Bactericidal Action of the Reactive Species …...discharge at atmospheric pressure (from [16]). molecules and particles present, often very rapidly. The use of the gas discharge to

IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 34, NO. 4, AUGUST 2006 1257

Bactericidal Action of the Reactive Species Producedby Gas-Discharge Nonthermal Plasma at

Atmospheric Pressure: A ReviewLindsey F. Gaunt, Clive B. Beggs, and George E. Georghiou

Abstract—Biological decontamination using a nonthermal gasdischarge at atmospheric pressure in air is the subject of signifi-cant research effort at this time. The mechanism for bacterial de-activation undergoes a lot of speculation, particularly with regardto the role of ions and reactive gas species. Two mechanisms havebeen proposed: electrostatic disruption of cell membranes andlethal oxidation of membrane or cytoplasmic components. Resultsshow that death is accompanied by cell lysis and fragmentation inGram-negative bacteria but not Gram-positive species, althoughcytoplasmic leakage is generally observed. Gas discharges can bea source of charged particles, ions, reactive gas species, radicals,and radiation (ultraviolet, infrared, and visible), many of whichhave documented biocidal properties. The individual roles playedby these in decontamination are not well understood or quantified.However, the reactions of some species with biomolecules are doc-umented otherwise in the literature. Oxidative stress is relativelywell studied, and it is likely that exposure to gas discharges inair causes extreme oxidative challenge. In this paper, a review ispresented of the major reactive species generated by nonthermalplasma at atmospheric pressure and the known reactions of thesewith biological molecules. Understanding these mechanisms be-comes increasingly important as plasma-based decontaminationand sterilization devices come closer to a wide-scale applicationin medical, healthcare, food processing, and air purification appli-cations. Approaches are proposed to elucidate the relative impor-tance of reactive species.

Index Terms—Bacteria, nonthermal plasmas, reactive oxygenspecies (ROS), sterilization, superoxide.

I. INTRODUCTION

A LMOST since the discovery of small air ions in the 1890’s[1], their biological significance and impact has fasci-

nated scientists. Whether air ions might influence physiologicalprocesses has been discussed at length in the literature for over ahundred years (for example see [2] and [3]). The study of ionicaction on microorganisms was pioneered in the 1930s whenbacteriostatic effects were observed with ion concentrations ofbetween 5 × 104 and 5 × 106 ions cm−3 [4].

Nonthermal plasma produced by a gas discharge produces acocktail of reactive molecules that continually react with other

Manuscript received November 21, 2005; revised April 10, 2006.L. F. Gaunt is with the School of Electronics and Computer Science,

University of Southampton, SO17 1BJ Southampton, U.K. (e-mail: [email protected]).

C. B. Beggs is with the School of Engineering, Design and Technol-ogy, University of Bradford, BD7 1DP Bradford, U.K. (e-mail: [email protected]).

G. E. Georghiou is with the Department of Electrical and Computer Engi-neering, University of Cyprus, Nicosia 1678, Cyprus (e-mail: [email protected]).

Digital Object Identifier 10.1109/TPS.2006.878381

Fig. 1. Kinetics of the inactivation process. A multislope survivor curve forPseudomonas aeuroginosa on a nitrocellulose membrane filter exposed to glowdischarge at atmospheric pressure (from [16]).

molecules and particles present, often very rapidly. The use ofthe gas discharge to generate ozone and thus disinfect waterhas been practiced for over a 150 years [5]. However, there hasbeen a recent resurgence in the study of the gas discharge fordisinfection technologies, and more specifically using nonther-mal plasma at atmospheric pressure [6]–[9]. The bactericidalagents generated may include reactive oxygen species (ROS),ultraviolet (UV) radiation, energetic ions, and charged particles.Among the ROS reported in air are ozone, atomic oxygen,superoxide, peroxide, and hydroxyl radicals.

In low-pressure plasma, UV radiation is thought to be themost important factor in sterilization [10], [11], but in at-mospheric pressure plasma most of the UV radiation producedis reabsorbed in the plasma volume, and not delivered to thetreated surface. Studies have shown that the kinetics of celldeath during plasma exposure is not indicative of UV radiation[12], [13].

Broad spectrums of microorganisms are susceptible toplasma exposure, including Gram-negative and Gram-positivebacteria, bacterial endospores, yeasts, viruses [14], and biofilms[15]. Reductions in bacteria viability of over 6 log10 are re-ported from short exposures of less than 30 s [7], [14], [16], anda total cell fragmentation is seen after longer exposures. In someinstances, the kinetics of cell death demonstrate single-slopesurvivor curves [13], [16], while in others multislope curves areseen, as shown in Fig. 1 [16], [17]. Multislope curves suggest

0093-3813/$20.00 © 2006 IEEE

Authorized licensed use limited to: Univ of Calif Los Angeles. Downloaded on April 13,2010 at 16:42:56 UTC from IEEE Xplore. Restrictions apply.

1258 IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 34, NO. 4, AUGUST 2006

Fig. 2. Transmission electron micrograph images of Gram-negativeEscherichia coli [(a) and (b)] and Gram-positive Staphylococcus aureus [(c)and (d)] bacteria before [(a) and (c)] and after [(b) and (d)] a 30-s exposure toglow discharge at atmospheric pressure. Bar = 1 µm (from [7]).

different inactivation methods are in effect, possibly triggeredby different reactive gas species.

Typical of most sterilization techniques, the rates and magni-tudes of cell killing in response to a particular treatment regimediffer for various species and strains of bacteria. Spores aregenerally more resistant than vegetative or actively growingbacteria, while there is no consensus on the relative susceptibil-ity of Gram-negative or Gram-positive bacteria. The supportingmedium also influences killing times, with inactivation beingslower on agar than on polypropylene [6].

Bacterial inactivation is accompanied by leakage of proteins,deoxyribonucleic acid (DNA) and ribonucleic acid (RNA) fromthe cellular cytoplasm [6]. These macromolecules are detectiblein the supernatant of a cell suspension of Gram-negative E. colitreated with atmospheric pressure plasma after 10 s, and after15 s of treatment for Gram-positive S. aureus. Observations ofthe physical damage inflicted on Gram-negative and -positivebacteria show marked differences. Transmission electron mi-croscopy, such as shown in Fig. 2, reveal that the continuityof the cellular envelope of E. coli is breached after 30 s oftreatment, while there is no apparent destruction of S. aureus[6], [7]. Impairment of some metabolic function, not resultingin cell death, has been reported from sublethal exposures,indicative of changes in enzyme activity [18], [19]. There aretwo main hypotheses for the mechanism of cell death caused bygas discharge, both involving lethal damage to cell membranestructures and ultimately leading to leakage of cyptoplasmiccontents or lysis. They are electrostatic disruption and oxi-dation of membrane components. The electrostatic disruptionmechanism [20] suggests that the total electric force caused byaccumulation of surface charge could exceed the total tensileforce on the membrane (Gram negative), the probability of

which is greater where some surface irregularities give regionsof higher local curvature. The tensile strength of the membraneis conferred by a murein layer, which is thicker in Gram-positive bacteria (∼ 15–18 nm) than Gram-negative bacteria(∼ 2 nm), meaning a lower accumulated charge would berequired for lysis of Gram-negative bacteria than Gram positive.

In the second mechanism, oxidation and damage of mem-brane or cellular components are suggested to be caused bythe energetic ions, radicals, and reactive species generated bygas discharge. Active radicals are generated directly in plasmaand diffuse to the cell surface, while reaction chemistry ina moisture layer on the cell surface can produce secondaryradicals. It is well documented that ROS have profoundlydamaging effects on cells through reactions with various bio-macromolecules [21]–[24]. The involvement of superoxide inthe bactericidal effects of a corona discharge is alluded to bythe protective effects demonstrated by superoxide dismutase(SOD) enzymes [25]. Ozone has well-recognized properties asa disinfectant while hydrogen peroxide acts as a bacteriostaticagent at concentrations of 25–50 µm [23].

ROS are generated by the normal respiratory process of aero-bic organisms. Hence, virtually all aerobic organisms, includingbacteria, demonstrate oxidative defense and repair mechanismson exposure to oxidizing agents such as those generated byair plasma [21], [22]. Indeed, both plant and animal hostshave adopted defense strategies which utilize ROS (super-oxide, hydrogen peroxide, hydroxyl radical) against invadingmicroorganisms [24]. Oxidative stress occurs when the levelsof exposure exceed the capacity of the cell defense systems.

The mechanism of microorganism inactivation by the gasdischarge is undoubtedly complex and heterogeneous in nature.In microbiology, even the process of determining cell death isnot straightforward. Inactivation or loss of viability occurs atthe point in time when vital cellular components suffer certainlevels of irreversible damage. Often this can initiate a chainreaction of supplementary damage unrelated to that directlycaused by the exposure. For this reason, the initial reactionsin the sequence leading to the death of the study organismsshould be identified in order to truly elucidate the disinfectionmechanism. Post hoc observations of bacteria eradicated byspecific treatments are not necessarily indicative. Chemicalreactions between the reactive species generated will continueafter death, while compounds released by dead cells may causefurther physical damage. Thus, while cell lysis may be correctlyobserved, loss of membrane integrity may have played no activerole in cell death. This point is exemplified in a study of Es-cherichia coli inactivation by ozone in which scanning electronmicroscopy (SEM) images show that noticeable changes to cellinteriors occur only after most of the cells in the sample havebecome nonviable [26]. Decomposition of dissolved ozonemeasurements suggested that only 25% of the ozone demandwas responsible for inactivation of practically all microorgan-isms present and that it was subsequent ozone exposure that ledto structural changes, membrane deterioration, and cell lysis.

This review of biological decontamination by gas-dischargenonthermal plasma at atmospheric pressure will focus onthe killing of microorganisms by oxidation and damage tomembrane or cellular components by the reactive gas species


GAUNT et al.: BACTERICIDAL ACTION OF THE REACTIVE SPECIES PRODUCED BY NONTHERMAL PLASMA 1259

generated. A discussion of the electrostatic disruption mecha-nism is given otherwise [20], [27]. An overview of the proposedinactivation agents generated by the gas discharge is presented.Their reactions with biomolecules and the implications of thesefor bacterial viability are discussed, drawing on the literature atlarge. Also discussed are the defense mechanisms demonstratedby bacteria on exposure to some reactive species.

II. NONTHERMAL-PLASMA PRODUCTION AT

ATMOSPHERIC PRESSURE THROUGH GAS DISCHARGES

Nonthermal plasma is increasingly playing a central part inmany areas of modern technology. To date, nonthermal, “cold,”or nonequilibrium plasma (operating at low neutral gas tem-perature, while the electron temperature can be much higher)has been successfully utilized in the area of material processingas well as pollution control, sterilization, disinfection, ozoneproduction, and surface processing [28].

The production and maintenance of plasma constitutes one ofthe main challenges in the area of plasma technology. Plasmais generated when enough energy is supplied to a neutral gas tocause charge production. This is achieved by ionization or pho-toionization when electrons or photons with sufficient energycollide with the neutral atoms or other molecules. The mostwidely used method for plasma generation utilizes the electricalbreakdown of a neutral gas (gas discharge) in the presence of anexternal electric field. In this case, electrons and other chargedparticles accelerate under the externally applied electric fieldand transfer their energy through inelastic collisions to otherparticles and atoms [29]. Discharges are classified as dc, ac,or pulsed discharges on the basis of the temporal behavior ofthe sustaining electric field. The spatial and temporal charac-teristics of plasma depend to a large degree on the particularapplication for which the plasma is utilized.

The majority of gas-discharge nonthermal-plasma processeshas, so far, operated at low pressure. Nevertheless, the latestdevelopments in plasma science have opened up a way to createnonthermal plasma at atmospheric pressure. The study of thisplasma has received much interest in recent years because ofthe need to replace expensive vacuum systems by simpler ones,thus leading to higher throughput and cost reduction [30]. Thecurrent major challenge is to produce, stabilize, and controlsuch plasma at atmospheric pressure.

Typical examples of the utilization of gas-discharge nonther-mal plasma at atmospheric pressure include the production ofozone for air cleaning [31], sterilization and gas treatment [32],pollution control, CO2 lasers [28], [33]–[35], surface treatment,high-efficiency excimer UV light sources, and plasma displaypanels [36]–[38].

Central to the production of nonequilibrium plasma at at-mospheric pressure lies the dielectric barrier discharge (DBD)(Fig. 3). DBDs, also known as silent discharges, provide asimple technology to establish nonthermal-plasma conditionsin atmospheric pressure gases [39]. As the emergence of theDBD is responsible for the renewed interest in atmosphericpressure discharges a brief overview of the DBDs and theiroperation is given next. More details and a review of theirpotential applications are outlined in [40].

Fig. 3. Some typical configurations of the DBD.

DBDs have been known for more than a century and wereoriginally developed for large-scale industrial ozone production[41]. In its simplest configuration, a DBD is the gas dischargebetween two electrodes, separated by one or more dielectriclayers and a gas-filled gap. When high ac voltage is appliedto the electrodes, the resulting electric field is adequate toproduce ionization of the gas in the gap. The radicals, ions,and electrons produced are attracted toward the electrodes ofopposite polarity and form a charge layer on the dielectricsurface. These charges cancel the charge on the electrodes sothat the electric field in the gap falls to zero, the discharge stops,and the current is limited. A weak current and hence low-powerdischarge is achieved. However, the activation of the workinggas in the discharge space enables the generated radicals andatoms to be applied to treat the surface layer. The early phasesof breakdown observed in barrier discharges are similar to thosewithout a dielectric, namely avalanche, streamer formation,and propagation between the electrodes [42]–[46] followed bystreamer decay. Fig. 4 shows the early phases of breakdownbetween two parallel electrodes for different instances. Thepresence of the dielectric barriers ensures that the current islimited so that the discharge does not transit to a spark or arcand the plasma therefore remains cold [40], [46].

Depending on the gas mixture, dielectric surface proper-ties, and operating conditions, vastly different modes, rangingfrom filamentary to completely diffuse barrier discharges, havebeen observed [28]. In most DBD configurations operating inatmospheric pressure gases, miniature filamentary dischargescalled microdischarges are formed [48] (Fig. 5). Most of theindustrial applications of DBDs operate in this filamentarymode [28].

Under controlled operating conditions, stable, uniform, non-filamentary, or glow discharges can be produced which areoften more effective for surface modification compared tofilamentary discharges [49]. In recent years, a uniform mode ofthe barrier discharge [50], the so-called atmospheric pressureglow (APG) discharge, has been reported for several gases and



Fig. 4. Contours of electron density in a 0.1-cm gap at 5.6-kV applied voltage. Critical avalanche and cathode streamer formation and propagation. The fourinstances correspond to the following times: (a) t = 3.64 ns, (b) t = 4.13 ns, (c) t = 6.32 ns, and (d) t = 6.49 ns (from [30]).

Fig. 5. Photograph depicting microdischarge formations (from [28]).

gas mixtures. The APG is preferred to the filamentary mode forapplications that require uniform plasma production such as thedeposition of uniform thin films or surface treatment [51]. Theirspatial appearance is characteristically diffuse and uniform, andtheir temporal features are reproducible and stable [47].

Both discharge types, filamentary and diffuse, have commonfeatures: the generation of cold nonequilibrium plasmas atatmospheric pressure and the strong influence of local fielddistortions caused by space charge accumulation.

III. BREAKDOWN MECHANISMS

At atmospheric pressure, breakdown occurs in the form ofmicrodischarges for the majority of cases in DBD configura-tions. Under certain circumstances, however, diffuse as well asglow discharges can be obtained [41]. The form of the dischargeplays an important role in the application and utilization ofplasma.

In any volume of gas, there exist free electrons that arecaused by cosmic rays, natural radioactivity, and the detach-ment of negative ions. Thus, if the electric field is high enough,these will accelerate and collide with molecules of the gas,thereby releasing more electrons, which in turn will do thesame, creating what is known as an electron avalanche. In thisway, electron numbers multiply [52], [53].

As long as the net charge is not sufficient to distort the fieldappreciably, the center of the avalanche moves with the electrondrift velocity appropriate to the applied field. If, during thelife of the avalanche, secondary electrons are released, thennew avalanches will be created and the total current will beamplified. Secondary electrons are released either by positiveions or UV photons hitting the cathode, or by photoionizationof the gas behind an avalanche. In this way, the current growsexponentially by what is known as the “Townsend breakdownmechanism” [25]. The Townsend mechanism is the first modelproposed to explain the initial phase of electrical breakdown ingases at high pressures. Diffuse discharges, recently observed in



DBDs, can be obtained at breakdown when there is a sufficientoverlap of simultaneously propagating electron avalanches tocause smoothing of transverse field gradients. It has repeatedlybeen demonstrated that diffuse discharges can also be obtainedin barrier discharge configurations at about atmospheric pres-sure and gap widths up to several centimeters [54]. Low-currentdiffuse barrier discharges are more like Townsend discharges,in which the charge density is so low that it has practically noinfluence on the applied field.

At the high pressures usually encountered in a DBD, how-ever, a different kind of breakdown takes place (streamer),due to the considerable space charge generated during the firstavalanche’s transit through the gap. The streamer mechanismwas initially proposed to explain the electrical breakdown ofovervolted gaps at near-atmospheric pressure [55], [56]. Essen-tially a streamer is an ionization wave. In front of the wave (thestreamer head) the separation of positive and negative chargedparticles shield the interior, and cause a sharp enhancement ofthe electric field over a limited region just outside the streamerhead. For fields near the breakdown threshold, the ionizationcoefficient is a strong function of the electric field, so that evenmodest field enhancements can result in substantial increasesin the ionization rate. If a mechanism such as transport, photo-emission, or photoionization exists that places a few free seedelectrons just in front of the streamer head, avalanching in thelocally enhanced field can cause the streamer to propagate withvelocities much higher than the peak electron drift velocity.Additionally, the ionization density in the streamer body canbuild up to values considerably larger than that necessary toinitiate streamer formation. This explains why at atmosphericpressures the breakdown mechanism in air and other gases isfound to be very fast (of the order of an avalanche transit time)and to consist of a filamentary discharge channel, rather than adiffuse distribution of avalanches.

At the time the streamer bridges the gap, a cathode-fallregion of high electric field and high positive-ion densities isestablished within a fraction of a nanosecond. Glow dischargesare characterized by a localized high-field cathode-fall region.At atmospheric pressure, the thickness of this high-field regionis about 10 µm. The current peak in a microdischarge is reachedat the time of cathode-layer formation. Immediately afterward,charge accumulation at the dielectric surface leads to a localcollapse of the electric field in the area defined by the surfacecharge. This self-arresting effect of the dielectric barrier limitsthe duration of a microdischarge to a few nanoseconds, andhence the dissipated energy and temperature rise remain verylow [46].

IV. CHEMISTRY OF ELECTRIC DISCHARGES

The properties of nonthermal plasmas are determined bycollisions between electrons and other plasma constituents. Inair, chemical reactions are mainly initiated by the impact ofelectrons with oxygen and nitrogen. Thus, among the primaryproducts of electron collision are atomic and metastable oxygenand nitrogen, with subsequent reactive collisions producing acocktail of neutral and ionic species. Atmospheric pressuredischarges differ from low-pressure plasma in that the chem-

TABLE ITYPICAL DENSITIES OF OXYGEN IONS, OXYGEN ATOMS, OZONE,

AND CHARGED SPECIES IN PLASMA DISCHARGES [56]

istry is dominated by reactive neutral species such as oxygenatoms, singlet oxygen, and ozone rather than ions [57], [58]. Inaddition to gas-phase processes, surface reactions should alsobe taken into account when considering the species to whichbiological samples are exposed. As this paper is concernedwith oxidative damage caused by gas discharges at atmosphericpressure in air or the presence of oxygen, the most significantreactive species are ozone (O3), atomic oxygen (O or O•−),superoxide (O2•−), peroxide (O−2

2 or H2O2), and hydroxylradicals (OH•). A full description of air plasma chemistry,including tables of relevant collision processes, is providedotherwise [59].

Table I summarizes the typical densities of some ion andcharge species in various plasma devices. In corona and DBDozone is the main reaction product, while in other plasmasources, oxygen atoms represent a larger proportion of thereactive species. A classical application of corona and DBD dis-charges are in the commercial production of ozone for air andwater purification. In a numerical investigation, the distributionof oxygen species generated by a coaxial corona potential wasmodeled [60]. Ozone was found to be the dominant species at aconcentration of 5 × 1018 cm−3 with singlet oxygen and atomicoxygen being about five to six orders of magnitude lowerin concentration. Ionized species were present at an averagedensity of only 1010 cm−3. In another example, a model ofozone generation in a DBD calculated that most of the chargedoxygen species formed were short lived, with O3 being theenduring reactive species [61].

In another study, time-resolved UV absorbance spectroscopy,optical emission spectroscopy, and numerical modeling meth-ods were used to investigate the reaction chemistry of at-mospheric pressure discharges [62]. The concentrations ofground state molecular oxygen, ground state oxygen atoms,metastable molecular oxygen, and ozone were quantified,considering the process conditions of gas pressure and thedistance from the plasma. The plasma generated 0.2−1.0 ×1016 cm−3 ground state oxygen atoms, between 2.0 × 1014 and1 × 1016 cm−3 metastable molecular oxygen, and 0.1−4.0 ×1015 cm−3 ozone. Ozone concentration increased after theplasma was extinguished due to reactions between oxygenatoms and oxygen molecules. This research is in agreementwith studies of DBD discharges showing that the chargedoxygen species are rapidly extinguished within micro- or mil-liseconds.



The presence of significant concentrations of atomic oxy-gen in a uniform glow discharge was demonstrated throughthe oxidation of oil of wintergreen by discharge effluvia [6].Oxidation of this agent would not be expected through reactionwith ozone.

Hydroxyl radicals OH• are highly reactive species that cancause significant damage to most biological molecules, and arereported amongst the reactive species in many types of plasma[6], [7], [16]. The presence of water vapor in a discharge feedgas results in the formation of OH• by electron impact dissoci-ation of H2O and by reactions of electronically excited oxygenatoms and nitrogen molecules [9]. It is probable that hydroxylradicals also form during surface reactions, for example throughthe reaction (1) between hydrogen peroxide and superoxide

O2 •− +H2O2 → O2 + OH • +HO−. (1)

The presence of hydroxyl radicals in the effluent of an RF-driven plasma has been observed using laser-induced fluo-rescent spectroscopy [63]. However, as this discharge was insaline, the outcome is likely to be different from a dischargein air. The presence of OH• in the microdischarges of a silentdischarge plasma reactor has also been imaged using OH•fluorescence [64]. It was demonstrated that OH• only existsin the channels formed by the transient discharges, and didnot diffuse out. The presence of water vapor in the feed gasfor the device promotes OH• production while simultaneouslyreducing ozone concentration [65]. As the water vapor reducesthe surface resistance and increases the dielectric capacity, totalcharge transfer is reduced. Dry air is used in the majorityof plasma devices as water vapor reduces the number of mi-crodischarges, and thus the plasma volume. For example, aglow discharge is operated below 14% relative humidity (RH)to obtain the desired uniform discharge across the electrodegap [6], [7].

The presence of superoxide (O2•−) in the effluvia of adischarge is difficult to detect because it is short-lived anddoes not accumulate. Its presence in ionized air is more oftenassumed where hydrogen peroxide is detected, as superoxide isa common precursor for this species. The radical can be gen-erated through the combination of an electron with an oxygenmolecule, and its chemistry can involve reaction with water toform hydrated cluster ions. Spontaneous decomposition occursthrough a dismutation reaction

2O2 •− +2H+ → H2O2 + O2. (2)

Evidence for the generation of superoxide during negativeair ionization is given by several authors [25], [66], [67].Its generation in the corona discharge was suggested by theprotection conferred by the enzyme SOD to bacteria exposedto negative corona [25]. Exposing the bacteria Staphylococcusalbus in solution to the products of a negative corona, withan estimated generation rate of 0.3 mmol O2•− per hour,substantial loss of viability was observed after 2 h and total lossseen after 5 h. Addition of SOD provided complete protectionfrom the corona treatment such that total viability was retained,suggesting the superoxide radical to be largely responsiblefor bacterial death. The accumulation of O2•− in a solution

treated with an electroeffluvial ionizer was also measured usingelectron paramagnetic resonance and found to be in the regionof 0.5−1.0 µm/min [67].

Hydrogen peroxide is a product of negative air ionization inthe presence of water or water vapor, and is thus generated bymost ionizers [66], [68]. It can also be formed on the surfaceof cells in the presence of both positive and negative air ions[69]. The concentration of hydrogen peroxide generated bythe corona discharge in air under conditions of increasing RHhas been measured using Fourier-transform infrared (FTIR)spectroscopy, showing that H2O2 concentration increases withRH [70]. In ordinary air at 40% RH the concentration was 0.46rising to 936 µg · l−1 at 96% RH. Only at 10% RH was theH2O2 concentration below the detectable limit.

V. OXIDATIVE DAMAGE IN BACTERIAL CELLS

A great deal of damage can be done to bio macromoleculesby oxygen radicals, but the damage that leads to cell death is notalways clear. The consequences of oxidative damage includethe following.

1) Adaptation of the cell by upregulation of defense systems.2) Cell injury involving damage to molecular targets such as

lipids, DNA, protein, and carbohydrate.3) Cell death usually arising due to excessive unrepaired

damage triggering cell death.It is important to consider that cell injury is not necessarilycaused directly by oxidative damage, but stress-related changesin ion levels or activation of proteases for example may them-selves be damaging. Damage to DNA, RNA, proteins, andlipids has been demonstrated in vivo and in vitro.

ROS are used by the immune systems of multicelled or-ganisms to counteract microbial growth [71]. Consequently,microorganisms have evolved specific multigene antioxidantdefense systems to neutralize their effects [72]. In many ways,the bactericidal action of ROS in cold plasma, mimics thatexperienced by bacteria undergoing phagocytosis in the humanbody. Phagocytic cells utilize respiratory bursts to produceO2•− and H2O2, which kill opsonized (made more suscep-tible to phagocytes) bacteria [73], [74]. Hydrogen peroxideis a bacteriostatic rather that bactericidal agent, dramaticallyincreasing the mean generation time of bacteria at micromolarconcentrations [23]. The toxicity of superoxide radicals, how-ever, derives from their great instability, which causes them toscavenge electrons from neighboring molecules, which in turnbecome radicals. The highly destructive OH• is also knownto be a product of stimulated phagocytes [73], and is thoughtto be generated by a metal-catalyzed reaction between super-oxide and hydrogen peroxide, known as the Fenton reaction(3) [22], [75]

O2 •− +H2O2Fe/Cu−−−−−−→ 2OH • +O2. (3)

Since the dismutation of superoxide produces hydrogenperoxide, it is likely that as the concentration of superoxideincreases, so concentrations of hydrogen peroxide and hydroxylradicals will also increase [22]. The hydroxyl radical is ahighly reactive species, having extremely high rate-constants



Fig. 6. Basic reaction sequence for lipid (L) peroxidation by free radicals.

for reactions with almost every type of molecule found inliving cells [21]. For example, the rate constant for its reac-tion with lecithin (an important membrane phospholipid) is5.0 × 108 m1−s−1 [21], [76]. Consequently, OH• has an av-erage diffusion distance of only a few nanometers, so is likelyto react with cell membrane components in gas exposure [22],[77]. Secondary radicals produced through the reactions ofOH• are inevitably less reactive.

A further reaction of superoxide in aqueous solution is toact as a base and accept a proton to form hydroperoxyl rad-ical (HO2•) (4). Since HO2• will also dissociate to releasethe proton an equilibrium is set up, the position of which ispH dependent. Within the physiological range of 6.4–7.5 thesuperoxide radical remains almost entirely in this form ratherthan becoming protonated [21]

O2 •− +H+ ↔ HO2 • . (4)

ROS can cause a great amount of damage to macromolecules.Biological targets for ROS include DNA, proteins, and lipids[78]. Although oxidative damage may be extensive, it is notalways clear the precise nature of the damage that leads tocell death [22]. Lipids in particular, are major targets duringoxidative stress [78]. Radicals attack the polyunsaturated fattyacids in the cell membrane, initiating lipid peroxidation, settingoff a destructive chain reaction, as shown in Fig. 6. Of theROS formed during oxidative stress, hydroperoxyl radicals(HOO•), superoxide radicals, singlet oxygen, and ozone canall initiate lipid peroxidation [22]. The lipid peroxidation chainreaction is initiated when a hydrogen atom is abstracted froman unsaturated fatty acid to form a lipid radical, which in turnreacts with molecular oxygen to form a lipid peroxyl radical(L-OO•). These radicals attack other unsaturated fatty acidsand abstract hydrogen atoms to form fatty acid hydroperoxides(L-OOH), thus perpetuating the chain reaction. Lipid perox-idation generates products, which are shorter than the initialunsaturated fatty acids [22], with the result that their abilityto rotate within the cell membrane is altered [79]–[81] andthe structural integrity of the membrane is compromised. Loss

of membrane integrity leads to an osmotic imbalance, whichmay ultimately result in cell lysis. During lipid peroxidationaldehydes can also be formed [78]. These in themselves arevery reactive and can damage proteins [77]. However, unlikeROS, aldehydes are longer lived and thus can attack targets,which are more distant from the initial free-radical event.

When proteins are oxidized, modifications occur to theamino acid side chains, with the result that the protein struc-ture is altered [78], [82], [83]. These modifications may leadto functional changes in the proteins, which can disrupt cellmetabolism, with potentially catastrophic effects. Proteins con-taining (Fe-S)4 clusters appear to be particularly sensitive toattack by superoxide radicals [84]. In general, metal-bindingsites in proteins appear to be especially sensitive to attack byROS [22]. When proteins are damaged it is necessary that theybe removed in order to prevent accumulation, which mightotherwise compromise cell metabolism [78]. There is evidenceto suggest that heavily oxidized proteins can inhibit the actionof protease, with the result that the oxidized proteins cannot bedegraded [85], [86].

Another target for ROS is DNA. ROS can cause numeroustypes of DNA lesions, through reactions with both bases andsugar moieties [22]. The hydroxyl radical is the most likelycandidate to produce DNA damage due to its extreme reactivity[87], attacking the sugar moiety and leading to DNA strandbreaks [22]. As OH• cannot diffuse through cells to reach theDNA due to its reactivity, it is likely that hydrogen peroxidemay serve as a diffusible “latent” form, reacting with metal ionsclose to DNA molecules to liberate the highly reactive OH•[75], [87]. In addition to the DNA damage caused directly byROS, intermediate radicals formed during lipid peroxidationalso react with DNA [22]. For example, decomposition ofpurine may occur as a result of H+ abstraction by fatty acid freeradicals or the fatty acid free radicals may add to the purinesto form bulky adducts [22]. Some of the stable end productsof lipid peroxidation, such as aldehydes have been shown tobe reactive with DNA, either by alkylating bases [88] or byforming intra- and interstrand cross-links [89].

Criteria have been proposed [90] for the implication ofreactive species in tissue injury in human disease. The criteriaare as follows.

1) The reactive species should be demonstrable at the site ofinjury.

2) The time course for the formation of reactive speciesshould be consistent with the time course of cell injury,precede or accompany it.

3) Direct application of the reactive species over the relevanttime course, at concentrations within the relevant rangeshould reproduce most or all of the injury observed.

4) Removing or inhibiting formation of the reactive speciesshould diminish the cell injury to an extent related to thedegree of inhibition of the oxidative damage caused.

These criteria could equally be used to implicate ROS inthe current subject. For example, reporting of the densitiesof reactive species at the exposed sample would be helpfulin determining the species influential in killing [criteria 1)].Similarly, the impact of specific reactive species could beimplied where cell death and damage is inhibited by the use



Fig. 7. Cell survival for three strains of Bacillus subtilis exposed to increasingdoses of H2O2. Open symbols show naive cells while closed symbols showcells demonstrating the peroxide stress response induced by preexposure to50-µm H2O2. The three strains are YB886 (wild-type) circles, YB1015 (repair-deficient strain) triangles and YB2001 (catalase deficient) squares (from [72]).

of spin traps to intercept radicals or of enzymes to neutralizespecific species [criteria 4)]. The latter approach was used tosuggest a critical role for superoxide in the bactericidal effectof corona, where SOD conferred complete protection [25].The presence of atomic oxygen at the reaction site of glowdischarge plasma was implied in a similar way, through a studyof the oxidation of the chemical warfare agent simulant: oil ofwintergreen [7].

The reactions between ROS and the various classes ofbiomolecules are relatively well understood, principally as aresult of studies into oxidative stress in aerobic organismsand of bacteria challenged by ROS by the immune systemsof multicelled organisms. These reaction schemes demonstratethe fatal nature of ROS exposure at micro- and nanomolarconcentrations, suggesting a definite role for ROS in biologicaldecontamination using nonthermal gas discharges. Principaldifferences are likely to arise in the concentrations of ROS towhich bacteria are exposed and from the simultaneous exposureto a cocktail of reactive species from air plasma.

VI. OXIDATIVE STRESS AND BACTERIAL DEFENCE

Aerobic bacteria have developed a complex array of preven-tative and reparative mechanisms to protect themselves from thedeleterious effects of oxidative damage. These cellular defensesinclude both enzymatic and nonenzymatic components. Twomain oxidative stress responses of bacteria are the peroxidestress response and the superoxide stress response. Althoughtotally independent of each other, both are accompanied byincreased DNA repair capacity [22]. The peroxide stress re-sponse to challenge by H2O2 is characterized by elevatedconcentrations of about 30 proteins above the basal level. Alevel of resistance is acquired by cells through preexposure to

H2O2, leading to enhanced survival ratios, as shown in Fig. 7.The superoxide stress response to O2•− is also mediated byelevated levels of about 30 proteins, which are mostly differentto those in the peroxide stress response. Again, preexposureleads to enhanced survival rates by means of acquired resistanceand repair capacity. The two mechanisms being independentconfer no cross-resistance, so that preexposure to H2O2 doesnot enhance survival ratios on exposure to O2•− and visa versa.

Probably the most important group of enzymes employed bybacteria to defend against oxidation are the SODs. Superoxideradicals are formed in small amounts during normal bacterialrespiration and by virtually all aerobic cells. O2•− is so toxicthat all organisms attempting to grow in atmospheric oxygenproduce SOD to neutralize it. Aerobic bacteria and facultativeanaerobes (growing aerobically) produce SOD, with whichthey convert superoxide into molecular oxygen and hydrogenperoxide according to (2).

The steady-state concentration of O2•− is about 10−9 to10−10 M in wild-type aerobically growing E. coli. In mutantstrains, which lack SOD activity, the concentration is muchhigher at around 5 × 10−6 M [91]. The presence of SOD in acell is therefore thought to reduce the steady-state concentrationby up to three orders of magnitude.

Based on the metal ligands there are three types of SOD:CuZnSOD, FeSOD, and MnSOD. FeSOD is primarily foundin prokaryotic cells, MnSOD is found in both prokaryotesand eukaryotes, while CuZnSOD is generally not found inbacteria [21], [22]. The rate constant for MnSODs is about1.8 × 109 m−1s−1 at pH 7.8 and decreases as pH becomesmore alkaline. In this respect, it is unlike the CuZnSODs ofeukaryotes, which catalyze the same reaction, because the rateconstant increases with alkalinity [21].

One of the products of superoxide dismutation, hydrogenperoxide, is itself a reactive species. Although thought to beless toxic than superoxide, hydrogen peroxide is nonethelessnoxious, and therefore bacteria have developed enzymes toneutralize it. The most familiar of these is catalase, whichconverts hydrogen peroxide into water and oxygen (5). Inbacteria, catalase destroys hydrogen peroxide with remarkablerapidity [22]. The reaction is a disproportionation (i.e., anoxidation–reduction) reaction, in which hydrogen peroxide isthe electron source. Consequently, the reaction does not requirean exogenous reducing source. The reaction is also exothermic,meaning that catalase can protect against the action of hydrogenperoxide even in energy-depleted cells

2H2O2 → 2H2O + O2. (5)

Bacteria also utilize peroxidases to neutralize hydrogenperoxide. Unlike catalase, however, peroxidases require nicoti-namide adenine dinucleotide (NADH) or nicotinamide adeno-sine dinucleotide phosphate (NADPH) as an electron source toreduce H2O2 to water.

On the reparative side, bacteria possess catalysts involved inthe repair of some protein and DNA damage. Bacteria possesscatalysts that are able to repair some covalent modificationsto the primary structure of proteins. For example, oxidationof the amino acid methionine to methionine sulfoxide can be



repaired by methionine sulfoxide reductase [78], [92]. A num-ber of DNA repair enzymes are also induced by oxidative stressin bacteria (reviewed in [93]). They are primarily involvedin two types of repair: base excision repair and nucleotideexcision repair [75], [94]. Base excision repair is achievedthrough the action of DNA glycosylase, which recognizes dam-aged bases and cleaves their respective glycosylic bonds [75].Nucleotide excision repair differs from base excision repairinsomuch that several consecutive nucleotide bases (includingany oxidized bases) are removed in a single action. Nucleotideexcision repair is facilitated by the enzymes endonuclease andexonuclease [75].

Repair of membrane lipids is less well documented. It ispossible that some repair may be effected by reduction of fattyacid hydroperoxides [22].

The discussion above has so far focused on the phenotypicalresponse of cells to oxidative stress. At a genomic level, bacte-ria possess sets of genes that regulate response to environmentalstresses. At least 13 different multigene systems are known tobe induced by various stress stimuli [22]. Although phenotyp-ically the peroxide and superoxide stress response systems areinterrelated, at a genomic level the two systems appear to beseparate and distinct. The expression of H2O2-induced defenseproteins is mediated by the transcriptional activator OxyR.Superoxide stress induces the production of defense proteins,which are for the most part different from those of the peroxidestress response. Positive regulation of these is mediated by thetwo-stage sox RS system.

VII. UV DAMAGE AND REPAIR IN BACTERIA

UV radiation has well-known bactericidal properties. Whenphotons of UV light strike biological cells, the energy in thephotons is absorbed by chromophores at discrete wavelengths.The biological impact of UV radiation is primarily due to theabsorption of photons by nucleic acids. DNA has an absorptionspectrum with a maximum in the region 260–265 nm and whichrapidly declines toward longer wavelengths [94]. Specific UVdisinfection devices emit upward of 86% of their light at254 nm, to coincide with this germicidal peak as closely aspossible [96]. Many researchers have demonstrated that whenUV light at 253.7 nm is absorbed by nucleic acids, pyrimidinelesions are readily formed [97]. Although purine bases are alsostrong absorbers of the UV photons, the quantum yield requiredto create dimmers is an order of magnitude higher for purinesthan pyrimidine bases [98]. While other photoproducts may becreated through UV irradiation, pyrimidine dimers are gener-ally considered to be the most important class. They are formedin double stranded DNA when two adjacent pyrimidine basesfuse together to form a dimer. Three types of pyrimidine dimerscan be formed, thymine–thymine (T <> T), thymine–cytosine(T <> C), and cytosine–cytosine (C <> C) dimers. DuringUV irradiation the frequency with which each type is formed,depends on the DNA base composition of the irradiatedbacteria [95].

Bacteria possess a range of DNA repair systems, permittingrapid recovery from sublethal UV damage [70]. In bacteria, UV-induced DNA lesions can be repaired by four main processes:

photoreactivation, nucleotide excision repair, recombinationrepair, and SOS repair (i.e., error-prone repair). Prokaryotesemploy are range of enzymes to facilitate these various repairmechanisms. Although not all bacteria species exhibit photore-pair mechanisms [97], photoreactivation is one of the principalmechanisms used by many species to repair DNA damage.Bacterial species which exhibit photoreactivation posses an en-zyme, DNA photolyase, which absorbs photons of visible lightto facilitate DNA repair [97]. Consequently, these species repairDNA damage much more efficiently under light conditions thanthey do in the dark, where excision repair is thought to be theprinciple mechanism employed. In bacteria, the various repairmechanisms are generally very efficient and rapid, meaning thatcell death only occurs when the UV photon hits are so numerousthat the bacterial repair mechanisms are overwhelmed.

In low-pressure plasma, UV radiation is thought to be themost important factor in sterilization, but in the atmosphericpressure plasma under consideration here, UV radiation gener-ated is reabsorbed such that lethal doses are not delivered. Forexample, the UV component of an atmospheric glow dischargewas not considered a significant agent of lethality because thedecontamination efficiency of the discharge was significantlygreater than that arising from the same duration of UV expo-sure [12]. In another report, the role of UV was discountedon the basis of killing kinetics. Similar killing kinetics wererecorded from an atmospheric glow discharge when Staphylo-coccus aureus was exposed wrapped in semipermeable bagsand unwrapped. As the porous bag blocked UV radiation yetallowed the passage of small reactive species, UV exposure wasassumed not to contribute to cell death [17].

VIII. OZONE DAMAGE OF BACTERIA

Ozone is a powerful oxidizing agent, which readily formsreactive OH and HOO radicals in aqueous solution. By allaccounts, ozone disinfection is a complex heterogeneous phe-nomenon, the mechanism of which remains little understooddespite decades of research effort. Ozone itself reacts withmany biomolecules. Protein damage is caused by reactions withdienes, amines, and thiols [22], and by cross-linking tyrosineresidues by oxidizing the –OH groups [21]. Cell wall targetssuch as fatty acids and peptides appear among the most likelytargets for gaseous ozone [99], [100]. Lipid peroxidation canbe stimulated by oxidation by ozone, resulting in a reduction ofmembrane fluidity. Single-strand DNA breaks have also beenobserved, which unrepaired can lead to extensive DNA damageand death [101], [102].

Disinfection of E. coli in wastewater with ozone has beenstudied in various semibatch and continuous-flow configura-tions. Bactericidal concentrations are in the 0.1–0.2 ppm range[103]. Although cell lysis is often observed following pro-longed ozoneation, is not necessarily a characteristic of theprocess [26]. Bacterial inactivation by ozone is a function ofozone concentration per viable bacteria, meaning that below athreshold concentration ozone has no effect on survival. Thiscan be complicated by the presence of other compounds withwhich ozone can react [26].



IX. CONCLUSION

In this paper, the role of oxidative stress and damage by ROSin bacterial neutralization by air plasma has been considered.There is strong evidence from research into physiological expo-sures that the ROS formed in nonthermal plasma at atmosphericpressure would cause various forms of cellular damage. Ifunrepaired, such damage could be lethal. The concentration ofROS and the cocktail of species implicated during gas dischargeexposure differ considerably from that of even the most extremephysiological conditions. Only further research will elucidatethe extent to which the reactions discussed here are involved inthe bactericidal action.

Biological and engineering challenges are still to be over-come in the development of effective atmospheric pressuredischarges for decontamination and disinfection applications.On the biological side, it remains critical to understand thepathway of damage leading to death for bacteria, viruses, andfungi. Without this knowledge, optimization of the decontami-nation process is reduced to an empirical affair. In elucidatingthe processes leading to cell death, the actual cause must beseparated from any physical changes that occur postmortem.One approach would be to study the sublethal effects of plasmaon cells, while the effects on individual cellular componentssuch as cell membranes, nucleic acids, proteins, and enzymescould also be considered. Furthermore, the relative effects ofplasma exposure on microbial and human tissues should be fur-ther compared where applications in patient care are consideredsuch as for dentistry and wound treatment.

On the chemical side, a deeper appreciation regarding therelative importance of the various reactive species involvedin bacterial neutralization is urgently needed. More detailedreporting on the nature and quantity of reactive species towhich test organisms are exposed would greatly assist in this.Although it is intuitive that the most reactive species or the mostabundant ROS would be influential, this may not necessarilybe so. The importance of understanding the role of specificreactive species in cell killing is highlighted by the case ofhydroxyl radicals. These are significantly implemented in thekilling of microorganisms, yet many plasma are operated indry air, where OH production is minimized. It can therefore behypothesized that the use of humidified air in plasma reactors,where this is possible, may promote bactericidal efficiencies.Implication of a reactive species in cell damage and death isimportantly not restricted to those species directly present ingas discharge effluent. The absence of hydroxyl radicals ingas discharge effluent does not preclude its involvement in thekilling process. Superoxide radicals in the effluent will reactwith moisture in and around a bacterium to form hydrogen per-oxide. This can form hydroxyl radicals on the surface of cellsor diffuse through the membrane and cytoplasm to form hy-droxyl radicals in proximity to nucleic acids, ultimately causingDNA damage.

A role for specific ROS in bactericidal activity could be im-plicated by demonstrating the presence of the reactive speciesat the site of injury. This can be done either by establishing thatthe time course of formation of the reactive species is consistent

with the time course of cell injury, by demonstrating that directapplication of relevant concentrations of the reactive speciesover a comparable time period reproduces most or all of thedamage observed or by demonstrating that the extent of injuryis diminished by removal of or inhibition of the formation ofthe reactive.

Use could also be made of experimental bacterial strains thatmay exhibit different phenotypical responses to the gas dis-charge. For example, SOD-deficient mutants of E. coli wouldbe more susceptible to gas discharge exposure than wild-typestrains if superoxide were a significant reactive species. Like-wise, up regulation of superoxide or hydrogen peroxide stressresponses (defense and repair) by preexposure to nonlethaldoses should confer a degree of protection. Indeed, studiesof the effects of sublethal exposures are likely to shed morelight on killing mechanisms because the complete pathway ofevents may be identified and not masked by damage causedafter the cells are actually dead. For this reason, it is importantto quantify the plasma dose required to achieve lethality invarious bacteria species. Some species and strains may bemore resistant than others to damage caused by ROS. Duringa disinfection process it is essential that pathogenic microbialspecies receive a lethal dose; otherwise the oxidative damagemay be repaired in time, invalidating the disinfection process.

It is clear that a truly multidisciplinary approach is neededin order to fully understand the biophysical and biochemicalprocesses. Evidence strongly suggests that the vulnerability ofcells and microorganisms to oxidation lies at the root of themechanism for cell death during exposure to nonthermal gasdischarges at atmospheric pressure. Further studies will showwhether it is possible for oxidative stress to be exploited as aweapon for biological decontamination.

REFERENCES

[1] J. Elster and H. Geitel, Wein Ber., vol. 94, 1890.[2] A. P. Krueger and E. J. Reed, “Biological impact of small air ions,”

Science, vol. 193, no. 4259, pp. 1209–1213, Sep. 1976.[3] A. P. Wehner, “Air ions: Physical and biological aspects,” in History of

Air Ion Research, J. M. Charry and R. I. Kavet, Eds. Boca Raton, FL:CRC, 1987, ch. 9, pp. 1–539.

[4] A. L. Tchijevsky, Aeroionization: Its Role in the National Economy(Russia). Washington, DC: U.S. Office Nav. Intell., 1960.

[5] W. Siemens, “Ueber die elektrostatische induction und die verzögerungdes stroms in flaschendräten,” Ann. Phys. Chem., vol. 102, p. 66, 1857.

[6] B. R. Gadri, J. R. Roth, T. C. Montie, K. Kelly-Wintenberg, P. P. Y. Tsai,D. J. Helfritch, P. Feldman, D. M. Sherman, F. Karakaya, and Z. Y. Chen,“Sterilization and plasma processing of room temperature surfaces witha one atmosphere uniform glow discharge plasma (OAUGDP),” Surf.Coat. Technol., vol. 131, no. 1–3, pp. 528–542, Sep. 2000.

[7] T. C. Montie, K. Kelly-Wintenberg, and J. R. Roth, “An overviewof research using the one atmosphere uniform glow discharge plasma(OAUGDP) for sterilisation of surfaces and materials,” IEEE Trans.Plasma Sci., vol. 28, no. 1, pp. 41–50, Feb. 2000.

[8] M. Laroussi, “Nonthermal decontamination of biological media byatmospheric-pressure plasma: Review, analysis and prospects,” IEEETrans. Plasma Sci., vol. 30, no. 4, pp. 1409–1415, Aug. 2002.

[9] M. Laroussi, K. H. Schoenbach, U. Kogelschatz, R. J. Vidmar, S. Kuo,M. Schmidt, J. F. Behnke, K. Yukimura, and E. Stoffels, “Applicationsof atmospheric pressure air plasma,” in Non-Equilibrium Air Plasma atAtmospheric Pressure, K. H. Becker, U. Kogelschatz, K. H. Schoenbach,and R. J. Barker, Eds. Bristol, U.K.: IOP, 2004, ch. 9, pp. 537–671.

[10] M. Moisan, J. Barbeau, M. C. Crevier, J. Pelletier, N. Philip, andB. Saoudi, “Plasma sterilization. Methods mechanisms,” Pure Appl.Chem., vol. 74, no. 3, pp. 349–358, Mar. 2002.



[11] N. Philip, B. Saoudi, B. C. Crevier, M. Moisan, J. Barbeau, andJ. Pelletier, “The respective roles of UV photons and oxygen atomsin plasma sterilization at reduced gas pressure: The case of N2–O2

mixtures,” IEEE Trans. Plasma Sci., vol. 30, no. 4, pp. 1429–1436,Aug. 2002.

[12] M. Laroussi, “Sterilization of contaminated matter with an atmosphericpressure plasma,” IEEE Trans. Plasma Sci., vol. 24, no. 3, pp. 1188–1191, Jun. 1996.

[13] H. W. Herrmann, I. Henins, J. Park, and G. S. Selwyn, “Decontaminationof chemical and biological warfare (CBW) agents using an atmosphericpressure plasma jet (APPJ),” Phys. Plasma, vol. 6, no. 5, pp. 2284–2289,May 1999.

[14] K. Kelly-Wintenberg, A. Hodge, T. C. Montie, L. Deleanu, D. Sherman,J. R. Roth, P. Tsai, and L. Wadsworth, “Use of a one atmosphere uniformglow discharge plasma to kill a broad spectrum of microorganisms,”J. Vac. Sci. Technol. A, Vac. Surf. Films, vol. 17, no. 4, pp. 1539–1544,Jul. 1999.

[15] M. Vleugels, G. Shama, X. T. Deng, E. Greenacre, T. Brocklehurst,and M. G. Kong, “Atmospheric plasma inactivation of biofilm-formingbacteria for food safety control,” IEEE Trans. Plasma Sci., vol. 33, no. 2,pp. 824–828, Apr. 2005.

[16] M. Laroussi, I. Alexeff, and W. L. Wang, “Biological decontaminationby nonthermal plasma,” IEEE Trans. Plasma Sci., vol. 28, no. 1, pp. 184–188, Feb. 2000.

[17] K. Kelly-Wintenberg, T. C. Montie, C. Brickman, R. J. Roth, A. K.Carr, K. Sorge, L. C. Wadsworth, and P. P. Y. Tsai, “Room tempera-ture sterilization of surfaces and fabrics with a one atmosphere uniformglow discharge plasma,” J. Ind. Microbiol. Biotechnol., vol. 20, no. 1,pp. 69–74, Jan. 1998.

[18] M. Laroussi, J. P. Richardson, F. F. Dyer, and F. C. Dobbs, “Biochemicalpathways in the interaction of non-equilibrium plasma with bacteria,” inProc. Electromed., Portsmouth, VA, 2001, pp. 33–34.

[19] ——, “Effects of non-equilibrium atmospheric pressure plasma on theheterotrophic pathways of bacteria and on their cell morphology,” Appl.Phys. Lett., vol. 81, no. 4, pp. 772–774, Jul. 2002.

[20] D. A. Mendis, M. Rosenberg, and F. Azam, “A note on the possibleelectrostatic disruption of bacteria,” IEEE Trans. Plasma Sci., vol. 28,no. 4, pp. 1304–1306, Aug. 2000.

[21] B. Halliwell and J. M. C. Gutteridge, Free Radicals in Biology andMedicine. Oxford, U.K.: Clarendon, 1985.

[22] S. B. Farr and T. Kogoma, “Oxidative stress responses in Escherichiacoli and Salmonella typhimurium,” Microbiol. Rev., vol. 55, no. 4,pp. 561–585, Dec. 1991.

[23] P. A. Hyslop, D. B. Hinshaw, I. U. Scraufstatter, C. G. Cochrane,S. Kunz, and K. Vosbeck, “Hydrogen peroxide as a potent bacteriosta-tic antibiotic: Implications for host defense,” Free Radic. Biol. Med.,vol. 19, pp. 31–37, Jul. 1995.

[24] J. A. Imlay, “Pathways of oxidative damage,” Ann. Rev. Microbiol.,vol. 57, pp. 395–418, 2003.

[25] E. W. Kellogg, M. G. Yost, N. Barthakur, and A. P. Kreuger, “Superoxideinvolvement in the bactericidal effects of negative air ions on Staphylo-coccus albus,” Nature, vol. 281, no. 5730, pp. 400–401, Oct. 1979.

[26] N. K. Hunt and B. J. Marinas, “Inactivation of Escherichia coli withozone: Chemical and inactivation kinetics,” Water Res., vol. 33, no. 11,pp. 2633–2641, Aug. 1999.

[27] M. Laroussi, D. A. Mendis, and M. Rosenberg, “Plasma interaction withmicrobes,” New J. Phys, vol. 5, no. 1, pp. 41.1–41.10, Apr. 2003.

[28] U. Kogelschatz, “Filamentary, patterned and diffuse barrier discharges,”IEEE Trans. Plasma Sci., vol. 30, no. 4, pp. 1400–1408, Aug. 2002.

[29] H. Conrads and M. Schmidt, “Plasma generation and plasma sources,”Plasma Sources Sci. Technol., vol. 9, no. 4, pp. 441–454, Nov. 2000.

[30] G. E. Georghiou, A. P. Papadakis, R. Morrow, and A. C. Metaxas,“Numerical modelling of atmospheric pressure gas discharges lead-ing to plasma production,” J. Phys. D, Appl. Phys., vol. 38, no. 20,pp. R303–R328, Oct. 2005.

[31] M. Kogoma and S. Okazaki, “Raising of ozone formation efficiency in ahomogeneous glow discharge plasma at atmospheric pressure,” J. Phys.D, Appl. Phys., vol. 27, no. 9, pp. 1985–1987, Sep. 1994.

[32] E. M. van Veldhuizen and W. R. Rutgers, “Pulsed positive coronasteamer propagation and branching,” J. Phys. D, Appl. Phys., vol. 35,no. 17, pp. 2169–2179, Sep. 2002.

[33] K. Zhang, B. Eliasson, and U. Kogelshatz, “Direct conversion of green-house gases to synthesis gas and C4 hydrocarbons over Zeolite HY pro-moted by a dielectric barrier discharge,” Ind. Eng. Chem. Res., vol. 41,no. 6, pp. 1462–1468, 2002.

[34] T. Hammer, “Applications of plasma technology in environmental tech-niques,” Control Plasma Phys., vol. 39, pp. 441–462, 1999.

[35] W. Manheimer, L. E. Sugiyama, and T. H. Stix, Eds., Plasma Scienceand the Environment. Woodbury, NY: Amer. Inst. Phys., 1997.

[36] D. Braun, V. Gibalov, and G. Pietsch, “Two-dimensional modelling ofthe dielectric barrier discharge in air,” Plasma Sources Sci. Technol.,vol. 1, no. 3, pp. 166–174, Aug. 1992.

[37] V. I. Gibalov and G. J. Pietsch, “The development of dielectric barrierdischarges in gas gaps and on surfaces,” J. Phys. D, Appl. Phys., vol. 33,no. 20, pp. 2618–2636, Oct. 2000.

[38] ——, “Dynamics of dielectric barrier discharges in coplanar arrange-ments,” J. Phys. D, Appl. Phys., vol. 37, no. 15, pp. 2082–2092,Aug. 2004.

[39] W. S. Kang, J. M. Park, Y. Kim, and S. H. Hong, “Numerical studyon influences of barrier arrangements on dielectric barrier dischargecharacteristics,” IEEE Trans. Plasma Sci., vol. 31, no. 4, pp. 504–510,Aug. 2003.

[40] U. Kogelschatz, “Industrial innovations based on fundamental physics,”Plasma Sources Sci. Technol., vol. 11, no. 3A, pp. A1–A6, Aug. 2002.

[41] ——, “Dielectric-barrier discharges: Their history, discharge physics andindustrial applications,” Plasma Chem. Plasma Process., vol. 23, no. 1,pp. 1–46, Mar. 2003.

[42] N. Y. Babaeva and G. V. Naidis, “Modelling of streamer propaga-tion,” in Electrical Discharges for Environmental Purposes, E. M. vanVeldhuizen, Ed. Commack, NY: Nova, 2000, pp. 21–48.

[43] ——, “Two-dimensional modelling of positive streamer dynamics innon-uniform electric fields in air,” J. Phys. D, Appl. Phys., vol. 29,pp. 2423–2431, 1996.

[44] M. S. Benilov and G. V. Naidis, “Modelling of low-current dischargesin atmospheric-pressure air taking account of non-equilibrium effects,”J. Phys. D, Appl. Phys., vol. 36, no. 15, pp. 1834–1841, Aug. 2003.

[45] G. E. Georghiou, R. Morrow, and A. C. Metaxas, “The two-dimensionalsimulation of streamers using the FE-FCT method,” J. Phys. D, Appl.Phys., vol. 33, no. 3, pp. L27–L32, Feb. 2000.

[46] G. E. Georghiou, A. P. Papadakis, R. Morrow, and A. C. Metaxas,“Numerical modelling of atmospheric pressure gas discharges leading toplasma production,” J. Phys. D, Appl. Phys, vol. 38, no. 20, pp. R303–R328, Oct. 2005.

[47] F. Massines, P. Segur, N. Gherardi, C. Khamphan, and A. Ricard,“Physics and chemistry in a glow dielectric barrier discharge at at-mospheric pressure: Diagnostics and modelling,” Surf. Coat. Technol.,vol. 174/175, pp. 8–14, Sep./Oct. 2003.

[48] B. Eliasson and U. Kogelschatz, “Modelling and applications of silentdischarge plasmas,” IEEE Trans. Plasma Sci., vol. 19, no. 2, pp. 309–323, Apr. 1991.

[49] R. Dorai and M. J. Kushner, “A model for plasma modification ofpolypropylene using atmospheric pressure discharges,” J. Phys. D, Appl.Phys., vol. 36, no. 6, pp. 666–685, Mar. 2003.

[50] F. Massines, A. Rabehi, P. Decomps, R. B. Gadri, P. Segur, andC. Mayoux, “Experimental and theoretical study of a glow dischargeat atmospheric pressure controlled by dielectric barrier,” J. Appl. Phys.,vol. 83, no. 6, pp. 2950–2957, Mar. 1998.

[51] J. J. Shi and M. G. Kong, “Cathode fall characteristics in a DC at-mospheric pressure glow discharge,” J. Appl. Phys., vol. 94, no. 9,pp. 5504–5513, Nov. 2003.

[52] J. M. Meek and J. D. Craggs, Electrical Breakdown of Gases. Hoboken,NJ: Wiley, 1978.

[53] H. Raether, Electron Avalanches and Breakdown in Gasses. London,U.K.: Butterworth, 1964.

[54] S. Okazaki, M. Kogoma, M. Uehara, and Y. Kimura, “Appearance of astable glow discharge in air, argon, oxygen and nitrogen at atmosphericpressure using a 50 Hz source,” J. Phys. D, Appl. Phys., vol. 26, no. 5,pp. 889–892, May 1993.

[55] H. Raether, “Die Entwicklung der Electronenlawine in denFunkenkanal,” Z. Phys., vol. 112, pp. 464–489, 1939.

[56] L. B. Loeb and J. M. Meek, “The mechanism of spark discharge in airat atmoshpheric pressure,” J. Appl. Phys., vol. 11, no. 7, pp. 438–447,Jul. 1940.

[57] A. Schütze, J. Y. Jeong, S. E. Babayan, J. Park, G. S. Selwyn, andR. F. Hicks, “The atmospheric-pressure plasma jet: A review andcomparison to other plasma sources,” IEEE Trans. Plasma Sci., vol. 26,no. 6, pp. 1685–1694, Dec. 1998.

[58] M. Shibata, N. Nakano, and T. Makabe, “Effect of O2(a1∆g) on plasma

structures in oxygen radio frequency discharges,” J. Appl. Phys., vol. 80,no. 11, pp. 6142–6147, Dec. 1996.

[59] K. Becker, M. Schmidt, A. A. Viggiano, R. Dressler, and S. Williams,“Air plasma chemistry,” in Non-Equilibrium Air Plasma at AtmosphericPressure, K. H. Becker, U. Kogelschatz, K. H. Schoenbach, and R. J.Barker, Eds. Bristol, U.K.: IOP, 2004, ch. 4, pp. 124–182.



[60] F. Pontiga, C. Soria, and A. Castellanos, “Electrical and chemical modelof negative corona in oxygen at atmospheric pressure,” J. Electrostat.,vol. 40/41, pp. 115–120, Jun. 1997.

[61] B. Eliasson, M. Hirth, and U. Kogelschatz, “Ozone synthesis from oxy-gen in dielectric barrier discharges,” J. Phys. D, Appl. Phys., vol. 20,no. 11, pp. 1421–1437, Nov. 1987.

[62] J. Y. Jeong, J. Park, I. Henins, S. E. Babayan, V. Tu, G. S. Selwyn,G. Ding, and F. R. Hicks, “Reaction chemistry in the afterglow of anoxygen–helium atmospheric pressure plasma,” J. Phys. Chem., vol. 104,no. 34, pp. 8027–8032, 2000.

[63] K. R. Stalder, D. F. McMillen, and J. Woloszko, “Electrosurgicalplasma,” J. Phys. D, Appl. Phys., vol. 38, no. 11, pp. 1728–1838,Jun. 2005.

[64] J. J. Coogan and A. D. Sappey, “Distribution of OH within silentdischarge plasma reactors,” IEEE Trans. Plasma Sci., vol. 24, no. 1,pp. 91–92, Feb. 1996.

[65] Z. Falkenstein and J. J. Coogan, “Microdischarge behaviour in the silentdischarge of nitrogen–oxygen and water–air mixtures,” Phys. D, Appl.Phys., vol. 30, no. 5, pp. 817–825, Mar. 1997.

[66] N. I. Goldstein, R. N. Goldstein, and M. N. Merzlyak, “Negative air ionsas a source of superoxide,” Int. J. Biometeorol., vol. 36, no. 2, pp. 118–122, 1992.

[67] M. N. Kondrashove, E. V. Grigorenko, A. V. Tikhonov, T. V. Sirota,A. V. Temnov, I. G. Stavrovskaja, N. I. Kosyakova, N. V. Lange, andV. P. Tikhonov, “The primary physico-chemical mechanism for the bene-ficial biological/medical effects of negative air ion,” IEEE Trans. PlasmaSci., vol. 28, no. 1, pp. 230–237, Feb. 2000.

[68] O. Challenger, J. Braven, D. Harwood, D. Rosen, and G. Richardson,“Negative air ionisation and the generation of hydrogen peroxide,” Sci.Total. Environ., vol. 177, no. 1, pp. 215–219, Jan. 1996.

[69] I. Digel, T. Demirci, A. T. Artmann, K. Nishikawa, D. Porst, andG. M. Artmann, “Free radical nature of the bactericidal effect of plasma-generated cluster ions (PCIs),” in Proc. Int. Symp. Cellular Eng. andNanosensors, Sep. 2004, pp. 982–983.

[70] G. Richardson, S. A. Eick, D. Harwood, K. G. Rosen, and F. Dobbs,“Negative air ionisation and the production of hydrogen peroxide,”Atmos. Environ., vol. 37, no. 26, pp. 3701–3706, Aug. 2003.

[71] B. Halliwell, J. M. C. Gutteridge, and C. E. Cross, “Free radicals, an-tioxidants and human disease—Where are we now,” J. Lab. Clin. Med.,vol. 19, no. 6, pp. 598–620, Jun. 1992.

[72] D. K. Bol and R. E. Yasbin, “Characterization of an inducible oxidativestress system in Bacillus subtilis,” J. Bacteriol., vol. 172, no. 6, pp. 3503–3506, Jun. 1990.

[73] B. M. Babior, “The respiratory burst of phagocytes,” J. Clin. Invest.,vol. 73, no. 3, pp. 599–601, Mar. 1984.

[74] T. Seres, R. G. Knickelbein, J. B. Warshaw, and R. B. Johnston, “Thephagocytosis-associated respiratory burst in human monocytes is associ-ated with increased uptake of glutathione,” J. Immunol., vol. 165, no. 6,pp. 3333–3340, Sep. 2000.

[75] E. S. Henle and S. Linn, “Formation, prevention and repair of DNAdamage by iron/hydrogen peroxide,” J. Biol. Chem., vol. 272, no. 31,pp. 19 095–19 098, Aug. 1997.

[76] M. Anbar and P. Neta, “Compilation of specific biomolecular rateconstants for the reactions of hydrated electrons, hydrogen atoms andhydroxyl radicals with inorganic and organic compounds in aqueoussolution,” Int. J. Appl. Radiat. Isot., vol. 18, no. 7, pp. 493–523,Jul. 1967.

[77] A. Singh and H. Singh, “Time scale and nature of radiation biologicaldamage—Approaches to radiation protection and post irradiation ther-apy,” Prog. Biophys. Mol. Biol., vol. 39, no. 2, pp. 69–107, 1982.

[78] E. Cabiscol, J. Tamarit, and J. Ros, “Oxidative stress in bacteria andprotein damage by reactive oxygen species,” Int. Microbiol., vol. 3, no. 1,pp. 3–8, Mar. 2000.

[79] J. Mead, “Free radical mechanisms of lipid damage and consequencesfor cellular membranes,” in Free Radicals in Biology, W. A. Pryor, Ed.New York: Academic, 1976, pp. 51–68.

[80] R. McElhaney, “The effects of membrane lipids on permeability andtransport in prokaryotes,” in Structure and Properties of Cell Mem-branes, G. Benga, Ed. Boca Raton, FL: CRC, 1985, pp. 75–91.

[81] K. M. Humphries and L. I. Szweda, “Selective inactivation of α-ketoglutarate dehydrogenase: Reaction of lipoic acid with 4-hydroxy-2-nonenal,” Biochemistry, vol. 37, no. 45, pp. 15 835–15 841, Nov. 1998.

[82] B. S. Berlett and E. R. Stadtman, “Protein oxidation in aging, diseaseand oxidative stress,” J. Biol. Chem., vol. 272, no. 33, pp. 20 313–20 316,Aug. 1997.

[83] E. R. Stadtman, “Protein oxidation and aging,” Science, vol. 257,no. 5074, pp. 1220–1224, Aug. 1992.

[84] P. R. Gardner and I. Fridovich, “Superoxide sensitivity of the Escherichiacoli 6-phosphogluconate dehydratase,” J. Biol. Chem., vol. 266, no. 3,pp. 1478–1483, Jan. 1991.

[85] R. T. Dean, S. Fu, R. Stocker, and M. J. Davies, “Biochemistry andpathology of radical-mediated protein oxidation,” Biochem. J., vol. 324,no. 1, pp. 1–18, May 1997.

[86] T. Grune, T. Reinheckel, and K. J. A. Davies, “Degradation of oxidizedproteins in mammalian cells,” FASEB J., vol. 11, no. 7, pp. 526–534,1997.

[87] L. J. Marnett, “Oxyradicals and DNA damage,” Carcinogenesis, vol. 21,no. 3, pp. 361–370, Mar. 2000.

[88] D. Segerback, “Alkylation of DNA and hemoglobin in the mouse follow-ing exposure to ethene and ethene oxide,” Chem. Biol.-Interact., vol. 45,no. 2, pp. 139–151, Jul. 1983.

[89] F. W. Summerfield and A. L. Tappel, “Determination by fluorescencequenching June of the environment of DNA crosslinks made by mal-ondialdehyde,” Biochem. Biophys. Acta., vol. 740, no. 2, pp. 185–189,1983.

[90] B. Halliwell and M. Whiteman, “Measuring reactive species and oxida-tive damage in vivo and in cell culture: How should you do it and whatdo the results mean?,” Br. J. Pharmacol., vol. 142, no. 2, pp. 231–255,May 2004.

[91] J. A. Imlay and I. Fridovich, “Assay of metabolic superoxide productionin Escherichia coli,” J. Biol. Chem., vol. 266, no. 11, pp. 6957–6965,Apr. 1991.

[92] G. S. John, N. Brot, J. Ruan, H. Erdjument-Bromage, P. Tempst, H.Weissbach, and C. Nathan, “Peptide methionine sulfoxide reductasefrom Escherichia coli and Mycobacterium tuberculosis protects bacteriaagainst oxidative damage from reactive nitrogen intermediates,” in Proc.Nat. Acad. Sci. U.S.A., Aug. 2001, vol. 98, pp. 9901–9906.

[93] B. Demple and L. Harrison, “Repair of oxidative damage to DNA:Enzymology and biology,” Annu. Rev. Biochem., vol. 63, pp. 915–948,1994.

[94] D. L. Croteau and V. A. Bohr, “Repair of oxidative damage to nuclearand mitochondrial DNA in mammalian cells,” J. Biol. Chem., vol. 272,no. 41, pp. 25409–25412, Oct. 1997.

[95] W. Harm, Biological Effects of Ultraviolet Radiation. Cambridge,U.K.: Cambridge Univ. Press, 1980, ch. 3.

[96] C. A. Lawrence and S. S. Block, Disinfection, Sterilization and Preser-vation. Philadelphia, PA: Lea & Febinger, 1968.

[97] C. B. Beggs, “A quantitative method for evaluating the photoreactivationof ultraviolet damaged microorganisms,” Photochem. Photobiol. Sci.,vol. 1, no. 6, pp. 431–437, Jun. 2002.

[98] J. Jagger, Introduction to Research in Ultraviolet Photobiology. Engle-wood Cliffs, NJ: Prentice-Hall, 1967, ch. 4.

[99] W. Doroszkiewicz, I. Sikorska, and S. Jankowski, “Studies of the in-fluence of ozone on complement-mediated killing of bacteria,” FEMSImmunol. Med. Microbiol., vol. 9, no. 4, pp. 281–285, Oct. 1994.

[100] W. A. Pryor, M. M. Dooley, and D. F. Church, “Mechanisms of thereactions of ozone with biological molecules: The source of toxic effectsof ozone,” in Advances in Modern Environmental Toxicology, M. G.Mustafa and M. A. Mehlman, Eds. Ann Arbor, MI: Ann Arbor Sci.,1983.

[101] C. Hamelin, F. Sarhan, and Y. S. Chung, “Induction of deoxyribonucleicacid degradation in Escherichia coli by ozone,” Experentia, vol. 34,no. 12, pp. 1578–1579, Dec. 1978.

[102] K. Ishizaki, K. Sawadaishi, K. Miura, and N. Shinriki, “Effect of ozoneon plasmid DNA of Escherichia coli in situ,” Water. Res., vol. 21, no. 7,pp. 823–828, Jul. 1987.

[103] W. T. Broadwater, R. C. Hoehn, and P. H. King, “Sensitivity of threeselected bacterial species to ozone,” Appl. Microbiol., vol. 26, no. 3,pp. 391–393, Sep. 1973.

Lindsey F. Gaunt was born in East Sussex, U.K., in1974. She received the B.Sc. degree in biology andthe Ph.D. degree in applied electrostatics, both fromthe University of Southampton, Southampton, U.K.,in 1995 and 2000, respectively.

She joined the School of Electronics and Com-puter Science, University of Southampton in 1995and now holds a post as School Research Fellow.Her current research interests include bioelectro-statics, the biological effects of electric fields anddischarges, and electrical disinfection methods.

Dr. Gaunt is a member of the Institute of Physics (IOP) and HonorarySecretary of the Electrostatics Group of the IOP.



Clive B. Beggs received the B.Sc. degree in buildingengineering from Leeds Polytechnic, Leeds, U.K., in1981, the M.Phil. degree in refrigeration engineeringfrom the University of Sheffield, Sheffield, U.K.,in 1991, and the Ph.D. degree in energy engineer-ing from DeMontfort University, Leicester, U.K.,in 1995.

He is currently Professor of medical technology atthe University of Bradford, Bradford, U.K., and wasformally, in 1996–2005, a Senior Lecturer in aerobi-ological engineering at the University of Leeds. He

has spent many years researching the use of engineering methods to control thespread of infection and has investigated the effects of ultraviolet light and airions on airborne bacteria. He has a particular interest in the epidemiology ofinfection and the strategic use of interventions to break transmission pathways.

George E. Georghiou received the B.A., M.Eng.,M.A., and Ph.D. degrees from the University ofCambridge, Cambridge, U.K., in 1995, 1996, 1997,and 1999, respectively.

He is currently an Assistant Professor at theDepartment of Electrical and Computer Engineer-ing, University of Cyprus, Cyprus. Prior to this, hewas the undergraduate course leader in electricalengineering at the Department of Electronics andComputer Science, University of Southampton, anda Research Advisor for the Electricity Utilization,

University of Cambridge. He continued his work at the University of Cam-bridge in the capacity of a Fellow at Emmanuel College for three moreyears (1999–2002). He has a special interest in the development of numericalalgorithms (both serial and parallel codes) applied to the characterization ofmultiphysics problems, such as deoxyribonucleic acid sequencing, electronicmanipulation of nanoparticles, lightning, and gas discharges. His research in-terests lie predominantly in the utilization of electromagnetic fields and plasmaprocesses for environmental, food processing and biomedical applications,biomicroelectromechanical systems, nanotechnology, and power systems.


Documents

Bactericidal Action of the Reactive Species …...discharge at atmospheric pressure (from [16]). molecules and particles present, often very rapidly. The use of the gas discharge to