th is emitted X ray has an energy characteristic of its parent element, which can then be identified This sums up the process of Energy Dispersive X ray spectroscopy. The primary goal o f the SEM and EDS studies was to test the effect of prolonged plasma generation on the dielectric substrate. Fresh FR4 and SC devices were taken and subjected to one of the protocols described below. Once the protocol for each device was repeated for the r equisite number of cycles, they were then prepped for SEM imaging and EDS analysis. Table 4 2. Description of protocols that the different devices were subjected to, prior to SEM Testing. Protocol# Description of Protocol No. of cycles 1 Clean device, n o plasma, serves as control 0 2 Clean device, powered continuously for 20 minutes 1 3 E. coli and powered for 2 minutes. 5 (=10 min of plasma) 4 E. coli and powered for 2 minutes. 10 (=20 min of plasm a) 5 E. coli and powered for 2 minutes. 25 (=50 min of plasma) It should be noted that for protocol#2, a continuous plasma run time of 20 minutes was used to mimic protocol#4, minus the constant inoculation with E. coli The
85 goal of doing so was to compare the effect of prolonged plasma exposure on a clean device versus an inoculated device. Figure 4 8 (A E) below depicts SEM images taken of the dielectric surface of the different FR4 devices (corresponding respectively to pr otocols 1 5) at 2000x magnification. Figure 4 8 (F J) depicts images of the electrodes of the same devices at 2000x magnification. Figure 4 9 represents similar images for the SC devices. The white scale bar for all images, at the top left corner, is 20 Figure 4 8 SEM images of FR4 devices at 2000x magnification. Images A E correspond to the SEM images of the dielectric surface of the devices used for protocols #1 #5 respectively. Images F J correspond to the SEM images of the electrode surface o f the same devices. The white scale bar at the top Comparing Figures 4 8 (A) and (B) there does not appear to be much of a modification of the dielectric substrate due to plasma generation for 2 0 minutes in (B), as compared to the di electric substrate of a new device (A). Comparing 4 8 (C) (E), there seems to be a deposition of a grainy material on the dielectric surface. The grainy material can be either the accumulation of salts or molecules from cell debris from
86 constant deposition of the E. coli sample or the accumulation of electrode sputter due to repeated plasma generation. Figure 4 9 SEM images of SC devices at 2000x magnification. Images A E correspond to the SEM images of the dielectric surface of the devices used for pro tocols #1 #5 respectively. Images F J correspond to the SEM images of the electrode surface of the same devices. The white scale bar at the top left Comparing Figures 4 9 (A) (E) above this same behavior is not noticed for SC devices. Ho wever, SEM imaging focuses on analyzing a single cross section of the surface being examined. Hence the absence of the grainy material in Figures 4 9 (C) (E) could also be because the particular cross section being analyzed is devoid of the grainy material. The important point to conclude from Figure 4 8 and Figure 4 9 (A) (E) is that the different protocols do not seem to affect the dielectric surface, except for the deposition of a grainy material, in some of the cases. Similarly, comparing electrode sur faces for FR4 devices (4 8 (F J)) and SC devices (4 9 (F J)), it is observed that electrode surfaces (F G) appear to remain the same, while electrode surfaces (H J) appear segmented or sputtered (in the case of
87 FR4 and SC). Since this segmented/sputtered a ppearance is not noticed in the case of the FR4/SC device powered solely for 20 minutes, without E. coli deposition, it is likely that the constant deposition of the E. coli culture leads to electrode corrosion. EDS studies were used to identify the eleme ntal composition of the dielectric surface of the various plasma devices, after they were subjected to the different protocols. In the EDS analysis shown below for a FR4 device (Figure 4 10 ) and SC device (Figure 4 11 ), the absence of a way of determining relative ratios of elements makes it difficult to derive any information about the elemental variation among the different electrode surfaces and the different dielectric surfaces. Figure 4 10 (A) displays a high concentration of bromine (Br) and trace con centrations of tin (Sn). Bromine is often used to enhance flame resistant properties in in FR4 laminates, which is why it is detected in the higher concentrations on the EDS plot. Figure 4 10 EDS analysis of the A) dielectric surface B) electrode surf ace for a FR4 device The FR4 dielectric, used for making the devices, is copper clad with a copper (Cu) layer overlaid with a tin (Sn) finish. It is possible that milling of this Cu+Sn layer
88 during device manufacture led to Sn residues on the surface, whic h explains the trace concentrations of Sn on the EDS plot too. Figure 4 10 (B) on the other hand displays a high concentration of Sn (due to the electrode surface) and Chlorine (Cl). Figure 4 11 (A) shows a high concentration of Silicon (Si), which is li kely a component of the SC dielectric surface. Figure 4 11 (B) shows a high concentration of copper (Cu), which is expected since the electrode is made up of copper. For the sake of simplicity, EDS plots for the other protocols have not been shown. Howeve r, from these plots, additional information such as the variation of elemental ratios from device to device or the detection of additional salts deposited on the devices due to continuous plasma generation is not derived. Figure 4 11 EDS analysis of t he A) dielectric surface B) electrode surface for a SC device Thus SEM and EDS analysis of a variety of FR4 and SC devices subjected to a number of protocols (Table 4 2) do not highlight clearly the substrate modifications taking place due to the different protocols. While some information is provided in terms
89 of elemental signatures and electrode appearance, it seems that the different protocols do not produce any consistent dielectric substrate modification both in FR4 and SC. This is very puzzling since it has been observed that prolonged sterilization experiments using SC plasma devices leads to inconsistent sterilization results. In fact, the SEM /EDS study was designed in order to identify and study the development of this inconsistency. However, even a fter 25 sterilization cycles (~50 minutes of plasma exposure), the reason for this inconsistency is not apparent. This inconsistency is not noted in the case of FR4 plasma devices. Alternative dielectrics like p olymethyl methacrylate (PMMA, or more common ly, acrylic), Teflon or Kapton or gorilla glass are of interest, but are not available with pre applied electrode material A preliminary study was conducted using PMMA dielectrics. PMMA devices were fabricated using acrylic slabs. Figure 4 1 2 Compa rison between a FR4 device and a PMMA device. A and B depict the unpowered FR4 and PMMA devices respectively. C and D depict the same device, powered.
90 The current electrode design was fabricated using copper adhesive tape and adhered to the PMMA slab (elec trode dimension 2.4 x 2. 4 cm 2 ). Figure 4 1 2 above demonstrates such a device compared with a similar FR4 device. Such PMMA devices were subjected to sterilization tests to obtain a survival curve in either case. 40 l of yeast was used as the inoculation volume. Complete sterilization was noted in 90s with the PMMA devices However the method of fabrication and testing for the PMMA devices was crude. Additionally when the devices were dipped in ethanol to disinfect them before sterilization experiments, th e adhesive of the copper tape dissolved, thus leading to a distorted electrode shape. This introduced an uncertainty in the experimental results. However, t he aim of the tests with PMMA was to explore plasma sterilization capabilities using a different die lectric (other than FR4 and SC). This confirmation was provided by the PMMA sterilization tests. 4. 4 Input Power and F requency The input power to the plasma devices was calculated via the measured input voltage and the input current. Input voltage and cur rent are measured using a Tektronix P6015A high voltage probe, a current probe (Corona Magnetics Inc.) and an Agilent DSO1004 Oscilloscope Power (P) was then calculated by the formula (4 4) Where V and I are the measured input voltage and current respectively and N is the number of cycles over which voltage and current were measured. The input power measured for a clean device versus an inoculated de vice, for FR4 and SC, is shown below in Figure 4 1 3 In order to plot this power over time, a
91 clean/inoculated device was powered at 14 kHz, 12 kV pp for 2 minutes, over the course of which, voltage (V) and current (I) were sampled every 15s. Figure 4 1 3 Comparison of the temporal variation of input power for clean and inoculated devices in the case of FR4 and SC dielectric. From Figure 4 1 3 the input power absorbed by the clean FR4 device was observed to be greater than that absorbed by the clean SC device. This measured power remained almost constant over the entire 2 minute interval. It was also observed that the input power absorbed by the inoculated FR4 device is greater than that absorbed by the inoculated SC device. However this measured power v aries over the 2 minute interval. Initial measured input power is low, increasing over the 2 minute interval. Also studied was the dependence of sterilization effectiveness on the input voltage. Input voltages were varied from 6 kV pp to 14 kV pp in incr ements of 2 kV pp Figure 4 1 4 given below shows the variation of average measured input power versus each input voltage tested for a clean FR4 and SC device.
92 Figure 4 1 4 Comparison of the average measured input power (W) for each input voltage (kV pp ) Evaluating the trend lines of the plots in the above figure, an expected quadratic dependence is observed (P V 2 ). For each input voltage, an inoculated device was powered for 2 minutes and then subjected to the post processing protocol described in Section 3.5.2 Tests were performed in triplicate. The variation of input power over time, for different voltages in the case of an inoculated FR4 and SC plasma device is given below in Figure 4 1 5 Figure 4 1 5 shows similar trends in the case of both FR4 and SC. At V= 6 ,8, 10 kV pp plasma is barely generated and the liquid bacterial sample deposited on the device does not evaporate at all, which explains the minimal variance in power observed for these voltage values. At V= 12 and 14 kV pp the deposited bacterial sample starts evaporating soon after 30s and is completely evaporated at 60s for FR4. Correspondingly, in Figure 4 1 5 (A), a rise in input power is observed until 6 0s, after which power remains constant. However, in Figure 4 1 5 (B), a continuous rise in input power is observed, since the sample is constantly evaporating in the case of SC.
93 Figure 4 1 5 T emporal variation of input power for different input voltages using inoculated devices for A) FR4 and B) SC. Figure 4 1 6 Dependence of sterilization effectiveness on input voltage (V)
94 Sterilization tests at each of the different input voltages were als o conducted in order to understand the dependence of sterilization efficiency on input power. This is shown above in Figure 4 16. Since the power varies over time for an inoculated device, Figure 4 1 6 has been plotted with the X axis corresponding to inpu t voltage and the Y axis corresponding to number of survivors ( log 10 ). From Figure 4 1 6 it is concluded that lower voltages (6,8,10 kV pp ) are not effective both in the case of FR4 and SC. In the case of SC, 12 kV pp seems to be inconsistently effective, but 14 kV pp produces complete bacterial inactivation in all the trials considered. Figure 4 1 7 Sterilization results at frequency f= 60 kHz A) Temporal variation of input power for an inoculated FR4 and SC device (f= 60 kHz, V= 9 10 kV pp ) B) Depende nce of sterilization on input voltage at f= 60 kHz, for FR4 and SC
95 Similarly, in order to understand the effect of varying input frequency on sterilization, tests were conducted at a higher frequency (60 kHz) using the portable plasma sterilization setup, described in Section 3.3 At such a high frequency, the setup was only capable of operating at a maximum voltage of 10 kV pp Hence sterilization tests were conducted at 9 and 10 kV pp Figure 4 1 7 (A) above shows the temporal variation of input power at 60 kHz, V= 9,10 kV pp for both FR4 and SC devices. The input voltage and current were sampled every 30s over a 2 minute interval. The trend of increasing input power w ith respect to time, for an inoculated case, is noted here also. Figure 4 1 7 (B) shows th e survival curves obtained for these tests. In the case of FR4, an input voltage of 9 kV pp leads to complete bacterial inactivation in 2 out of 3 cases. An input voltage of 10 kV pp leads to complete bacterial inactivation in all three cases. In the case of SC, an input voltage of 9 kV pp does not lead to bacterial inactivation in any of the cases. However 10 kV pp leads to complete bacterial inactivation in 2 out of 3 trials, thus implying that 10 kV pp might be the threshold sterilization input voltage for SC plasma devices An input voltage sligh tly higher than 10 kV p p or a sterilization time slightly higher than 2 minutes might be enough to ensure repeatable, complete bacterial inactivation in the case of SC. 4. 5 Operating P ressure The aim of the plas ma sterilization experiments conducted at pressures lower than normal atmospheric pressure was to evaluate whether DBD surface plasma sterilization was enhanced at lower pressures. Low pressure experiments were conducted in a vacuum chamber. The chamber w as made out of acrylic and constructed such that it can be evacuated to pressures as
96 low as 20 Torr. For each experiment, an inoculated device was placed inside the chamber, the chamber sealed and the air evacuated to the requisite pressure, using an air p ump. Plasma was generated in this low pressure environment for the requisite sterilization time Once this was done, p ressure was slowly increased back to atmospheric pressure, chamber opened and the device removed and subjected to the post processing protocol described in Chapter 3 There were a number of difficulties associated with using the vacuum chamber at reduced pressures for DBD surface sterilization experiments. The total time required for placing the inoculated device into the chamber, evacuating the chamber and powering the device takes ~1.5 minutes. In this time, there is the possib ility that some of the micro organisms in the bacterial sample on the device may be adversely affected due to the low pressure environment inside the chamber Also, due to operating constraints imposed on the ozone meter, ozone measurements were not possi ble. Another operating constraint was the limitation on operating pressure, due to device design. As the pressure is lowered, the plasma glow becomes more diffuse, as shown in Figure 4 1 8 Figure 4 1 8 Images of the devices at A) 760 Torr B) 500 Torr C) 400 Torr
97 Plasma glow is confined to the electrode surface area. At 500 Torr, plasma glow extends a little beyond the electrode surface area and at 400 Torr, it has extended almost to the edge of the dielectric surface. It was noticed that for P<350 Tor r, the plasma glow extended beyond the edge of the dielectric surface causing the formation of an electric arc This is because at P<350 Torr, the extremely high voltage is applied close to the edge of the device, in which case, electrons travelling from t he cathode (grounded electrode) take the path of least resistance to the anode (powered electrode) and hence arc over the edge of the device. Thus operating pressure had to be limited to 400 Torr, in order to work with the plasma devices currently being us ed. The spectroscopic signatures obtained during the operation of the plasma devices at reduced pressures was observed to be similar to the spectroscopic signature of DBD plasma at atmospheric pressure, except that the intensity of the emitted spectra at the peak wavelengths was found to increase with decreasing pressure. More on the spectroscopic results will be discussed in Chapter 5. Inoculated devices were placed in the vacuum chamber and powered for two 19 below shows the results from the two trials for (A) FR4 and (B) SC In the plot s below, the st erilization behavior is observed to be similar to that at atmospheric pressure for both dielectrics. It is to be noted here that unlike Figure 4 6, after plasma exposure for 2 minutes, Figure 4 19 (B) shows incomplete sterilization using SC plasma devices at reduced pressures This is to be expected. As discussed in Section 4.3, SC plasma devices show an inconsistency in sterilization behavior after 2 minutes of plasma exposure. Hence while fresh devices were used for the experimental
98 results shown in Figu re 4 6, devices that had been subjected to multiple sterilization cycles were used for results shown in Figure 4 19 (B). If a similar device were to be used to obtain a survival curve at atmospheric pressure, complete sterilization would not be seen after 120s of plasma exposure. Figure 4 1 9 Sterilizati o n behavior at reduced pressures for A) FR4 and B ) SC plasma devices. Two sterilization times (t= 60s and 120s) were tested using E. coli at the reduced pressure s Sterilization behavior at reduced press ures is found to be similar to that at atmospheric pressure. For FR4, at 500 Torr, incomplete bacterial inactivation is achieved after 2 minutes on Trial#2. However taking into account the complete bacterial inactivation achieved for both trials at P=400 Torr, for 2 minutes, it is believed that that the
99 observation at P=500 Torr for FR4 is an experimental outlier, rather than indicative of sterilization behavior at reduced pressure. 4.6 Discussion Th is Chapter outlined a number of parametric studies, aime d at characterizing the sterilization efficiency of the AC, RF DBD surface plasma used in this study. A variety of pathogens are subjected to plasma generated at 14 kHz, 12 kV pp The time taken for complete sterilization is determined in the case of each pathogen. Using FR4 plasma devices, this sterilization time is determined to be 90s 120s for E. coli and y east. Using B. subtilis cells, this sterilization time is determined to be 4 minutes. However the medium in which B. subtilis cells are cultured is o bserved to make a difference. Using G. stearothermophilus spores, this sterilization time is determined to be 20 minutes. A range of other pathogens (as shown in Table 4 1) are also tested, most of which are completely sterilized within 3 minutes of plasma exposure. Additionally doubling the volume of the E. coli sample deposited on the surface of the plasma device is observed to extend the sterilization time by ~30s. However, the number of CFU in 40 same as the number of CFU Hence it seems that the sterilization time is dependent on the volume of the liquid in the E. coli sample. This is an importa nt insight into the mechanism of surface DBD plasma sterilization and will be discussed more in Chapter 5. Two different dielectric materials (FR4 and SC) are compared in terms of sterilization efficiency. It is observed that the sterilization time in the case of SC is longer than that in the case of FR4. One hypothesis to explain this can be drawn on the basis
100 of the dielectric constant, which is the only differentiating parameter between both dielectric materials. The d ielectric constant of FR4 is ~ 30 % h igher than the SC. The dielectric constant of a material is the ratio of amount of electrical energy stored in a material by an applied voltage, relative to that stored in vacuum. Any of the devices described in this study can be considered as a parallel p late capacitor, using a dielectric ( 4 5 ) 0 = absolute permittivity of air, A= surface area of the top surface of the device and Hence, for the plasm a devices considered in this paper, energy stored in the device is directly proportional to the capacitance of the system, which in turn is directly proportional to the dielectric constant of the material. T he FR4 device, which has a higher dielectric cons tant than the SC device, has more energy stored in the dielectric layer, which explains the results noted in Figure 4 6 i.e. complete sterilization is achieved faster (t= 90s) for FR4 as compared to semi ceramic (SC) dielectric (t= 120s). Chapter 5 will f urther demonstrate the differences between FR4 and SC plasma devices, all of which can be ascribed to the difference in absorbed power between both devices, which in turn can be attributed to difference in dielectric constant (as shown above). However sinc e only two dielectric materials have been tested, the validity of this theory has to be tested using additional dielectric materials.
101 Furthermore, it is also noticed that the SC plasma devices do not ensure consistent sterilization behavior (Figure 4 6) Prolonged sterilization cycles using the SC devices are observed to increase this inconsistency. The SEM analysis of both dielectric substrates, for different protocols (describe d in Section 4. 3 ), demonstrates that while the dielectric surface is not visib ly altered by repeated plasma generation (both in the clean and inoculated cases), the electrode surface presents a highly segmented/sputtered appearance in the inoculated cases. This observation holds true for the electrodes on both FR4 and SC devices. Th e reason for this is not immediately apparent but can be speculated to be due to the corrosive action of salts used in the preparation of the bacterial sample. However the SEM analysis does not readily highlight why the SC plasma devices are found to demon strate an inconsistency in terms of sterilization behavior. Different input voltages and different input frequencies, for both FR4 and SC, are tested in order to understand the dependence of sterilization effectiveness on input power and frequency In ord er to compare the sterilization performance of these devices, a parameter independent of device surface area as well as sterilization time required needs to be selected. Electrode surface area of the plasma devices is ~5.76 cm 2 Considering the temporal va riation of input power over time for inoculated FR4 and SC devices in Figure 4 15, it is evident that at lower input voltages ranging from 6 10 kV pp, temporal variation of input power is not significant. At higher input voltages such as 12 and 14 kV p p, this temporal variation is more evident.
102 Equation 4 4 describes the formula used for calculation of power (P). Thus for a be calculated by the formula 4 6 Where n= t/ t. Thus for a sterilization time t= 2 minutes= 120s, if power P i was measured at intervals of 15s, then n= 120/15= 8. Once this average power has been calculated, then the average power density (P den ) can be calculated by the formula 4 7 Where A is the electrode surface area of the plasma device. Finally plasma dose can be calculated by multiplying P den with the treatment time [ 20 ] Figure 4 20 below gives a plot of the number of survivors ( N) after exposure to varying energy flux injected by the plasma (plasma dose) in the inoculated FR4 and SC devices. Figure 4 20. Dependence of bacterial inactiva tion on the applied plasma dose
103 From Figure 4 20 above, it is evident that am energy flux > 285 J/cm 2 is required for complete bacterial inactivation in the case of E. coli. This plasma dose is even greater in the case of SC plasma devices, as compared to FR4 plasma devices. This plasma dose is an intrinsic parameter that can be calculated and u sed to compare the sterilization performance of different plasma devices in the case of different microorganisms. From Figure 4 1 7 (A) and (B), it is observed that plasma generation at a higher input frequency of 60 kHz and an input voltage of 9 and 10 k V pp leads to higher input power densities in the case of FR4 as compared to SC. Plasma doses calculated in a manner described above yielded values greater than 285 J/cm 2 for both 9 and 10 kV pp in the case of FR4. In the case of SC, plasma doses calculate d at 9 and 10 kV pp were less than 200 J/ cm 2 Consequently at 10 kV pp complete bacterial inactivation is observ ed in all three trials for FR4. In the case of SC plasma devices the SC devices need to be powered at a voltage higher than 10 kV pp or powe re d at 10 kV pp for t>2 minutes to ensure consistent complete sterilization Lastly, sterilization experiments are also conducted at reduced pressures of 400, 500 and 600 Torr. The sterilization trends in these cases are found to be similar to those observed at atmospheric pressures. However comparing spectrum data at these different pressures, intensity peaks at similar wavelengths are found. These intensity values are found to increase with decreasing pressure. This increase in intensity does not seem to he lp sterilization, as the sterilization trends noted in Figure 4 17 are similar to those noted at atmospheric pressure. The spectra for the different pressures are described in further detail in Chapter 5, along with a discussion of the different
104 components involved in DBD plasma generation and their role in influencing plasma sterilization.
105 CHAPTER 5 UNDERSTANDING THE MECHANISM OF DBD SURFACE PLASMA STERILIZATION Chapter 4 was aimed at varying the different parameters involved in plasma generation and s tudying the effect of this variation on plasma sterilization. The current chapter focuses on the fundamental question: How does DBD surface plasma sterilization work? Diagnostic data here refers to the spectroscopic, ozone and temperature data obtained du ring plasma generation, both with a clean and inoculated device, in the case of both dielectrics (FR4 and SC). In the sterilization process, bacteria could be getting irradiated by the UV photons or could be chemically reacting with one or more of the reac tive species produced during plasma generation. Though DBD plasma is low temperature, non uniform average surface temperatures have been detected in the range of 300 340 K ( 27 67 o C ) and 293 313 K ( 20 40 o C ) for FR4 and SC respectively The upper end of this range is harmful to many bacteria, but not bacterial spores [ 96 ] It remains to be determined whether this also plays a role in DBD surface plasma sterilization Additionally experiments isolating each of these agents and examining their individual effect on bacterial concentrations are also required in order to determine the role of each agent in the plasma sterilization process. The current chapter summarizes work done in these areas 5.1 Spectroscopic S tudies The spectroscopic signature of the generated DBD plasma was determined using the Ocean Optics USB 20 00+ spectrometer. The setup of the spectrometer has been described in Chapter 3, Section 3.1.
106 Spectroscopic data was measured in several different scenarios. Clean and inoculated devices were powered continuously for 2 minutes (120s) in the case of both F R4 and SC. The spectrometer samples the spectroscopic data every 10s. The recorded spectral data in the case of a clean and inoculated FR4 device is given below in Figure 5 1. The same, for the case of a clean and inoculated SC device is given below in Fig ure 5 2. For the sake of simplicity, only the spectroscopic signature at the plasma treatment intervals of 30, 60, 90 and 120s are given in the plots below. Spectra in Figure 5 1 and Figure 5 2, for the clean and inoculated cases are similar Figure 5 1. Spectral signature of A) a clean FR4 device and B) an inoculated FR4 device Y axis lists emission intensity in arbitrary unit
107 Figure 5 2. Spectral signature of A) a clean SC device and B) an inoculated SC device Y axis lists emission intensity in arbitrary unit The intensity peaks at different wavelengths correspond to photonic emission by different chemical species. Once the intensity peaks and their corresponding wavelengths are identified, the corresponding emitting chemical species can be det ermined [ 97 ] In Figure 5 1 and Figure 5 2 there are four principle intensity peaks: 31 5 25 nm, 336.67 nm, 357.24 nm and 379. 52 nm. All four correspond to transitions taking place in the 2 nd positive system of N 2 (C 3 u B 3 g ). The peak intensities are seen at wavelengths corresponding to N 2 which makes sense considering that N 2 makes up over 78% of air. T here are no reactive oxygen species produced according to the spectrum. Mao et al. [ 98 ] showed that this was due to the fact that the process of
108 excited collision of electrons with Nitrogen (N 2 ) is far stronger than that of collisions with oxygen (O 2 ) when reduced electric field (E/N) varies from 100 Townsend (Td) to 1000 Townsend (Td) Simulating the process of a DC discharge using mont e carlo simulation methods, they calculated that the probability of an electron colliding with N 2 was 30 times more likely than that of an electron colliding with oxygen (O 2 ) Choi et al. [ 99 ] analyzed the spectrum of a pul sed DBD discharge (volume discharge configuration at 1 kHz frequency). While they noted similar intensity peaks for N 2 in the 300 400 nm range, they also noted intensity peaks for O atoms at wavelengths higher than 394 nm. A nalysis of the spectrum recorded by other authors [ 68 ] for an air plasma (in volume discharge configu ration) demonstrated no significant wavelengths below 285 nm. In Chapter 1, the lethal effects of UV radiation at 254 nm and the sterilizing effect of VUV radiation have been described in great detail. Since the spectral signature in Figure 5 1 and Figure 5 2 shows no noticeable wavelengths below 290nm, it is unlikely that shortwave UV radiation (200 300 nm) plays a major role in surface DBD plasma sterilization. Figure 5 3 and 5 4 are expanded version s of Figure 5 1 and Figure 5 2, focusing on the spectral signature in the range 330 350 nm. This is the wavelength range in which the highest intensity peak is observed for each of the spectra in Figure 5 1 and 5 2. From Figure 5 3 (A) and 5 4 (A) below it is evident that p lasma generated using a clean FR4 and SC device for a 2 minute time interval, shows similar intensity values at all times. However it is evident from Figure 5 3 (B) and 5 4(B) that there is a temporal variation of spectroscopic intensity. Intensity at 30s is the least while intensity at 120s is the highest.
109 Figure 5 3 Expanded version of (A) Spectral signature of a clean FR4 device (B) Spectral signature of an inoculated FR4 device. Figure 5 4 Expanded version of (A) Spectral signature of a clean SC device (B) Spectra l signature of a n inoculated SC device.
110 Spectroscopic measurements were also recorded using inoculated and clean FR4 and SC plasma devices at reduced pressures. The vacuum chamber in our lab is built out of acrylic and is designed to be evacuated to pressures as low as 20 Torr. However, due to the design of the plasma device, it was only possible to operate the plasma devices at P > 400 Torr. Hence experiments were conducted at 400, 500 and 600 Torr. Figure 5 5. Spectroscopic comparison of (A) FR4 devices and (B) SC dev ices at reduced pressure Figure 5 5 above presents the spectroscopic signature of (A) an inoculated FR4 and (B) an inoculated SC device, when powered at 14 kHz, 12 kV pp inside the vacuum chamber at P= 400,500,600 Torr. For reference, the spectroscopic s ignature of a similar FR4 and SC device, powered at atmospheric pressure, is also shown. The
111 devices were powered for a total time of 2 minutes and s pectroscopic data was sampled at every 10s. The spectroscopic signature at one particular time point (t= 12 0s) and for one particular wavelength range (330 350 nm) is shown in Figure 5 5. At reduced pressures the spectroscopic signature is observed to be similar to the ones shown in Figure 5 1 and 5 2, much due t o reduction of pressure. In particular, no new short UV wavelength peaks appear. However 400, 500 and 600 Torr correspond to 0.53, 0.66 and 0.79 atm respectively, which is still close to atmospheric pressure Hence a drastic change in spectrum is not expec ted. In Figure 5 5, which depicts a single intensity peak in the wavelength range of 330 350 nm, different intensities are noted at different pressures. The intensity at 400 Torr is highest and the intensity at atmospheric pressure is lowest. The gradual ly increasing trend in intensity (depicted in Figure 5 3 and 5 4) mimics the gradually increasing trend observing in measured input power (discussed in Section 4.4). As with the case of the measured input power, here also, the temporal variation of intensi ty is observed to correspond to the gradual evaporation of the liquid sample. Also, the spectroscopic intensity in the case of the FR4 devices is far greater than that of the SC devices (both clean and inoculated). Considering the points noted above, the i ntensity difference in both cases raises the question of whether spectroscopic intensity (i.e. photons) plays an important role in plasma sterilization. Moisan et al. [ 16 ] discussed photo desorption (etching of bacteria by the emitted UV photons) as one of the possible mechanisms in plasma sterilization. The question then is whether photo desorption plays a role in DBD surface plasma sterilization.
112 The comparative ratio of spectroscopic intensities can also be used to simulate rotational and vibr ational plasma temperatures [ 100 101 ] However, the spectroscopic intensity measured here is basically a measure o spectroscopic intensities obtained cannot identify the population of excited species or photons, for the current setup of the spectrometer. Hence it is difficult to determine the role of photo desorption, if any, in t he surface plasma sterilization process. The spectra discussed in this study have been obtained using emission spectroscopy methods. Another method to measure in situ particle concentrations in an atmospheric pressure DBD plasma setup is to use absorption spectroscopy. Spaan et al. [ 102 ] and W inter et al. [ 1 03 ] use absorption spectroscopy to measure nitric oxide (NO) and ozone (O 3 ) concentrations respectively in a DBD plasma. The advantage of such a method, as compared to emission spectroscopy is that the experimental setup can be designed so as to measure the concentration of one particular species during plasma generation. Thus elusive species such as ozone and oxygen emissions can be detected with much more confidence using absorption spectroscopy. This section outlined spectroscopic data obtained during plasma generation using clean and inoculated FR4 and SC devices. Spectrscopic data at reduced pressures was also examined. The data shows that dominant intensity peaks are obtained at wavelengths corresponding to transitions in the N 2 2 nd positive system. This leads to the following questions: a) Are the reactive nitrogen species being produced the only chemically erosive species acting in the sterilization process? b) Could ozone, be acting synergistically to enhance sterilization? Section 5.2 discusses t hese two questions in more detail.
113 5.2 Ozone S tudies DBDs are known ozone generators. Ozone formation is a two step process that starts with the dissociation of O 2 molecules by the electrons in a micro discharge [ 104 ] : (5 1) And a subsequent three body reaction ( 5 2) where M is a third reaction partner. M can be a pure oxygen species or even a nitrogen species, acting as a catalyst. Eliasson [ 104 ] hypothesize d that the formation time of oxygen atoms is much smaller than that of ozone. The efficiency of ozone production in a micro discharge depends very much on its strength, which is influenced by parameters such as the gap spacing between electrodes, operating pressure, dielectric material etc. [ 105 ] The strength of the micro discharge is further characterized by the relative concentration of oxygen atoms ([O]/[O 2 ]) in the channel. As this ratio increases, i.e the streng th of the micro discharge increases, the efficiency of the conversion of oxygen atoms [O] to ozone [O 3 ] decreases. Thus Kogelschatz et al. [ 105 ] state that the optimum micro discharge strength for ozone generation is a compromise between avoiding energy losses to ions and still achieving a reasonable efficiency of conve rsion of oxygen atoms to ozone. Eliasson et al. [ 104 ] also state that ozone forms at a faster rate than the rate at which it migh t dissociate (equation 5 3) or its concentration might get diluted due to recombination (equation 5 4). Ozone Dissociation: (5 3) Ozone Recombination: (5 4)
114 If the oxygen concentration is too high (or the tempe rature is low), recombination ( 5 4 ) is the dominant reaction. If the temperature is too high, thermal decomposition ( 5 3) is the dominant chemical reaction. As discussed in Section 1.3.2, the role of ozone or more generally, the r ole of reactive oxygen spe cies in plasma sterilization is still widely debated. This section concentrates on isolating the role of ozone in DBD surface plasma sterilization. T h e plasma device was enclosed within a sterilization chamber made of a crylic. Rates of ozone production/dec ay while operating this DBD plasma source were determined. The plasma device was also placed within sterilization chambers of different volumes and powered in order to study the dependence of ozone production/decay on chamber volume. Additionally, the effe ct of exposing bacterial ( E. coli ) concentrations to the ozone thus produced was also analyzed. In doing so, the lethal amount of ozone necessary to completely inactivate E. coli within 7 minutes was determined. In order to prove that this inactivation ef fect was due to the ozone produced during DBD plasma generation, ozone production was inhibited during plasma generation. Known concentrations of E. coli were exposed to the ozone produced in this case too, but no significant inactivation was noted, thus l eading to the conclusion that the ozone produced during plasma generation was indeed responsible for bacterial inactivation. As previously described in Section 3.1 a 2B Tech Ozone monitor was used to measure the ozone levels at fixed time intervals wit hin a closed chamber. 5.2.1 Characterization of Ozone Production a nd Decay During DBD Plasma Generation Sterilization chambers of four different volumes were used for the experiments in this paper. The dimensions of these are listed below:
115 Chamber#1 1 840 in 3 Chamber#2 1680 in 3 Chamber#3 3360 in 3 Chamber#4 27072 in 3 The volume ratio of the chambers is 1:2:4:32, w ith respect to Chamber#1. In order to characterize the ozone production trends during DBD plasma generation, preliminary experiments consisted of placing a plasma device in Chamber#4, powering the device over a time interval of 2 minutes and measuring the ozone levels at different locations (both along the X axis and Y axis) of the chamber. The various locations at which ozone was measured inside the chamber are shown in Figure 5 6 For all measurements in Figure 5 6 the floor of the chamber. Figure 5 6 Schematic of Chambe r#4. The grey square in the middle represents the plasma device. The black short lines represent the different locations at which apart), along the X and Y axis.
116 The aim of such an e xperiment was to get an idea of the spatial variation of ozone levels inside Chamber#4. This spatial variation is given below in Figure 5 7 (A) and (B). Figure 5 7 (A) shows the spatial variation of ozone production along the X axis of the chamber while Figu re 5 7 (B) shows the spatial variation of ozone production along the Y axis of the chamber. While ozone data is sampled every 10s, for the sake of simplicity only the data at 30s, 60s, 90s and 120s ha s been plotted and shown. Figure 5 7. Spatial variat ion of ozone distribution (A) along the X axis in the sterilization chamber ( B ) along the Y axis in the sterilization chamber Figure 5 7 ( A ) demonstrates that the highest amount of ozone is produced in the upper right quadrant of the chamber (as labeled in Figure 5 6 ). Typically the levels of ozone noted on the right hand side ( RHS ) of the chamber are more than those noted on
117 the left hand side (L HS ). Figure 5 7 ( B ) indicates that the distribution of ozone produced along the Y axis does not follow a clear cut trend, as along the X axis. This bias in ozone levels towards the R.H.S of the chamber can be explained due to the electrodynamic force ( ) produced as a result of the existent electric field during plasma generation where F is the electrodynamic force, q is the charge and E is the produced electric field. Previous research [ 106 107 ] elucidates further on the generation of this electrodynamic force and its dependence on the various plasma input parameters as well as its effect on fluid momentum. Most importantly, the direction of this force is from the powered electrode to the grounded electrode. As a result of this electrodynamic force (from left to right), the flow is pushed to the right, as shown in Figure 5 8(A ). When the configuration is fl ipped, as shown in Figure 5 8 ( B ), the direction of force is from right to left. Accordingly a higher level of ozone is detected on the L.H.S of the chamber in this configuration. Figure 5 8 T he two different configurations in which the device is placed. It is noted that in ( A axis). In ( B ), when the
118 Next, the trend of ozone production and decay during DBD plasma generation and the dependence of this production/decay on chamber volume are evaluated. In order to do this, a clean FR4 device was taken and placed inside the given sterilization right of the device (measu red from the center point of the device). The device was powered for 2 minutes, turned off and the setup was allowed to rest for an additional 5 minutes, before repeating the same protocol using the same device for the next sterilization chamber. Productio n/decay profiles were obtained in this manner for all four sterilization chambers. Analyzing the measured ozone data, using trend fitting tools, Equation (5) and (6) are derived. The total ozone concentration ([O 3 ]) and time (t), during plasma generation follow a power law relationship, as shown below. (5 5) W here t (total time of plasma generation i.e. ozone production) = t 1 + t 2 and A, B, C, D are constants. In equation (5 5 time segments 1 2 1 2 time during which ozone is still produced, but at a much slower rate than during t 1 (90s). Similarly, the decay of ozone in all four chambers is observed to follow a logarithmic trend of the form given below: (5 6) where is the initial measured concentration of ozone and is the ozone decay constant.
119 A, B, C, D and are listed in Table 5 1 below for each chamber. Table 5 1 Values of the constants used i n Equation (5 5 ) ( 5 6) Chamber# A B C D [O 3 ] 0 (ppm) (s) 1 7.08 0 0.65 8 8.692 0. 59 7 174.94 0 909.091 2 4.999 0. 55 7 8.735 0.408 63.958 1250 3 9.216 0.465 12. 11 6 0. 30 9 31. 87 9 1250 Figure 5 9. Comparison of ozone concentrations in all four chambers. T he plasma device is powered at 0 min and turned off at 2 minutes, after which the setup is allowed to sit for another 5 minutes Thus, in Figure 5 9 above, both measured ozone data as well as curve fitted data from Equations (5 5 ) and ( 5 6) have been plotte d. In the case of measured ozone data, error bars have been plotted with respect to standard deviation. In Figure 5 9 it is noticed that the plots obtained using trend fitting tools are most accurate in the case of Chamber#1. For Chamber#2 and #3, these p lots do not follow the exact same trend as the measured data. This implies that as the volume of the chamber increases, it becomes increasingly difficult to predict the rate of production/decay of ozone. It is for
120 this reason precisely, that constants for Chamber#4 have not been listed in the above table. Owing to its comparatively larger volume, it is highly difficult to evaluate accurate trends of production/decay in Chamber#4. It must also be noted that the values of these constants apply to the ozone le vels measured only at the probe position described above. These are not indicative of the ozone production/decay throughout the rest of the chamber. Plotting the measured ozone levels with the corresponding chamber volume, at different time points during the 7 minute interval produces a correlation such as the one given below in Figure 5 10 Here 5 time points have been plotted: 60s, 120s (during plasma generation) and 240,360 and 420 s (plasma is turned off). Both chamber volume (X axis) and ozone levels (Y axis) have been plotted on a log 2 scale for easier comparison. Figure .5 10. Correlation of the ozone levels with the corresponding chamber volumes. The different time points (60,120,240,360,420s) represent ozone measurements in each chamber at that particular time point.
121 From Figure 5 10 it is observed that ozone levels measured in Chamber#1 at any time are the highest. During the plasma generation phase (60,120s), ozone levels produced ozone towards the right (as shown in Figure 5 8 ), it seems that the bulk of produced ozone is continuously pushed towards the right and in an upward direction during plasma generation. Since both Chamber#2 and #3 have the same height, even though the latter is twice as long as the former, the ozone produced seems to dissipate at an identical rate in both chambers, which is why similar amounts of ozone are measured in both cases. This implies that during plasma generation (and thus, acti ve ozone production), chamber height is the factor that determines the amount of ozone present in the chamber. This also explains why the ozone concentration in Chamber#1 is almost twice that of Chamber#2. Chamber#1 is half the height of Chamber#2 (same le ngth and breadth), which means that the produced ozone has even less volume to dissipate in. Hence it is measured by the probe in greater concentrations. During the decay phase, once the plasma is turned off (240,360,420s), it is observed that the rate of decay of ozone in Chamber#1 is almost double that in Chamber#2. The same trend applies to #2 when compared with #3. This implies that larger the chamber volume, faster the rate of decay of ozone. This is reinforced by the values of [O 3 ] 0 and given i n Table 5 1. In Figures 5 7, 5 9 and 5 10 measured ozone levels have been listed in terms of ppm, for the sake of easy comparison. At any instant, these ozone levels are measured by one ozone probe, placed at the same position in each chamber. Figure 5 9 seems to indicate that Chamber#1 contains the highest ozone concentrations while Chamber#4
122 contains the least ozone concentrations. But it is the same FR4 plasma generator producing ozone for the same amount of time. Thus in order to understand the total concentration of produced ozone in each chamber, the total number of ozone molecules in eac h chamber is calculated. T he total number of molecules of ozone in each chamber can be calculated by first converting ppm to mg/m 3 calculating the number of moles o f ozone/volume in each chamber, multiplying this number by the volume of the chamber and subsequently calculating number of molecules of ozone in each chamber (1 mole= 6.023x10 23 molecules). Figure 5 11 given below shows the total number of molecules of oz one (calculated) in each chamber at t= 60,120,240,360 and 420 s Figure 5 11. Correlation of the total number of ozone molecules present in each chamber at t= 60,120,240,360,420s to the respective chamber volumes. The X axis is plotted on a log 2 scale f or easier comparison. #1 #4 represents the different chambers.
123 From Figure 5 11 above, it is evident that at any given point, the total number of ozone molecules in any chamber ranges between 10 22 23 This variation may be due to the increase in chamber s ize. When the ozone production of the two dielectric materials compared for sterilization purposes, FR4 and SC, was examined during plasma generation, it was observed that the ozone production in the case of FR4 was marginally higher than that in the case of SC. This comparison is shown below in Figure 5 12 Fi gure 5 12. Comparison of ozone concentrations during 7 minutes for FR4 versus SC. The plasma device is powered at 0 min and turned off at 2 minutes, after which the setup is allowed to sit for a nother 5 minutes In Figure 5 12 it is observed that the rate s of production and decay of ozone in the case of SC is similar to that in the case of FR4. Comparing the ozone concentration at different times for FR4 and SC, it is noted that on an average, th e ozone concentration in the case of FR4 is ~25.8% higher than in the case of SC.
124 In this section, an overview of the trends in ozone production and decay during DBD plasma generation was provided. O wing to the volume of the sterilization chamber, the oz one produced dissipates at different rates. This also means that if an inoculated device were to be placed at one constant position in each chamber, it would be exposed to varying concentrations of ozone. This helps determine the maximum level of ozone nee ded in order to completely inactivate E. coli The effect of exposing E. coli to this produced ozone was studied. The results of such a study are listed in the next section. 5.2.2 The Effect o f Ozone Produced During DBD Plasma Generation o n E Coli As ment ioned before, DBD plasma generation produces UV photons, charged particles and neutrals. Since DBD plasma produces high concentrations of ozone, the effect of exposing E. coli concentrations to th is produced ozone was tested. Figure 5 13. Experimental schematic for the ozone exposure tests. On the left is the plasma generator, which is used to generate plasma and produce ozone inside the chamber. On the right is the inoculated device/glass slide, which is E. coli and exposed to the ozone produced by th e plasma generator The schematic for the ozone exposure experi ments is given above in Figure 5 13 height as the devices so as to accurately capture the ozone levels that the ino culated
125 device is subject to. Judging by the ozone distribution in Figure 5 8 the ozone probe was placed on the R HS of the chamber, in order to measure the maximum amount of ozone in the chamber. For these tests, a clean FR4 device is selected as the pla sma generator or ozone generator. Three kinds of inoculated substrates (having equal surface areas) are tested: FR4, SC and E. coli (corresponding to a concentration of 10 6 8 CFU ). Once the inoculated substrate is placed next to the ozone generator, the chamber is closed. The inoculated substrate is allowed to sit in the chamber for total time (t 3 ) and a resting time (t 4 ). Hence t EXP = t 3 +t 4 At the end of t EXP the chamber is opened, the inoculated substrate removed and subjected to the post processing protocol described in Section 3.5 .2. Experiments are triplicated to ensure reproducibility. For the results given below in Figure 5 14, Sterilization Chamber#1 was used. Both a FR4 plasma generator and SC plasma generator were used to produce ozone. T hree kinds of inoculated substrates (FR4, SC, glass slides) were tested in the case of FR4 and SC plasma generator. T hree exposure times (t E XP = 2, 7, 32 min) were tested in the case of FR4 plasma generator. EXP 3 = 2 min and t 4 = 0, 5 and 30 minutes respectively However an exposure time of 7 minutes only was tested in the case of the SC plasma generator Ozone concentrations measured in the case of SC plasma generator were typically 20 30 ppm lesser than that measured in the case of FR4 plasma generator. Figure 5 14 given below depicts the results of such a test. Error is listed in terms of the standard deviati on from the mean of measurements.
126 Figure 5 14 : Inactivation plots due to ozone exposure with ( A ) F R4 plasma generator and ( B ) SC plasma generator In Figure 5 14(A), it is evident that in the case of FR4 plasma generator, ozone exposure is highly lethal to E. coli In the case of inoculated FR4 substrate, starting from an initial E. coli concentration of 10 8 CFU a ~3 log 10 reduction, ~6 log 10 reduction and complete reduction in E. coli concentration is noted at 2, 7 and 32 minutes respectively. A single log 10 implies a 1/10 th decrease in bacterial concentration i.e. 10 7 CFU reduces to 10 6 CFU In the case of inoculated SC substrate, similar log 10 reduction is noted at 2, 7 and 32 minutes, although complete inactivation is not noted at 32 minutes. However in the case of inoculated glass slides, incomplete inactivatio n is noted in all three cases.
127 In Figure 5 14(B) i.e. in the case of SC plasma generator, ozone exposure causes incomplete inactivation of E. coli in the case of all three inoculated substrates From the ozone concentrations noted for the FR4 Plasma generator and the SC plasma generator in Figure 5 12 and the sterilization data noted for both in Figure 5 14 it is evident that 120 150 ppm of O 3 is sufficient to cause a major reduction in E. coli concentration after 7 minutes of ozone exposure. It is conjectured that the 7 minute E. coli concentration was exposed to the same level of ozone for t EXP slightly longer than 7 minutes, complete bacteri al inactivation could be seen This conjecture is made in light of the fact that for the inoculated FR4 substrate, after 7 minutes of O 3 exposure, 4 out of 6 trials showed complete bacterial inactivation. It must be noted here that in the case of the inoc ulated glass slides, exposure to ozone produced using either a FR4 plasma generator or SC plasma generator does not seem to make a difference. This indicates that there is a substrate dependence that determines the effect of ozone on E. coli When glass sl E. coli it is visibly evident that the bacterial sample deposited on the glass slide clumps into random droplets on the surface of the glass slide. This is because glass is a far less hydrophilic surface than FR4 or SC, t hus making it difficult for liquid to adhere to it. Hence this uneven clumping of E. coli on the glass slide might be leading to shielding of the underlying bacteria, which would explain reduced killing in the case of glass slides. The results of exposing E. coli ozone in Chambers#2 #4 are shown below in Figure 5 15 Owing to the similar levels of inactivation noted in the case of inoculated FR4 and SC substrate in Figure 5 14 for the following cases, only an inoculated FR4
128 substrate is considered. As is evident, after 7 minutes, a nearly complete inactivation is noted in the case of Chamber#1. In Chambers#2 #4, after 7 minutes, only a reduction of 1 4 log 10 is noted. Figure 5 15. Inactivation plots due to ozone exposure in the different chambers using a FR4 plasma generator and an inoculated FR4 substrate This dependence of inactivation due to ozone exposure on chamber volume is explained by ozone concentrations noted in Figure 5 9 wherein it is observed that, using a FR4 plasma generator, an ozone p robe placed on the RHS of the chamber measures highest ozone concentration in Chamber#1 and lowest in Chamber#4. Accordingly, bacterial inactivation due to ozone exposure is highest in Chamber#1. In all other chambers, the reduction in bacterial concentrat ion is not significant. This further implies that there exists a threshold value of ozone concentration required for a major reduction in bacterial concentration, which in this case seems to be 120 150 ppm. The next step was to evaluate whether ozone prod uced during plasma generation is truly responsible for the almost complete bacterial inactivation noted in
129 Figure 5 14(A). In order to evaluate this, ozone produced was inhibited in the following two ways 1) Using activated charcoal to inhibit ozone produc tion 2) Generating plasma using Nitrogen (N 2 ) as the discharge gas A fixed amount of Activated charcoal (MarineLand Black Diamond ) was placed on the plasma generator and the plasma generator subsequently operated. In doing so, the produced ozone was dire ctly adsorbed by the charcoal and ozone levels were immediately reduced by around 98%. For subsequent tests, care was taken to adjust this amount of activated charcoal on the plasma device to maintain the same reduced levels of ozone. This comparison of l evels of ozone produced with and without charcoal for Chamber#1 is shown below in Figure 5 16 For both cases, a clean FR4 plasma device was placed in Chamber#1, powered at 0 min and turned off at 2 minutes, after which the setup was allowed to rest for a nother 5 minutes O zone concentration (on the Y axis) is shown on a log 10 scale. Figure 5 16. Comparison of ozone production with and without charcoal for Chamber#1. Plasma is turned off at 2 min.
130 As is evident from Figure 5 16 the addition of charc oal on top of the device during plasma generation leads to a reduction of ozone concentration by a factor of 100 For all the sterilization tests testing the effect of exposing E. coli concentrations to ozone produced in cases with and without charcoal, th e amount of charcoal on the device was adjusted so as to maintain the reduced ozone concentration as shown in Figure 5 16. For the purpose of this study such a configuration will be referred to as a e tests and FR4 devices were used as the modified plasma generators Figure 5 17. Inactivation plots due to ozone exposure in Chamber#1 with and without charcoal. A clean FR4 device was used as the ozone generator and covered with charcoal when needed. The device was powered for 2 minutes. When the inoculated substrate is exposed to this modified plasma generator for 7 minutes and then post processed, negligible reduction in E. coli concentration is observed, thus proving that the ozone produced during plasma generation is responsible for inactivation of E. coli This inactivation effect with and without charcoal, in the case of an inoculated FR4 substrate, is shown above in Figure 5 17 The wide disparity in bacterial inactivation is immediately eviden t and proves that the reduced
131 ozone concentrations due to adsorption by activated charcoal do not effectively kill E. coli when compared to the case of no charcoal. While Figure 5 17 demonstrates that ozone produced during DBD plasma generation is respons ible for bacterial inactivation, an alternative experiment to prove this was also conducted. DBD Plasma was generated using Nitrogen (N 2 ) as the working gas. A smaller was then evacuated to an absolute pressure of 0.0978 atm, following which Nitrogen gas ( Airgas Inc., UN1066, 99.0% N 2 ) was introduced into the chamber until the pressure in the chamber rose to 1 atm. This process was repeated four times to maintain a majority N 2 environment. The aim of such an intensive method of flushing out all the air and filling the chamber with nitrogen was to ensure that very low oxygen levels remain in the chamber in order to inhibit ozone production during DBD plasma generation. Using laws of partial pressures, this percentage of nitrogen, at the end of 4 flushes was calculated to be 99.998 + 0.78 %. Hence the aim of maintaining a pure N 2 environment was accomplished fairly well. The experimental protocol in these tests, unlike the prev ious ozone exposure tests, consisted of inoculating select FR4 devices with 20 l of E. coli (initial concentration= 10 8 CFU ), placing them in the sterilization chamber, sealing the chamber, flushing the chamber with N 2 four times and then powering the dev ice for the test was replicated thrice, using the same E. coli sample to ensure repea tability. The
132 comparison of sterilization results from plasma generation using discharge gas as air versus N 2 are given below in Figure 5 18 Figure 5 18. DBD Surface Plasma Sterilization, comparing air and nitrogen as the discharge gas Two sterilizat ion times (60s and 120s) are tested, using E, coli as the test pathogen. In Figure 5 18 above, in the case of air, complete bacterial inactivation is noted within 120s. In the case of N 2 starting from an initial concentration of 10 6 CFU at the end of 12 0s of plasma generation, only a 1 log 10 reduction in E. coli concentration is noted. However, the initial concentration of E. coli used was 10 8 CFU which implies that the flushing of the chamber four times as well as evacuation of the chamber to extremel y low pressure causes a 2 log 10 reduction in E. coli concentration. Nevertheless, even with a pre plasma concentration of 10 6 CFU it is evident that plasma generation in Nitrogen (hence, plasma generation in the absence of ozone) does not cause much of a reduction in E. coli concentration. Thus from the results of the Charcoal tests as well as Nitrogen tests, we conclude that ozone produced during plasma generation is capable of inactivation of E. coli on prolonged exposure times.
133 5.3 Temperature Studies Section 1.2 discusses the classification of laboratory plasmas into two categories: high temperature plasmas and low temperature plasmas. In high temperature plasmas, the high temperature serves to equilibrate the high temperature of the electrons, hence e stablishing local thermal equilibrium (LTE). But in low temperature plasmas, the heavy particles in plasma (ions, neutrals ) are at a much lower temperature than the electrons, which are typically high energy particles. Hence these plasmas are classified as non LTE and are typically used for low temperature applications, such as plasma sterilization. The effect of heating of the surface due to plasma generation and its subsequent contribution to sterilization has been examined  U sing a CO 2 plasma in th e low pressure regime they observed a higher destruction efficiency at 60 o C as compared to that at 15 o C A more fundamental analysis of the rotational and vibrational temperatures during plasma generation was conducted by Ayan et al. [ 108 ] who analyzed the relationship between rotational and vibrational temperatures and the plasma power density. They conclude that the gas temperature depends only on average power, which means that increasing the average power might increase the ave rage temperature, resulting in an increase in surface temperature of the treated tissue. However, they do not explicity evaluate the effect of surface temperature on sterilization. Hence for the DBD plasma setup used for sterilization in this thesis, it w as necessary to (a) u nderstand the kind of surface temperatures attained during plasma generation (b) e xamine the effect of temperature only on E. coli in order to determine the role temperature plays in the plasma sterilization process.
134 In order to do t his, an infrared camera (FLIR A320) wa s used to measure the temperature of the surface during plasma generation. FLIR A320 The infrared camera uses an uncooled micro bolometer to detect infrared radiation and converts it into an electronic signal, which is then processed to produce a thermal image that can be processed to obtain surface temperature. In order to obtain the thermographic image of each plasma device, while it was being operated, the plasma device was powered for 2 minutes, during which the rmographic images of the plasma device were obtained by the Infrared camera at a sampling rate of 0.5 Hz. After turning off the plasma device, the camera continued to record images for another 2 minutes, thus yielding a total of 48 frames. These images are transferred in real time to a computer, wherein they are subjected to additional data processing. Since the IR camera works by detecting the heat radiated off the surface, the temperatures obtained depend on the thermal emissivity coefficient, which need s to be calculated accurately. This is done by placing thermocouples on the device and heating up the device on a VWR Scientific hot plate. Data from the thermocouple is sampled via a LabView interface. The hot plate is allowed to heat up to a fixed known temperature, after which it is switched off and allowed to cool down. While it is cooling down, temperature is sampled via the thermocouple attached to the device on the hotplate and the IR camera simultaneously. Thus, the temperature detected by the thermocouple is the known, actual temperature (T a ) and the temperature detected by the IR camera is the detected temperature (T IR
135 via the software provided by the IR Camera (ExaminIR) using the Stefan Boltzmann law: ) ( 5 7 ) The LHS of this equation remains the same while the RHS varies for the temperature measured by the thermocouple versus the temperature measured by the IR camera. Hence, by a process of comparison, the emissivit y coefficient is calculated and applied to correct for T IR For FR4 and SC, this emission coefficient was calculated as 0.9097 + 0.03 and 0.929 + 0.03 respecti v ely Once the emissivity coefficients for temperature correction were obtained, a number of differe nt scenarios were tested in order to obtain temperature data. First a set of three different clean FR4 and SC devices wa s powered for a 2 minute interval, wherein temperature wa s sampled every 5s for each device. Secondly, a set of three different FR4 and SC devices, each inoculated with 20 l of E. coli were powered for a 2 minute interval, wherein temperature wa s sampled for each device again. For each of these devices, once plasma is turned off at 2 minutes, the device was allowed to rest for another 2 minutes. Temperature wa s sampled during these 2 minutes also. Figure 5 19 below shows two sets of images. (A) (D) is a depiction of the temperature of the electrode surface area at four different time s (t= 3 0s,60s,90s,120s). (E) (H) is a depiction of the same for an inoculated FR4 device. From Figure 5 19 it is observed that for a clean FR4 device, the surface temperature at 120s is marginally higher than that at 30s.The average surface temperature at 120s is observed to be ~16 o C higher than that at 30s. The highest temperatures are noted at the edge of the electrodes. The electrodes themselves are made of copper. Thermal conductivity of
136 copper (355 W/mK) is far greater than that of FR4 (0.25 W/mK) (Report on Thermal considerations for surface mount layout s, Texas Instruments). Hence the copper electrode conducts away heat at a much faster rate than the FR4 dielectric, which explains the cooler electrode temperatures. F or an inoculated device at 30s (5 19 (E)), the temperature is fairly low ( average surface temperature= 28 o C 34 o C ) due to the presence of the liquid bacterial sample on the surface. By 60s, this liquid has evaporated and temperature slowly starts increasing, reaching the maximum value at 120s. However this temperature profile does not resemble the one for a clean device at 120s The difference between average surface temperature at 30s and 120s for the inoculated case is calculated to be ~25 o C Figure 5 19 Variation of temperatures at t= 30s, 60s,90s, 120s for Clean FR4 device (A D) and In oculated FR4 device (E H)
137 Figure 5 20 below shows the same two sets of images for a clean and inoculated SC device respectively. Surface temperatures are considerably lower in th e case of SC, as compared to FR4. In the case of the inoculated device, the e ffect of liquid bacterial sample on the surface is demonstrated very clearly. In the case of the inoculated SC device, the liquid sample does not evaporate until the very end of the 2 mi nute interval. This is depicted in Figure 5 20 (E) (H), where surface temperature remains on the lower side even at 90s. The difference between average surface temperature at 30s and 120s for the clean and inoculated SC cases is calculated to be ~11 o C and 22 o C respectively. Figure 5 20 Variation of temperatures at t= 30s 60s,90s, 120s for Clean SC device (A D) and Inoculated SC device (E H) To get a better idea of the variation of temperature over time for the two configurations (clean and inoculated ), the average temperature over the entire electrode surface area at ev ery 5s, over a total of 2 minutes was calculated. Figure 5 2 1 below
138 shows this variation of average temperature for a clean and inoculated device (FR4 and SC). The variation of temperature thus shown in Figure 5 21 is a clearer graphical representation of the temperature trends noted in Figure 5 19 and Figure 5 20. Figure 5 2 1 Comparison of average surface temperatures during plasma generation for clean and inoculated FR4 and SC devices, measured using an infrared camera. A clean FR4 device has the high est average surface temperature. An inoculated FR4 device has very low surface temperatures initially. However as has been demonstrated in Figure 5 1 9 after the first 30s, this temperature rises rapidly. The rise in temperature is less steep than in the c ase of the inoculated SC device, as corroborated by the images in Figure 5 20. In order to determine the role of temperature in sterilization, a separate set of experiments was conducted wherein the devices were heated up to fixed temperatures, inoculated with 20 l of E. coli and allowed to rest on the hot plate for 2 minutes. The thermal conductivity of FR4 material used in this study is not readily available (although literature sources commonly cite a value of 0.25 0.3 W/m.K) while manufacturer
139 specifications c ite the thermal conductivity of SC as 0.71 W/mK Accordingly the SC devices were seen to heat up faster than the FR4 devices. Once the device reached the required temperature, the bacterial sample was deposited on it and the heating of the device was conti nued for another 2 minutes. Following this, each of the devices was subjected to standard post processing protocols and incubated for 48 hours. Colony counts were performed to estimate the number of surviving organisms Two types of temperatures were teste d using each dielectric. The average and maximum surface temperature s during plasma generation were measured both for FR4 and SC. Average and m aximum temperatures measured in the case of FR4 were ~54 o C and ~ 77 o C respectively. Average and maximum temperatur es measured in the case of SC were ~39 o C and ~ 72 o C respectively Thus in order to test the effect of temperature alone on E. coli the inoculated FR4 and SC devices were first heated up to their respective average temperatures The results of such a test, using the average temperature in each case, are given below in Figure 5 2 2 Figure 5 2 2 Sterilization plots analyzing the effect of temperature on inoculated FR4 and SC devices. In each case, the inoculated devices were heated up to the average temper ature measured during plasma generation.
140 Figure 5 23 below shows the results of the same test, using the maximum temperature in each case. For the sake of simplicity, results for both inoculated FR4 and SC devices are shown on the same plot. Figure 5 23 Sterilization plots analyzing the effect of temperature on inoculated FR4 and SC devices. In each case, the inoculated devices were heated up to the maximum temperature measured during plasma generation. I noculated FR4/SC devices were heated to average a nd maximum temperatures in each case so as to subject the bacterial concentration to surface temperatures measured during plasma generation. This was done in order to mimic the role of temperature during plasma generation. Of course, due to the design of t he hotplate, variability is introduced in terms of maintaining a constant temperature for the 2 minute interval. However care was taken to maintain the temperature within 2 o C 3 o C of the desired temperature. It must also be noted that during plasma generati on, the surface temperature is non uniform, which means that the whole surface is not at the same temperature at any given point during plasma generation. However, by heating up an inoculated device on a hot plate, the whole device is heated to a uniform t emperature. Hence, bacterial samples in these tests are actually being subjected to a much more
141 uniform (and hence greater) heating effect than they would be subject to during plasma generation. As is evident from Figure 5 2 2 and Figure 5 2 3 subjecting t he bacterial concentrations to even such uniform high temperatures does not seem to produce a significant reduction in bacterial concentrations. This implies that temperature does not play a major role in plasma sterilization. However it is possible that t emperature could be acting synergistically with another plasma component to influence the sterilization process. From the spectroscopic, ozone and temperature data discussed above, there seems to be a critical point during the sterilization cycle i.e. the complete evaporation of the liquid bacterial sample deposited on the dielectric surface. Once complete evaporation occurs, the spectroscopic intensity, emitted ozone levels and surface temperatures all increase significantly. Corresponding to this, a rapi d drop in bacterial concentrations, leading to complete inactivation, is also observed T he next section discusses the significance of this critical point more elaborately. 5. 4 Microbiological A nalysis The previous sections concentrated on characterizing th e DBD plasma being used for sterilization in this study and analyzing how the different plasma components might play a role in the mechanism of plasma sterilization. The current section takes a different approach in understanding this mechanism. The intera ction of plasma with cells is explored through a number of different microbiological techniques in order to obtain insight into which cell component might be affected by plasma exposure. These techniques include microscopic analysis of plasma treated cells fluorescence measurements etc.
142 5.4.1 Evaluation of M embrane D amage by Live/Dead B a cLight TM Assay The plate counts provide information about the number of viable microorganisms remaining on the sterilization device, but do not give information about the fate of individual cells. Ozone and other oxidative species target unsaturated lipids and lipopolysaccharides (LPS) in G cell membranes resulting in lethal damage and leakage of cell contents before complete destruction of vegetative cells is seen [ 109 110 ] T o examine membrane damage, LIVE/DEAD Technologies, Carlsbad, California) were used Th e s e kit s use a two color fluorescence assay to determine the integrity of bacterial membranes. The two dyes both stain nucl eic acids, but differ in their spectral range and ability to penetrate cells. SYTO 9, a green fluorescent dye that penetrates both damaged and intact cell membranes, and propidium iodide, a red fluorescent dye that cannot penetrate intact membranes and th us only stains cells with lethal membrane damage. The propidium iodine reduces the fluorescence seen from SYTO 9 when both are present. If cells are damaged to the point of lysis and thus lose their nucleic acids, they will not stain with either dye [ 52 70 ] These stained bacterial suspensions can then be observed under a f luorescence microscope or fluorescence can be measured in microtiter plates. In order to visualize the E. coli effect of plasma exposure on E. coli cells, different inoculated FR4 devices were plasma treated for t= 0,30,60,90 and 120s and sealed into ster ile bags containing 1 ml of phosphate buffered saline (PBS). Each device was agitated thoroughly for 20s, after which the supernatant from each bag was removed and subjected to a fluorescent staining protocol. F or each time point, 5 l of the stained
143 bact erial suspension, was then trapped between a slide and a coverslip and examined using a Leica DM4000 fluorescence microscope. Figure 5 2 4 Fluorescence Images obtained of the different cell suspensions after exposure to plasma for t= 0,30,60,90 and 120 s respectively. A E correspond to t= 0 120s respectively.
144 The excitation/emission maxima for the SYTO 9 stain and propidium iodide are 480/500 nm and 490/635 nm respectively. The corresponding set of images is given above in Figure 5 24 (A) (E). The diffe rent images A E correspond to t= 0,30,60,90 and 120s respectively. The fluorescent images confirm what is expected: at: no bacterial inactivation at 0s (the picture is filled with green mostly), negligible inactivation at 30s (minimal red flecks are observ ed), followed by partial inactivation at 60s and 90s (similar concentrations of red and green) and complete inactivation at 120s (all red on the slide) From Figure 5 2 4 very little concentration of cells is noticed on each individual fluorescence image. This is due to the fact that the initial concentrations of cells were very low, to start out with. Additionally the lag time between plasma treatment of cells and fluorescence imaging was 3 5 hours, owing to time taken for preparation protocols When cell suspensions are allowed to rest for such a long time, a portion of the cells tends to lyse, which is why lower cell concentrations are noted. Figure 5 2 5 An example of the calibration curve calculated by measuring Ratio G/R for cell suspensions with di fferent proportions of live/dead cells.
145 The ratio of live/dead cells can also be calculated using the measurement of fluorescent intensity of the cell suspension at each time point and correlating it to the percentage of live cells in the suspension (usin g a calibration curve like the one shown above in Figure 5 25). This calibration curve is used to determine the proportion of live/dead cells in the cell suspensions obtained after different plasma exposure times. This is accomplished via measurement of t he fluorescence in a micro plate reader. The same protocol as before is followed, except that the supernatant at each time point, after being treated with a fluorescent staining protocol, is subsequently mixed with filter sterilized dH 2 O. 100 l of this pr epared cell suspension is pipetted into separate wells of a 96 well flat bottom microplate. Three such sets of wells are prepared to account for standard deviation. With the excitation wavelength centered at about 485 nm, the fluorescence intensity at an emission wavelength corresponding to green (~530 nm ) is measured for each we ll of the entire plate. Similarly, using the same excitation wavelength, the fluorescence intensity at an emission wavelength corresponding to red ( R ~ 630 nm ) is measured for each well again. The ratio of intensities (Ratio G/R ) is derived by dividing the fluorescence intensity measured at G by the fluorescence intensity measured at R This ratio is then compared to existing calibration curves, in order to calculate the proportion of live bacteria. An example of this calibration curve is shown above in Figure 5 25 Thus, calculating Ratio G/R for the different cells suspensions obtained after plasma exposure times t= 0,30,60,90 and 120s, and plotti ng this Ratio G/R for the different time points, Figur e 5 2 6 below is obtained. In this figure, the mean of data from
146 all three replicates of microplate readings is shown. Error is plotted as standard deviation of the three replicates for each time point. F igure 5 26 shows similar phasic behavior, as observed in survival curves for E. coli (Figure 4 1). There exists a slight decrease in Ratio G/R at 30s (implying that a small population of the cells is dead), followed by a larger reduction in this ratio at 60 90 and 120s. It is also evident from this figure, that the majority of the cells are killed between 30 60s, thus corroborating the D values (time taken for 90% reduction in cell population) obtained in Figure 4 4 Since the cell suspensions used at diffe rent time points here were obtained after plasma exposure using inoculated FR4 devices, the below values of Ratio G/R make sense. If inoculated SC devices were used instead, a much less phasic behavior would be expected in Figure 5 26 with the major drop in Ratio G/R occurring between 100 120s. Figure 5 26 Plot of the Ratio G/R calculated after different plasma exposure times Comparing Figure 5 25 and 5 26 the percentage of live cells at 60, 90 and 120s seems to be ~0%. Notice that Ratio G/R at t= 0s is ~ 1, which roughly corresponds to 3 0% live bacteria in the cell suspension. As reiterated before, this is due to the relatively
147 lesser concentration of cells in the original cell suspensions used for fluorescence imaging. The presence of both red and green staining cells indicates that damage to cell membranes is contributing to overall cell mortality. If the cells were ruptured due to electrostatic tension or were completely etched away from the plasma treatment, then Ratio G/R would have been difficult to d etermine at t=60 120s since ruptured cells cannot be stained by either dye Thus it is likely that the cells treated by plasma are inactivated due to damage to the cell membrane or one of the cell organelles rather than cell rupture by plasma treatment. Further analysis in this direction would require SEM imaging of the plasma treated cells 5.4.2 Mutation Studies Section 5.1 demonstrates that the spectral signature of the generated DBD plasma does not contain any lethal UV radiation, specifically UV C ra diation (Figure 1 10). UV C radiation is especially important as it can cause dimerization of DNA strands in cells, thus affecting base pairings and causing mutations during DNA replications. The objective of the mutation studies was to definitively determ ine whether damage due to UV C radiation is a factor during DBD p lasma s terilization. Rifampin is a bactericidal antibiotic drug of the rifamycin group. Media containing Rifampin is known to select for the growth of rifampin resistant (Rif (r)) mutants [ 111 ] Rifampin resistant cells arise at about 1 in 10 14 cells from naturally occurring point mutations DNA damage due to UV light exposure increases the rate at which resistance muta t ions occur. Hence E. coli cells treated with plasma for the sterilization times cited in Figure 4 1 were cultured both on regular LB agar as well as agar containing Rifampin The expectation was that if the E. coli cells were being damaged
148 by UV C radiation during plasma generation, then a higher percentage of Rifampin resistant colonies would be seen on the Rifampin plates, as compared to unexposed cultures However this was not observed. Instead the CFU concentration on both sets of plates appeared to be similar. This proved that damage to the E. coli cells was not due to DNA damage, which in turn proved that shortwave UV radiation was likely not responsible for cell death d ue to DBD plasma exposure. 5.4.3 Microscopic Analysis of P lasma I nteraction with B. subtilis B iofilms A biofilm is an architecturally complex community of bacterial cells. Biofilms harbor multiple cell types i.e. in the same biofilm, individual cells can f ollow different developmental pathways, resulting in heterogeneous populations. Hence one group of cells can perform one distinct function and localize to a separate region within the biofilm, while another can perform another distinct function and localiz e to a different regions. This property of biofilms makes them very popular for laboratory research. Research in plasma interaction with biofilms picked up pace, as recently as 2007 onwards ( discussed further in Section 1.3.3) The differentiation in cell function can be highlighted by the use of different fluorescence proteins [ 112 ] Two different cell phenotypes are monitor ed for changes producing cell phenotype. Changes in the motile phenotype are determined by monitoring the expression of the hag operon, which encodes flagellin, a major protein component of flagella (responsible for locomotion in cells). Changes in the matrix producing phenotype are determined by monitoring the expression of the tapA operon, which encodes the primary protein component of the extra cellular matrix [ 93 94 113 ] A dual reporter strain for both of these genes was use d in which genes en coding
149 fluorescent proteins are fused to promoters of the tapA and hag genes. A tapA YFP (yello fluorescent protein) fluorescence reporter construct was integrated into the amy E locus and a hag CFP (cyan fluorescent protein) was integra ted into the lacA operon. The objective of microscopically examining the biofilm exposed to plasma was to determine which cell phenotype is affected due to plasma exposure: the motile cells or the matrix producing cells. Since fluorescent reporter strains of B. subtilis NCIB360 were readily available, these were used for the microscopic analysis. B. subtilis c ells were stored at 80 o C and streaked directly from frozen stocks to LB agar. After 12 hours of incubation at 37 o C, 3 ml of LB liquid medium was in oculated with cells from an isolated colony. This inoculated medium was incubated in a shaker at 37 o C for 3 hours, until OD was approximately 1. A 3 l drop of this culture was then deposited at the center of a 60 mm MSgg (minimal salts, glycerol glutamate ) agar plate and incubated at 37 o C for 24 48 hours. This led to the formation of a biofilm. The biofilm thus grown was transferred to an inverted microscope system ( Nikon Eclipse Ti) for brightfield and fluorescence imaging. Then, a plasma device was inver ted (comb like electrode inverted face down) and plasma generated at 40 50 kHz, 9 10 kV pp for 2 minutes. Both the input frequency and voltage varied because of the additional impedance introduced due to the device resting face down on the surface of the M Sgg agar. However care was taken to ensure that the frequency and input voltage remained within the stated limits. The biofilm was imaged before, during and after plasma exposure using a 4x objective. Images were taken along the equatorial slice of the ag ar plate and are shown below in Figure 5 27. The imaging order is (A) before plasma exposure (B) immediately
150 after plasma exposure for 1 minute (C) before Plasma exposure for the 2 nd time (D) immediately after plasma exposure for 1 minute (2 nd time) (E) 50 minutes after plasma exposure for 1 minute (2 nd time). Figure 5 27 Images of bio films before, during and after plasma exposure. CFP indicates motile cells and YFP indicates matrix producing cells. A E indicates the imaging order of the biofilms. Thu s the B. subtilis biofilm was exposed to plasma for 2 minutes in 1 minute intervals. As is observed from Figure 5 27 the CFP images do not show a significant variation in intensity, whereas the YFP show a gradual decrease in intensity from A E, signaling cell injury/death. When this mean intensity variation from A E is plotted graphically both for YFP and CFP channels Figure 5 28 below is obtained. In Figure 5 28, the % reduction in mean intensity values in frames B E, as compared to frame A is plotted b oth for CFP and YFP modes. As is evident, there is a drastic reduction in intensity values for the YFP mode, which implies further that the matrix producing cells are more damaged than the motile cells after 2 minutes of plasma exposure.
151 Figure 5 28 Mean intensity variation in CFP and YFP modes for imaging order A E. On the Y axis is plotted the % reduction in intensity ( ve because of the reduction). Points 1 5 correspond to frames A E. This result also matches the results seen with the B. subtili s sterilization experiments, describe in Figure 4 2 DBD plasma is observed to be more lethal to the B. subtilis strain cultured in MSgg media (complete inactivation in 4 minutes) rather than the one cultured in LB media (incomplete inactivation in 6 minu tes). MSgg media promotes the growth of the matrix producing phenotype while LB media promotes the growth of the motile phenotype Hence b oth the sterilization tests and the microscopic biofilm analysis confirm that plasma exposure is more lethal to cells expressing the matrix forming phenotype. This is an important insight in to determining the cell breakdown mechanism due to plasma exposure. 5.5 Discussion In this chapter, many different features of DBD plasma generation are discussed in an effort to unde rstand their role in plasma sterilization. The different features
152 discussed are the spectroscopic signature, ozone levels and surface temperatures measured during plasma generation. Additionally, an insight into the cell state after plasma exposure is also gained via microbiological studies. Spectroscopic data obtained highlights intensity peaks at wavelengths corresponding to transitions in the 2 nd positive system of N 2 No intensity peaks are making it unlikely that UV C radiation plays a major role in DBD plasma sterilization at atmospheric pressure. This observation is further confirmed by the Rifampin studies which do not show an increase in mutation rates The levels of ozone produced during DBD plasma generation and the dependence of the decay of this produced ozone on volume of the sterilization chamber is discussed in detail in Section 5.2.1. Such a comprehensive dataset helps understand and predict the rates of produced ozone and the time taken for the produced ozone to dissipate using a surface DBD plasma device. Ozone exposure tests, in which an initial concentration of E. coli is exposed to ozone produced during plasma generation, dem onstrate that this initial concentration is almost completely inactivated after 7 minutes of ozone exposure (in the case of an inoculated FR4/SC substrate). Additionally, it is determined that a threshold value (120 150 ppm) of ozone concentration is requi red for ~99.99% reduction in bacterial concentrations Activated charcoal is used to inhibit ozone production. Reducing E. coli concentrations to these reduced ozone concentrations leads to negligible reduction in bacterial concentrations (Figure 5 17). T his implies that the reduction in ozone levels due to activated charco al is what causes the reduced degree of lethality due to ozone
153 exposure. However the protocol for testing ozone exposure consists of inoculating a clean plasma device with E. coli and pl acing it next to a clean plasma device (which acts as the plasma generator). Thus while the latter is being powered, the former is being exposed to the plethora of reactive spec ies being produced It is not being exposed to the direct action of the plasma, as described before, contains a number of reactive species, especially reactive oxygen species like ozone (O 3 ), O 2 *(a 1 g ), O( 1 D), O( 3 P) It is also possible that the activated charcoal responsible for inhibit ion of ozone production is also responsible for the inhibition of other reactive species. However, keeping in mind that the average residence times for most of these reactive species is on the order of seconds and that m the plasma generator, it is considered highly likely that bacterial inactivation in such a case is due to the produced ozone. In order to further explore this theory, plasma is also generated in a pure nitrogen (N 2 ) atmosphere. A smaller vacuum chamber i s used as the sterilization chamber and before each experiment, the chamber is flushed with nitrogen sufficiently so as to maintain a ~99.98% N 2 environment. Survival curves obtained using such a setup (using a protocol similar to that used for Figure 4 1 ) show that plasma exposure in a Nitrogen (N 2 ) environment does not produce significant bacterial inactivation (Figure 5 18). This further supplements the idea that ozone produced during DBD plasma generation is likely responsible for bacterial inactivatio n. Of course, the way to confirm this would be to use additional reagents designed to detect reactive oxygen species in bacterial cell suspensions [ 114 ]
154 Temperature studies, conducted using an FLIR A320 IR camera are useful in determining the surface temperature distribution for the dielectric substrate during plasma generation. It is noted that surface temperatures are greater in the case of FR4 compared to SC. Additionally, experiments are also conducted in which the bacterial concentrations were heated to surface temperatures, similar to those observed during plasma generation. Survival curves obtai ned from these experiments demonstrated that pathogen inactivation due to effect of temperature alone was negligible (Figure 5 22 and 5 23) The physiological state of cells after plasma exposure is examined using a LIVE/DEAD bacterial viability kits, whic h helps visualize the proportion of live/dead cells after different intervals of plasma exposure. The proportion of the live/dead cells measured corresponds to the inactivation rates observed in Figure 4 1 This information also helps conjecture that a pos sible mechanism of plasma sterilization is damage to the cell membrane rather than rupture of the cells due to electrostatic tension. Finally the effect of liquid on the dielectric surface is examined. The evaporation of the liquid E. coli sample deposite d upon the device surface follows a pattern. Initially the bacterial sample deposited covers the entire electrode surface area, and plasma is visible only around the edges of the electrode. As time progresses, the sample begins to evaporate around the oute r edges of the electrode. Gradually, this evaporation begins to spread to other parts of t he electrode, until eventually plasma covers the entire electrode surface area. This usually occurs at around t= 6 0s for the FR4 dielectric and at t= 120s for the sem i ceramic (SC) dielectric. These two times are also the times
155 at which a steep drop in pathogen concentration is noted from the sterilization plots (Figure 4 1 ). Spectroscopic, ozone, temperature and power data uniformly show that plasma is repressed whil e visible bacterial sample is present on the plasma devices. Absorbed input power plots (Figure 4 13) shows the correspondence of the temporal variation of absorbed input power to the amount of bacterial sample on the plasma device. The spectroscopic inten sity peaks are noted at the same wavelengths at each time point; however, their intensities increase as the liquid evaporates (Figure 5 3 and 5 4). Similarly, it was observed that as the liquid sample evaporated, rate o f production of ozone remained low u ntil the entire liquid sample evaporated. Thereafter ozone production increases. Similarly surface temperature plots show the sudden increase in temperature once the entire sample evaporates (Figure 5 19 5 21). When the same number of organisms was deposi ted in a 40l volume instead of loss of viability, that was noted in Figure 4 1 was extended by about 30s in Figure 4 5 i.e. the rapid drop in E. coli concentration occ urs at t> 120s, as opposed to 90s in the case of the lower inoculation volume (20 l). This further promotes the theory that the amount of liquid bacterial sample deposited on the plasma device affects the sterilization time. Oehmigen et al. [ 115 ] r eported experiments wherein they examined the role of acidification in influencing antimicrobial activity due to DBD plasma exposure. They concluded that plasma treatment of non buffered liquids by indirect surface DBD results in acidification and thus, inactivation of suspended bacteria. When they tested the same
156 theory with buffered solutions, they noted that pH decrease is avoided and therefore, antimicrobial plasma activity is reduced. It is suggested that react ive species from the plasma generation are the cause of liquid acidification and bactericidal activity. E. coli and plasma ghly rinsed with 1 ml of Type 1 (ultrapure) Milli Q water. For each sample, the pH of the corresponding volume of water was measured using an Accumet AB 15 pH meter, which has an accuracy of + 0.01. This process was repeated for both FR4 and SC dielectr ic devices. Before measuring the pH, the meter was standardized using pH buffer solution. The pH of LB broth used to make the E. coli sample was measured as 7.16 and that of the E. coli sample itself was measured to be 6.77. The variation o f pH is given b elow in Figure 5 2 9 Figure 5 2 9 Plot of pH values, obtained by rinsing devices with Millipore water after plasma generation and measuring the pH value of this water in each case. Bo th FR4 and SC dielectrics are compared. Figure 5 29 indicates that the reduction of pH is greater in the case of FR4 as compared to SC. However unlike the strong reduction in pH values noted by Oehmigen et al. [ 115 ] there is not a strong pH change in our results (both FR4 and SC). Thus it is
157 most likely that rapid acidification plays some role but not a major one in bacterial cell death. Note that the pH value does not vary mu ch, except during the last 30s (for FR4) and not at all for SC. This is again indicative of the effect of the liquid on the dielectric surface. Since the liquid bacterial sample deposited on the dielectric substrate does not evaporate until the very end of the sterilization time interval (for both FR4 and SC) the pH does not change very much until the very end. This confirms that the liquid deposited on the dielectric substrate inhibits plasma generation Hence, the point at which all the liquid covering t he electrode evaporates and plasma covers the entire electrode surface area is the point at which there is a steep rise in input power, emitted ozone levels and spectroscopic intensity. It is also at this point, where the steep drop in pathogen concentrati on occurs, thus indicating that th ere is a threshold time point after which complete sterilization occurs Dobrynin et al. [ 63 ] explore the plasma dosage required for bacterial inactivation in cases with and without water. Their results show that the plasma dosage required for complete bacterial inactivation in cases with water is lower than that re quired for cases without water. They also conclude that the presence of water and direct plasma are both required to achieve fast inactivation and that this inactivation is highly dependent on the amount of water. In the present cases, it is found that as long as the liquid bacterial sample is present on the surface of the plasma device, plasma generation and therefore, sterilization is hampered. One way to explain this adverse effect is in terms of capacitance. I f the FR4/SC plasma device is considered a s a capacitor of capacitance (C 2 ), then the liquid layer
158 spread uniformly on top of the device can be considered as a second capacitor of capacitance (C 1 2 Thus the combined capacitance of this system would be (5 8) The impedance Z 1 of the liquid layer varies inversely with the amount of liquid present on the surface of the device i.e. as the liquid evaporates, impedance Z 1 decreases. Since Z 1 is inversely proportional to capacitance, C 1 (Z = it follows that as Z 1 decreases C 1 increases Following this, as C 1 increases the capacitance of the overall system (C) increases and thus, the energy stored in the system ( U= increases, proving that the amount of the liquid on the surface of the plasma device is actually detrimental to the performance of the plasma device as a sterilizer. This is mirrored in Figure 4 5 i.e. more the inoculation volume, more the amount of liqu id covering the electrode surface and hence more the sterilization time.
159 CHAPTER 6 CONCLUDING REMARKS AND RECOMMENDATIONS FOR FUTURE WORK Using a FR4 plasma device to generate DBD surface plasma it is determined that a sterilization time of 2 3 minut es i s required for complete inactivation of vegetative pathogens ( E. coli Yeast) Sterilization tests with G. stearothermophilus spores require d a sterilization time of 20 minutes for complete inactivation [ 116 ] Two types of dielectric materials are investigated: FR 4 (dielectric constant= 4. 7 ) and semi ceramic (dielectric constant = 3. 00 ). B oth kinds of substrate materials have not been investigated before in terms of plasma sterilization. It is observed that sterilization times are shorter and more reliable in the case of FR4 as compared to semi ceramic. Diagnostic data measured during plasma generation in both cases (spectral signature, emitted ozone levels, surface temperature, absorbed RF input power) also highlight ed FR4 as the superior dielectric, when compared to SC. In analyzing the dependence of sterilization on input power, as the input voltage is increased (i.e. input power increases), the sterilization beha vior becomes more and more consistent with complete sterilization being obtained at input voltages of 12 and 14 kV pp at an input frequency of 14 kHz. As the input frequency is increased to 60 kHz, it is observed that a lower input voltage of 10 kV pp is s ufficient to bring about complete sterilization in the case of FR4, but not in the case of SC plasma devices. In order to better compare the sterilization performance of devices of different dielectric materials, at different input voltages and frequencie s, an intrinsic parameter called plasma dose [ 20 ] is calculated For a given input voltage and frequency, for a given dielectric material, the plasma dosage can be used to determine the plasma treatment time required for complete sterilization.
160 Lastly, in terms of paramet er analysis, the effect of reduced pressures on sterilization effectiveness is also examined. Sterilization effectiveness is examined at P= 400,500 and 600 Torr, for plasma exposure times of 1 and 2 minutes. The sterilization effectiveness at the reduced p ressures is found to be similar to that at atmospheric pressure. Spectroscopic studies show ed that the spectral pattern characteristic of the generated DBD plasma shows intensity peaks at wavelengths characteristic of the 2nd positive system of N 2 FR4 and SC plasma devices show ed intensity peaks at same wavelengths, although they differ in intensity values shown at each wavelength. Since ar e observed, it is concluded that UV irradiation is likely not a mechanism in su rface plasma sterilization at atmospheric pressure. This is also confirmed by the Rifampin studies. The dissipation rates of ozone, during and after plasma generation using the different FR4/SC substrates as well as using sterilization chambers of differe nt sterilization volumes, are determined. It is observed that the FR4 substrate used produces greater levels of ozone as compared to the SC substrate, during plasma generation. It is also observed that the chamber volume determines the dissipation of produ ced ozone inside the chamber Ozone exposure tests, in which an initial concentration of E. coli is exposed to ozone produced during plasma generation, demonstrate d that this initial concentration is ~99.99% inactivated after 7 minutes of ozone exposure ( in the case of an inoculated FR4/SC substrate). The required levels of ozone for this inactivation are determined to be 120 150 ppm.
161 In order to prove that bacterial inactivation on exposure to air excited by plasma generation is due to the produced ozone two independent tests are conducted. Activated charcoal is used to adsorb the produced ozone, following which an exposure test demonstrated no significant bacterial inactivation Also plasma treatment of inoculated devices in a nitrogen rich environment also produced negligible reduction in bacterial concentration, which proved that when plasma was generated in a nitrogen rich environment (i.e. an environment low in ozone), sterilization efficiency was significantly reduced. Measurement of surface temper atures during plasma generation demonstrates average non uniform surface temperatures of 320 340 K in the case of FR4 and 310 315 K in the case of SC. Sterilization tests in which bacterial concentrations are heated to these observed surface temperatures, using a hot plate, showed that the effect of temperature alone was not enough for complete bacterial inactivation. Bacterial samples deposited on the dielectric surface and exposed to plasma possess a liquid component due to the culture broth used for bac terial growth. It is noted that as plasma is generated, the liquid sample slowly evaporates. The point at which the liquid completely evaporates has been observed to be the point at which an increase in spectroscopic intensity, ozone concentration, surface temperatures and absorbed input power is noticed. It is also at this point that a steep drop in pathogen concentration is observed. Thus it appears that the presence of liquid on the surface also determines the sterilization efficiency. Microscopic analy sis of B. subtilis biofilms exposed to plasma demonstrated that plasma exposure affected one particular cell function (matrix production) as compared
162 to another cell function (motility). This is important to know in order to determine the mechanism of cell death due to plasma exposure. On that note, fluorescence measurements also demonstrate that cell death is likely not due to rupturing of the cells due to electrostatic tension caused by charge accumulation on the cell membranes. In understanding the mech anism of plasma sterilization, different factors have been examined and tested separately to determine their influence on sterilization. While the spectroscopic data implies that UV radiation and photon intensity are not major factors in plasma sterilizati on, it is prudent to exercise caution before eliminating their roles. It is possible that the photons emitted during plasma generation might be etching the bacterial species (photo desorption), though probably in a minor capacity. Similarly, ozone exposur e tests indicate that exposure to highest concentrations of generated ozone for 2 minutes produces a ~4 log 10 reduction in bacterial concentrations. The lethal effect of ozone has been proven. Considering that the bacterial sample resting on the surface ab sorbs most of the ozone produced initially during plasma generation, it is quite possible that ozone plays a pivotal role in plasma sterilization. From the temperature studies, it is noted that heating an inoculated plasma device to an average surface tem perature (noted during plasma generation) produces only 1 2 log 10 reduction in bacterial concentrations Hence temperature and ozone tests point to the fact that surface temperature and ozone might be playing a synergistic role in the sterilization process The hypothesis of the synergistic role of ozone and heat in the process of surface plasma sterilization is drawn cautiously with the caveat that the roles of other reactive
163 oxygen species in the sterilization process have not been tested. It is also pos sible that for a plasma device inoculated with a certain amount of bacterial sample, when plasma is generated, the sample itself might absorb some of the other reactive oxygen species produced during plasma generation (superoxide ions, oxide ions etc.) whi ch could further damage the bacterial cells. However considering the short residence times of these ions and the more lethal effect of ozone (as shown by the ozone exposure tests), it is conjectured that while additional reactive oxygen species might play a minor role in the sterilization process, ozone plays a key role. The role of the other reactive oxygen species needs to be further analyzed with the help of a number of methods elaborated upon in Section 6.2. The volume of research discussed in this the sis has provided a substantial amount of information about the processes of surface plasma sterilization. It has also raised an additional number of interesting research avenues. 6.1 Scope of Technology Work outlined in this study has already proved the e ffectiveness of DBD surface plasma in sterilizing pathogens. O nly single plasma devices were tested. In order to further the scope of this technology, sterilization testing has to be conducted using larger sterilization arrays. These arrays cou ld be embedded with a grid like arrangement of the current plasma devices. In testing these arrays, the inherent problem to be dealt with will be the huge amount of ozone being produced during plasma generation. While a single plasma device can produce 120 150 ppm of ozone during a 2 minute interval in an enclosed space, a larger array made up of several of these devices will definitely produce more ozone. Thus the volume of the chamber in which such an array is enclosed will be important.
164 Additionally, in testing for sterilization, the entire array will be inoculated with the bacterial sample. The devices themselves will be arranged on the array in a grid like fashion, with a small gap between each row and each column of devices. During plasma generation, this gap will be inoculated, but not covered by plasma. This poses a problem in terms of sterilization and should be kept in mind, while designing the sterilization array. In fact, work is currently underway in building such an array and testing it for s terilization purposes. 6.2 Further analysis of the Mechanism of P lasma S terilization The microbiological information gained has highlighted that plasma exposure affects one cell function primarily. It has also highlighted a possible cause of cell death du e to plasma exposure. In order to view the damage to the cells up front, SEM analysis of the cells exposed to plasma is required. Additionally, the direction of research afforded by plasma interaction with biofilms, wherein it has been observed that plasm a affects a particular type of cell function, seems to be a very interesting direction of research to pursue In this aspect, knockouts (cell types with a particular gene subtracted) can be studied further to understand whether the lack of a particular gen exposure. On the plasma side, an important research avenue to pursue would be to determine the rotational, vibrational and translational temperatures attained during plasma generation. These temperatures can be simu lated using spectroscopic information or measuring using a more rigorous optical emission spectroscopy (OES) setup. Measurement of the rotational and vibrational temperatures during plasma
165 generation helps determine the energy densities of the different sp ecies produced in plasma generation, thus providing further insight into the sterilization mechanism. Additionally, in this study, the role of ozone in bacterial inactivation has been discussed extensively. The effect of exposing bacterial concentrations to different ozone concentrations produced during plasma generation has been discussed However owing to the highly coupled nature of the problem, it is a little harder to understand the role of ozone in the process of plasma sterilization itself. Howeve r, plasma generation also produces other reactive oxygen species (O, O O 2 + ). Due to the short reaction times of these species, it is much more difficult to detect these and analyze their role in plasma sterilization. However certain assays do exist that detect the oxidation of certain proteins that are integral to the cell structure [ 114 ] U sing these as says, it is possible to analyze the role of other reactive species in the process of plasma sterilization. 6.3 Numerical M odeling in P lasma S terilization Abundant literature exists in the domain of numerical modeling of dielectric barrier discharge (DBD) plasma. The concentration of different chemical species emitted in a plasma discharge can be numerically modeled using chemical reactions and plasma transport parameters [ 106 ] There also exists an abundance of literature in the modeling of cell breakdown the microbiological parameters involved, different mechanisms that can cause breakdown, threshold voltages that can irreversibly damage a cell. What is req uired is a coupling between the two problems: a single numerical model that simulates the penetration of plasma species into a bacterial cell, causing cellular breakdown. This numerical model should be able to map the cellular breakdown also, via chemical destruction of integral components of the bacterial cell or electrophoretic rupture of the cell wall.
166 For a numerical model resembling the structure of the cell wall, Dobrynin et al. [ 63 ] suggest the lipid peroxidation model as a valid model. Given below, in Figure 6 1 is a skeleton of chemical reactions that make up the lipid peroxidation model. Figure 6 1. Scheme of reactions used for Lipid Peroxidation [ 117 ] This scheme consists of seven coupled reactions. Reaction#1 is not used in the model, but is more of a starter reac tion that sparks off the other reactions. Note that that coupling factor between this model and plasma air chemistry is oxygen. More complex lipid peroxidation models exist, that employ reactions with other reactive oxygen species, which are again common t o plasma air chemistry models. Thus, a combination of such a model as shown in Figure 6 1 and a numerical plasma air chemistry model would have numerous coupled chemical reactions, all
167 happening at the same instant, thus adding to computational complexit y and time. The ionization rate is definitely dependent on other plasma parameters such as driving voltage, driving frequency, biological species being tested, electrode geometry etc. There is a huge dependency on biological species being tested because di fferent species take different times for complete inactivation to be achieved. However this is estimated by the factor which is the sterilization time. Thus numerically simulating such a model poses a hefty challenge. In an earlier work [ 118 ] an attempt was made to correlate the rate of ionization during plasma generation to the sterilization pattern noted in a stamp test. However more work is needed in trying to underst and the relation between the rate of ionization during plasma generation and its effect on influencing the sterilization time. The road ahead is full of interesting questions to answer. With further research, the day is not too far off, when surface plasm a sterilization will be implemented as a viable technology, promising safe, portable and fast sterilization.
168 REFERENCES 1. Tietjen L, Cronin W, McIntosh N (1992) Infection prevention for family planning service programs. JHPIEGO, Bal timore, MD. 2. Setlow P (2006) Spores of Bacillus subtilis: their resistance to and killing by radiation, heat and chemicals. Journal of applied microbiology 101: 514 525. 3. Russell A (1990) Bacterial spores and chemical sporicidal agents. Clinical microb iology reviews 3: 99 119. 4. Walker JT, Dickinson J, Sutton J, Raven N, Marsh P (2007) Cleanability of dental instruments implications of residual protein and risks from Creutzfeldt Jakob disease. British dental journal 203: 395 401. 5. Lee C H, Montville TJ, Sinskey AJ (1979) Comparison of the efficacy of steam sterilization indicators. Applied and Environmental Microbiology 37: 1113 1117. 6. Palenik C, Adams M, Miller C (1990) Effectiveness of steam autoclaving on the contents of sharps containers. Americ an journal of dentistry 3: 239. 7. Prolo DJ, Pedrotti PW, White DH (1980) Ethylene oxide sterilization of bone, dura mater, and fascia lata for human transplantation. Neurosurgery 6: 529 539. 8. Spaulding EH, Emmons EK, Guzara ML (1958) Ethylene oxide ster ilization. AJN The American Journal of Nursing 58: 1530 1531. 9. Borneff Lipp M, Okpara J, Bodendorf M, Sonntag H (1997) Validation of low temperature plasma (LPT) sterilization systems: Comparison of two technical versions, the Sterrad 100, 1.8 and the 10 0S. Hygiene und Mikrobiologie 3: 21 28. 10. Jacobs P, Kowatsch R (1993) Sterrad Sterilization System: a new technology for instrument sterilization. Endoscopic surgery and allied technologies 1: 57. 11. Jacobs PT, Lin S M (1987) Hydrogen peroxide plasma st erilization system. U.S. Patent No. 4,643,876 12. da Silva Aquino KA Sterilization by Gamma Irradiation. 13. Kowalski W, Bahnfleth WP (2000) UVGI design basics for air and surface disinfection. Heating/piping/air conditioning engineering 72: 100 110. 14. L erouge S, Wertheimer M, L'H Y (2001) Plasma sterilization: a review of parameters, mechanisms, and limitations. Plasmas and Polymers 6: 175 188. 15. Booth A (1999) Sterilization of medical devices: CRC Press.
169 16. Moisan M, Barbeau J, Moreau S, Pelletier J, Tabrizian M, et al. (2001) Low temperature sterilization using gas plasmas: a review of the experiments and an analysis of the inactivation mechanisms. International journal of Pharmaceutics 226: 1 21. 17. Laroussi M (1996) Sterilization of contaminated m atter with an atmospheric pressure plasma. Plasma Science, IEEE Transactions on 24: 1188 1191. 18. Koulik P, Begounov S, Goloviatinskii S (1999) Atmospheric plasma sterilization and deodorization of dielectric surfaces. Plasma chemistry and plasma processi ng 19: 311 326. 19. Kelly Wintenberg K, Hodge A, Montie T, Deleanu L, Sherman D, et al. (1999) Use of a one atmosphere uniform glow discharge plasma to kill a broad spectrum of microorganisms. Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films 17: 1539 1544. 20. Kalghatgi S, Kelly CM, Cerchar E, Torabi B, Alekseev O, et al. (2011) Effects of non thermal plasma on mammalian cells. PLoS One 6: e16270. 21. Fridman A (2008) Plasma chemistry: Cambridge University Press. 22. Bogaerts A, Neyts E Gijbels R, van der Mullen J (2002) Gas discharge plasmas and their applications. Spectrochimica Acta Part B: Atomic Spectroscopy 57: 609 658. 23. Conrads H, Schmidt M (2000) Plasma generation and plasma sources. Plasma Sources Science and Technology 9: 4 41. 24. Kogelschatz U (2003) Dielectric barrier discharges: their history, discharge physics, and industrial applications. Plasma chemistry and plasma processing 23: 1 46. 25. Gaunt LF, Beggs CB, Georghiou GE (2006) Bactericidal action of the reactive spec ies produced by gas discharge nonthermal plasma at atmospheric pressure: a review. Plasma Science, IEEE Transactions on 34: 1257 1269. 26. Haverkamp RG, Miller B, Free K (2002) Ozone production in a high frequency dielectric barrier discharge generator. Oz one: science & engineering 24: 321 328. 27. Chu PK, Chen J, Wang L, Huang N (2002) Plasma surface modification of biomaterials. Materials Science and Engineering: R: Reports 36: 143 206. 28. Shiloh J, Rosenberg A, Wurzberg E (2001) Modular dielectric barri er discharge device for pollution abatement. U.S. Patent No. 6,245,299
170 29. Kalghatgi SU, Fridman G, Fridman A, Friedman G, Clyne AM. Non thermal dielectric barrier discharge plasma treatment of endothelial cells; 2008. IEEE. pp. 3578 3581. 30. Fridman G, Friedman G, Gutsol A, Shekhter AB, Vasilets VN, et al. (2008) Applied plasma medicine. Plasma Processes and Polymers 5: 503 533. 31. Fridman G, Peddinghaus M, Balasubramanian M, Ayan H, Fridman A, et al. (2006) Blood coagulation and living tissue steriliza tion by floating electrode dielectric barrier discharge in air. Plasma Chemistry and Plasma Processing 26: 425 442. 32. Fridman G, Shereshevsky A, Jost MM, Brooks AD, Fridman A, et al. (2007) Floating electrode dielectric barrier discharge plasma in air pr omoting apoptotic behavior in melanoma skin cancer cell lines. Plasma Chemistry and Plasma Processing 27: 163 176. 33. Cooper M, Fridman G, Staack D, Gutsol AF, Vasilets VN, et al. (2009) Decontamination of surfaces from extremophile organisms using nonthe rmal atmospheric pressure plasmas. Plasma Science, IEEE Transactions on 37: 866 871. 34. Klmpfl TG, Isbary G, Shimizu T, Li Y F, Zimmermann JL, et al. (2012) Cold atmospheric air plasma sterilization against spores and other microorganisms of clinical int erest. Applied and environmental microbiology 78: 5077 5082. 35. Heinlin J, Morfill G, Landthaler M, Stolz W, Isbary G, et al. (2010) Plasma medicine: possible applications in dermatology. JDDG: Journal der Deutschen Dermatologischen Gesellschaft 8: 968 97 6. 36. Jiang C, Chen MT, Gorur A, Schaudinn C, Jaramillo DE, et al. (2009) Nanosecond pulsed plasma dental probe. Plasma Processes and Polymers 6: 479 483. 37. Sladek RE, Stoffels E, Walraven R, Tielbeek PJ, Koolhoven RA (2004) Plasma treatment of dental c avities: a feasibility study. Plasma Science, IEEE Transactions on 32: 1540 1543. 38. Menashi WP (1968) MENASHI TREATMENT OF SURFACES. U.S. Patent No. 3,383,163 39. Boucher RMG (1980) Seeded gas plasma sterilization method. U.S. Patent No. 4,207,286 40. Bithell RM (1982) Package and sterilizing process for same. U.S. Patent No.4,321,232
171 41. Ratner BD, Chilkoti A, Lopez GP (1990) Plasma deposition and treatment for biomaterial applications. Plasma deposition, treatment and etching of polymers: 463 516. 42 Baier R, Carter J, Sorensen S, Meyer A, McGowan B, et al. (1991) Radiofrequency gas plasma (glow discharge) disinfection of dental operative instruments, including handpieces. The Journal of oral implantology 18: 236 242. 43. Park BJ, Lee D, Park J C, Le e I S, Lee K Y, et al. (2003) Sterilization using a microwave induced argon plasma system at atmospheric pressure. Physics of Plasmas 10: 4539. 44. Trompeter F J, Neff WJ, Franken O, Heise M, Neiger M, et al. (2002) Reduction of Bacillus subtilis and Asper gillus niger spores using nonthermal atmospheric gas discharges. Plasma Science, IEEE Transactions on 30: 1416 1423. 45. Lee K Y, Joo Park B, Hee Lee D, Lee I S, O Hyun S, et al. (2005) Sterilization of Escherichia coli and MRSA using microwave induced arg on plasma at atmospheric pressure. Surface and Coatings Technology 193: 35 38. 46. Nelson CL, Berger TJ (1989) Inactivation of microorganisms by oxygen gas plasma. Current Microbiology 18: 275 276. 47. Roth JR, Tsai PP, Liu C (1995) Steady state, glow disc harge plasma. U.S. Patent No.5,387,842 48. Yu H, Xiu ZL, Ren CS, Zhang JL, Wang DZ, et al. (2005) Inactivation of yeast by dielectric barrier discharge (DBD) plasma in helium at atmospheric pressure. Plasma Science, IEEE Transactions on 33: 1405 1409. 49. Sun Y, Qiu Y, Nie A, Wang X (2007) Experimental research on inactivation of bacteria by using dielectric barrier discharge. Plasma Science, IEEE Transactions on 35: 1496 1500. 50. Tanino M, Xilu W, Takashima K, Katsura S, Mizuno A. Sterilization using die lectric barrier discharge at atmospheric pressure; 2005. IEEE. pp. 784 788. 51. Deng S, Cheng C, Ni G, Meng Y, Chen H (2008) Bacterial inactivation by atmospheric pressure dielectric barrier discharge plasma jet. Japanese Journal of Applied Physics 47: 700 9. 52. Joshi SG, Paff M, Friedman G, Fridman G, Fridman A, et al. (2010) Control of methicillin resistant Staphylococcus aureus in planktonic form and biofilms: A biocidal efficacy study of nonthermal dielectric barrier discharge plasma. American journal o f infection control 38: 293 301.
172 53. Hury S, Vidal D, Desor F, Pelletier J, Lagarde T (1998) A parametric study of the destruction efficiency of Bacillus spores in low pressure oxygen based plasmas. Letters in applied microbiology 26: 417 421. 54. Lerouge S, Wertheimer M, Marchand R, Tabrizian M, Yahia LH (2000) Effect of gas composition on spore mortality and etching during low pressure plasma sterilization. Journal of biomedical materials research 51: 128 135. 55. Ying J, Chunsheng R, Zhilong X, Dezhen W, Younian W, et al. (2006) Comparison of yeast inactivation treated in He, Air and N2 DBD plasma. Plasma Science and Technology 8: 720. 56. Wrobel A, Lamontagne B, Wertheimer M (1988) Large area microwave and radiofrequency plasma etching of polymers. Plasm a chemistry and plasma processing 8: 315 329. 57. Kao KC, Blank G, Madrid F (1996) Microwave plasmas for low temperature dry sterilization. Biomaterials 17: 1273 1277. 58. Bol'Shakov A, Cruden B, Mogul R, VS Rao M, Shama S, et al. (2004) Radio frequency ox ygen plasma as a sterilization source. AIAA journal 42: 823 832. 59. Lerouge S, Fozza A, Wertheimer M, Marchand R, Yahia LH (2000) Sterilization by low pressure plasma: the role of vacuum ultraviolet radiation. Plasmas and Polymers 5: 31 46. 60. Moreau S, Moisan M, Tabrizian M, Barbeau J, Pelletier J, et al. (2000) Using the flowing afterglow of a plasma to inactivate Bacillus subtilis spores: Influence of the operating conditions. Journal of Applied Physics 88: 1166 1174. 61. Sohbatzadeh F, Colagar AH, Mir zanejhad S, Mahmodi S (2010) E. coli, P. aeruginosa, and B. cereus bacteria sterilization using afterglow of non thermal plasma at atmospheric pressure. Applied biochemistry and biotechnology 160: 1978 1984. 62. Philip N, Saoudi B, Crevier M C, Moisan M, B arbeau J, et al. (2002) The respective roles of UV photons and oxygen atoms in plasma sterilization at reduced gas pressure: the case of N2 O2 mixtures. Plasma Science, IEEE Transactions on 30: 1429 1436. 63. Dobrynin D, Fridman G, Friedman G, Fridman A (2 009) Physical and biological mechanisms of direct plasma interaction with living tissue. New Journal of Physics 11: 115020.
173 64. Kong MG, Kroesen G, Morfill G, Nosenko T, Shimizu T, et al. (2009) Plasma medicine: an introductory review. New Journal of Physi cs 11: 115012. 65. Rastogi RP, Kumar A, Tyagi MB, Sinha RP (2010) Molecular mechanisms of ultraviolet radiation induced DNA damage and repair. Journal of nucleic acids 2010. 66. Halfmann H, Denis B, Bibinov N, Wunderlich J, Awakowicz P (2007) Identificatio n of the most efficient VUV/UV radiation for plasma based inactivation of Bacillus atrophaeus spores. Journal of Physics D: Applied Physics 40: 5907. 67. Moisan M, Barbeau J, Crevier M C, Pelletier J, Philip N, et al. (2002) Plasma sterilization. Methods a nd mechanisms. Pure and Applied Chemistry 74: 349 358. 68. Laroussi M, Leipold F (2004) Evaluation of the roles of reactive species, heat, and UV radiation in the inactivation of bacterial cells by air plasmas at atmospheric pressure. International Journal of Mass Spectrometry 233: 81 86. 69. Laroussi M, Mendis D, Rosenberg M (2003) Plasma interaction with microbes. New Journal of Physics 5: 41. 70. Stoffels E, Sakiyama Y, Graves DB (2008) Cold atmospheric plasma: charged species and their interactions with cells and tissues. Plasma Science, IEEE Transactions on 36: 1441 1457. 71. Laroussi M (2005) Low Temperature Plasma Based Sterilization: Overview and State of the Art. Plasma Processes and Polymers 2: 391 400. 72. Broadwater WT, Hoehn R, King P (1973) Sensitivity of three selected bacterial species to ozone. Applied Microbiology 26: 391 393. 73. Sweet F, Kao M S, Lee S Hagar WL, Sweet WE (1980) Ozone selectively inhibits growth of human cancer cells. Science 209: 931 933. 74. Efremov N, Adamiak BY, Blouchin V, Dadashev SJ, Dmitriev K, et al. (2000) Action of a self sustained glow discharge in atmospheric pressure air o n biological objects. Plasma Science, IEEE Transactions on 28: 238 241. 75. Ohkawa H, Akitsu T, Tsuji M, Kimura H, Kogoma M, et al. (2006) Pulse modulated, high frequency plasma sterilization at atmospheric pressure. Surface and Coatings Technology 200: 58 29 5835. 76. Mastanaiah N, Saxena U, Johnson JA, Roy S. Inactivation of Yeast cells using Dielectric Barrier Discharge; 2010. AIAA 2010 1340, 48th AIAA Aerospace Sciences Meeting, Orlando, FL.
174 77. Brelles Mario G (2010) Bacterial biofilm inactivation by g as discharge plasmas. Biological and environmental applications of gas discharge plasmas Nova Science Publisher, New York, ISBN: 978 971. 78. Sladek R, Filoche S, Sissons C, Stoffels E (2007) Treatment of Streptococcus mutans biofilms with a nonthermal atm ospheric plasma. Letters in applied microbiology 45: 318 323. 79. Vandervoort KG, Abramzon N, Brelles Marino G (2008) Plasma interactions with bacterial biofilms as visualized through atomic force microscopy. Plasma Science, IEEE Transactions on 36: 1296 1 297. 80. Lee MH, Park BJ, Jin SC, Kim D, Han I, et al. (2009) Removal and sterilization of biofilms and planktonic bacteria by microwave induced argon plasma at atmospheric pressure. New Journal of Physics 11: 115022. 81. Kumar R, Saurav S, Titov E, Levin D, Long R, et al. (2011) Thermo structural studies of spores subjected to high temperature gas environments. International Journal of Heat and Mass Transfer 54: 755 765. 82. Gallagher MJ, Vaze N, Gangoli S, Vasilets VN, Gutsol AF, et al. (2007) Rapid inact ivation of airborne bacteria using atmospheric pressure dielectric barrier grating discharge. Plasma Science, IEEE Transactions on 35: 1501 1510. 83. Akishev Y, Grushin M, Karalnik V, Trushkin N, Kholodenko V, et al. (2008) Atmospheric pressure, nonthermal plasma sterilization of microorganisms in liquids and on surfaces. Pure and Applied Chemistry 80: 1953 1969. 84. Kutasi K, Pintassilgo C, Coelho P, Loureiro J (2006) Modelling of a post discharge reactor used for plasma sterilization. Journal of Physics D : Applied Physics 39: 3978. 85. Pintassilgo C, Loureiro J, Guerra V (2005) Modelling of a N2 O2 flowing afterglow for plasma sterilization. Journal of Physics D: Applied Physics 38: 417. 86. Babaeva NY, Kushner MJ (2010) Intracellular electric fields produ ced by dielectric barrier discharge treatment of skin. Journal of Physics D: Applied Physics 43: 185206. 87. Tripathi P, Beaussart A, Andre G, Rolain T, Lebeer S, et al. (2012) Towards a nanoscale view of lactic acid bacteria. Micron 43: 1323 1330. 88. Ant ifungals ( http://www.sigmaaldrich.com/technicaldocuments/articles/biofiles/antifungals.html )
175 89. Lectures on Microbial Growth ( http://www.readbag.com/aesop rutgers dbm microbialgrowth ) 90. Niemel S (2002) Uncertainty of quantitative determinations derived by cultivation of microorganisms: Centre for Metrology and Accreditation Helsinki. 91. Kearns DB, Chu F, Branda SS, Kolter R, Losick R (2005) A master regulator for biofilm formatio n by Bacillus subtilis. Molecular microbiology 55: 739 749. 92. Kearns DB, Losick R (2003) Swarming motility in undomesticated Bacillus subtilis. Molecular microbiology 49: 581 590. 93. Branda SS, Gonzlez Pastor JE, Ben Yehuda S, Losick R, Kolter R (2001) Fruiting body formation by Bacillus subtilis. Proceedings of the National Academy of Sciences 98: 11621 11626. 94. Branda SS, Gonzlez Pastor JE, Dervyn E, Ehrlich SD, Losick R, et al. (2004) Genes involved in formation of structured multicellular communi ties by Bacillus subtilis. Journal of Bacteriology 186: 3970 3979. 95. Gordon R, Haynes W, Pang C (1989) The genus bacillus. US Department of Agriculture handbook: 109 126. 96. Mackey B, Miles C, Parsons S, Seymour D (1991) Thermal denaturation of whole ce lls and cell components of Escherichia coli examined by differential scanning calorimetry. Journal of general microbiology 137: 2361 2374. 97. Pearse RWB, Gaydon AG, Pearse RWB, Gaydon AG (1976) The identification of molecular spectra: Chapman and Hall Lon don. 98. Mao Z, Dong L, Ran J. Spectrum of dielectric barrier discharge in atmosphere; 2005. International Society for Optics and Photonics. pp. 319 324. 99. Choi JH, Han I, Baik HK, Lee MH, Han D W, et al. (2006) Analysis of sterilization effect by pulsed dielectric barrier discharge. Journal of electrostatics 64: 17 22. 100. Staack D, Farouk B, Gutsol AF, Fridman A (2007) Spatially resolved temperature measurements of atmospheric pressure normal glow microplasmas in air. Plasma Science, IEEE Transactions on 35: 1448 1455. 101. Staack D, Farouk B, Gutsol AF, Fridman AA (2006) Spectroscopic studies and rotational and vibrational temperature measurements of atmospheric pressure normal glow plasma discharges in air. Plasma sources science and technology 15: 81 8.
176 102. Spaan M, Leistikow J, Schulz Von Der Gathen V, Dbele H (2000) Dielectric barrier discharges with steep voltage rise: laser absorption spectroscopy of NO concentrations and temperatures. Plasma Sources Science and Technology 9: 146. 103. Winter J, Dnnbier M, Schmidt Bleker A, Meshchanov A, Reuter S, et al. (2012) Aspects of UV absorption spectroscopy on ozone in effluents of plasma jets operated in air. Journal of Physics D: Applied Physics 45: 385201. 104. Eliasson B, Hirth M, Kogelschatz U (1987) Ozone synthesis from oxygen in dielectric barrier discharges. Journal of Physics D: Applied Physics 20: 1421. 105. Kogelschatz U, Eliasson B, Hirth M (1988) Ozone generation from oxygen and air: discharge physics and reaction mechanisms. 106. Singh KP, Ro y S (2007) Modeling plasma actuators with air chemistry for effective flow control. Journal of applied physics 101: 123308 123308 123308. 107. Singh KP, Roy S (2008) Force approximation for a plasma actuator operating in atmospheric air. Journal of Applied Physics 103: 013305 013305 013306. 108. Ayan H, Fridman G, Staack D, Gutsol AF, Vasilets VN, et al. (2009) Heating effect of dielectric barrier discharges for direct medical treatment. Plasma Science, IEEE Transactions on 37: 113 120. 109. Cho M, Kim J, K im JY, Yoon J, Kim J H (2010) Mechanisms of Escherichia coli inactivation by several disinfectants. Water Research 44: 3410 3418. 110. Zhang Y, Wu Q, Zhang J, Yang X (2011) Effects of ozone on membrane permeability and ultrastructure in Pseudomonas aerugin osa. Journal of applied microbiology 111: 1006 1015. 111. Norton MD, Spilkia AJ, Godoy VG (2013) Antibiotic resistance acquired through a DNA damage inducible response in Acinetobacter baumannii. Journal of bacteriology 195: 1335 1345. 112. Vlamakis H, Agu ilar C, Losick R, Kolter R (2008) Control of cell fate by the formation of an architecturally complex bacterial community. Science Signaling 22: 945. 113. Kearns DB, Chu F, Rudner R, Losick R (2004) Genes governing swarming in Bacillus subtilis and evidenc e for a phase variation mechanism controlling surface motility. Molecular microbiology 52: 357 369. 114. Joshi SG, Cooper M, Yost A, Paff M, Ercan UK, et al. (2011) Nonthermal dielectric barrier discharge plasma induced inactivation involves oxidative DNA damage and
177 membrane lipid peroxidation in Escherichia coli. Antimicrobial agents and chemotherapy 55: 1053 1062. 115. Oehmigen K, Hhnel M, Brandenburg R, Wilke C, Weltmann KD, et al. (2010) The role of acidification for antimicrobial activity of atmospher ic pressure plasma in liquids. Plasma Processes and Polymers 7: 250 257. 116. Mastanaiah N, Johnson JA, Roy S (2013) Effect of Dielectric and Liquid on Plasma Sterilization using Dielectric Barrier Discharge Plasma. PLoS One (In Press) 117. Abuja PM, Ester bauer H (1995) Simulation of lipid peroxidation in low density lipoprotein by a basic" skeleton" of reactions. Chemical research in toxicology 8: 753 763. 118. Mastanaiah N, Wang CC, Johnson J, Roy S. A computational diagnostic tool for understanding plasm a sterilization; 2011. pp. 2011.
178 BIOGRAPHICAL SKETCH In 2003, Navya began her undergraduate studies at the Indian Institute of Technology (IIT), Madras, India, where she went on to receive her Bachelor of Technology (B.Tech) degree in aerospace e ngineer ing. After finishing her undergraduate curriculum in 2007, Navya received a partial scholarship from the University of Florida, to pursue a Master of Scie nce (M.S.) degree in aerospace e ngineering. Under the guidance of Dr.Subrata Roy, she began working to wards her doctoral degree in aerospace e ngineering in 2008.