IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 41, NO.
1, JANUARY 2013
Gas Flow Effect on E. coli and B. subtilis Bacteria
Inactivation in Water Using a Pulsed Dielectric
Barrier Discharge
Benjamín Gonzalo Rodríguez-Méndez, Member, IEEE, Alma Neli Hernández-Arias, Régulo López-Callejas,
Raúl Valencia-Alvarado, Antonio Mercado-Cabrera, Rosendo Peña-Eguiluz, Member, IEEE,
Samuel Roberto Barocio-Delgado, Arturo Eduardo Muñoz-Castro, and Aníbal de la Piedad-Beneitez
Abstract—This paper presents an experimental study of the
inactivation in water of two representative classes of bacteria,
Gram-negative Escherichia coli and Gram-positive Bacillus sub-
tilis, using pulsed dielectric barrier discharges (PDBDs) in a coax-
ial arrangement. To this purpose, an adjustable plasma source
supplies 25-kV/500-Hz pulses, 30 μs long, at atmospheric pressure,
with total energy consumption estimated at about 100 mJ/pulse.
The inactivation effect of a PDBD on these types of microor-
ganisms has been previously studied in dependence on an oxy-
gen gas flow mode (null, continuous, and modulate). The results
have shown a significant bacterial reduction rate from 108 to
103 cells/mL with E. coli and from 107 to 103 cells/mL with
B. subtilis. The inactivation effectiveness is substantially similar
in both kinds of bacteria, although some data suggest a greater
susceptibility of the Gram-negative E. coli to plasma exposure.
Plasma diagnostics was carried out using optical emission spec-
troscopy whereby the OH radical and reactive oxygen species
formation rates in solution were found and the level of ozone
produced by the discharge was monitored. Finally, a kinetics
model was developed to characterize the chemical species taking
place in nonthermal plasma inactivation processes in water.
Index Terms—Atmospheric-pressure plasmas, discharge, mod-
eling, pulsed power supply (PPS), spectroscopy.
I. I NTRODUCTION
N ONTHERMAL PLASMAS (NTPs) at atmospheric pres-
sure are finding uses in diverse industrial and research
applications, with particular interest in environmental, medi-
cal, and biological areas [1], [2]. Different investigations have
shown that the NTP is a promising technology in the treatment,
modification, or processing of media and materials, mainly
because it represents an efficient source of chemical species
(electrons, ions, and active radicals), the gas phase is main-
Manuscript received August 9, 2012; revised November 15, 2012; accepted
November 17, 2012. Date of publication December 20, 2012; date of current
version January 4, 2013. This work was supported by CONACYT, México.
B. G. Rodríguez-Méndez, A. N. Hernández-Arias, R. López-Callejas,
R. Valencia-Alvarado, A. Mercado-Cabrera, R. Peña-Eguiluz, S. R. Barocio-
Delgado, and A. E. Muñoz-Castro are with the Instituto Nacional de Inves-
tigaciones Nucleares, México 11801, México (e-mail: benjamin.rodriguez@
inin.gob.mx; nelly.alama@hotmail.com; regulo.lopez@inin.gob.mx; raul.
valencia@inin.gob.mx; antonio.mercado@inin.gob.mx; rosendo.eguiluz@
inin.gob.mx; samuel.barocio@inin.gob.mx; arturo.munoz@inin.gob.mx).
A. de la Piedad-Beneitez is with the Instituto Tecnológico de Toluca,
Metepec 52140, México (e-mail: adelapiedad@hotmail.com).
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TPS.2012.2230343
0093-3813/$31.00 © 2012 IEEE
147
tained at room temperature, and the electrode configuration can
be widely varied. These features allow that this technology
can be used for a wide range of applications. In particular,
the plasmas generated by dielectric barrier discharge (DBD)
have shown a high effectiveness in bacterial inactivation, both
in gas and surfaces, whereas pulsed corona discharge (PCD)
plasma is generally applied in liquid media [1], [3]. Pulsed dis-
charges generated in liquid produce electromagnetic radiation,
shock waves, electric fields, and chemical species that facilitate
collateral physicochemical action which increases the synergy
of the bacterial inactivation process [4]. Mechanisms of the
bactericidal properties of the electrical discharge in water have
been proposed, both theoretically and experimentally, assuming
local and nonlocal effects, yet the central mechanism of bac-
terial inactivation remains unknown. A proposed mechanism
of bacterial inactivation by electrical discharges is the one that
suggests that chemical action degrades or inactivates biological
cells, showing a given amount of residual disinfectant capa-
bility even when the plasma has been turned off, maintaining
an inactivation ability [5]. However, the chemical mechanisms
are not well understood, and this field is being extensively
investigated [6], [7].
High-voltage power supply is currently applied to generate
electric discharges. When this high voltage is applied between
two electrodes without isolation, an electric arc can be de-
veloped causing heating and high energy consumption. If a
dielectric barrier layer is placed between them, this barrier
limits the current, avoiding the arc formation and, in some
cases, forming a homogeneous plasma. This is a temporary
phenomenon, and in order to sustain the DBD plasma, the
applied voltage has to change over time. With this, a transient
NTP is generated in the gap [8].
In this paper, we present the results of a pulsed DBD (PDBD)
applied to bacterial inactivation in an aqueous environment
using a high-voltage pulsed power supply (PPS) in a cylindrical
coaxial geometry reactor. The strains considered in this study
were the two representative classes of bacteria: Gram-negative
Escherichia coli (E. coli) and Gram-positive Bacillus subtilis
(B. subtilis). We use an initially constant oxygen (O2 ) gas flow
injected to culture suspension samples in the reactor, in order
to compare them to the null gas flow sample case. After that,
we proposed a modulated gas flow in order to compare the
influence of either: null, constant, and modulated gas flows on
148
Fig. 1. Experimental layout for PDBD in water.
the bacterial inactivation process effectiveness. The chemical
species generated in a water environment with O2 are important
to understand better the mechanism of bacterial inactivation.
Thus, the respective optical emission spectroscopy (OES) and
chemical kinetics model results will be presented, and the
proposed inactivation mechanism will be discussed.
II. E XPERIMENTAL S ETUP AND M ETHODS
A. Electrical Setup
The high-voltage pulsed discharge was provided by a com-
pact PPS in a flyback configuration. The main circuit of the
PPS consists of a direct current source, a power insulated-gate
bipolar transistor coupled by means of a pulse transformer. This
design allows a pulse repetition rate of 100–2000 Hz, a 1–30-μs
pulsewidth, and 0–30-kV amplitudes and can be adjusted at any
time [9]. The reactor cathode was made of stainless steel in
coaxial configuration (30 cm long with an inner diameter of
2.54 cm) whose symmetry axis is occupied by a tungsten rod
covered with alumina as dielectric barrier. In order to provide
the gas supply, a gas inlet and a commercial MKS mass flow
controller have been assembled. High-purity O2 (99.99%) is
used as discharge gas at atmospheric pressure, creating bub-
bles into the culture liquid. The main electrical diagnostics
include an oscilloscope (TDS 2024, Tektronix Inc.), a high-
voltage probe (Tektronix P6015A, 1000 × 3.0 pF–100 MΩ),
and a current transformer (Stangenes 2-0.1 W). In order to
identify the chemical species in the discharge, we used a
spectrometer (Jaz OceanOptics) with a maximum optical reso-
lution of ∼0.3 nm (FWHM) performed over the 270–870-nm
wavelength range and an ultraviolet (UV) absorption ozone
meter (Ozone Monitor model 460 L, Teledyne Instruments)
operating at 253.7 nm. Both outputs are processed and recorded
on a computer. The experimental layout for PDBD in water is
shown in Fig. 1.
B. Preparation of Biological Samples
The effectiveness of PDBD bacterial inactivation in water
was measured with cells of E. coli (ATCC 8739) and vegetative
cells of B. subtilis (ATCC 6633). As for E. coli, a strain is inoc-
IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 41, NO. 1, JANUARY 2013
ulated in 5 mL of Luria-Bertani (LB) liquid broth and incubated
for 24 h at 37 ◦ C, being continuously stirred (overnight culture).
The culture was centrifuged twice for 10 min at 5000 r/min. The
sedimented bacteria were put in 5 mL of distilled and sterilized
water. In the case of B. subtilis bacteria, 5 mL of trypticase-
soy-broth solution was inoculated with a strain and incubated
for 24-h periods at 37 ◦ C as to obtain the reference sample
following the method described earlier.
In order to determine the initial concentration of both kinds
of cells obtained after overnight culture, an aliquot of 100 μL
at 1 : 102 dilution from the original sample was made, and the
count of cells was performed by a phase contrast microscope
in a Neubauer improved chamber (depth of 0.1 mm and a
0.04-mm2 area per square). The concentration in cells per
milliliter (cells/mL) was recorded.
The number of E. coli cells counted was approximately
107 cells/mL. The original concentration can also be ap-
proached by multiplying it by the dilution factor, leading to
∼109 cells/mL. A count of ∼105 cells/mL of B. subtilis was
obtained, giving a final concentration of ∼107 cells/mL. The
concentration for each experiment was obtained following this
method and adjusted by standard progressive dilutions.
C. Experimental Method
The power supply output was established as 25 kV at 500 Hz,
with an ∼30-μs pulsewidth. As for E. coli bacteria experi-
ments, 15-mL distilled water samples containing controlled
concentrations of 103 , 104 , 106 , and 108 cells/mL were placed
into the reactor. Ten experiments were carried out with each
concentration, whereas an untreated sample was kept as a
reference. The reactor gas inlet was kept closed during exper-
iments without gas flow. In order to improve the production
of reactive oxygen species (ROS), two cases were considered:
First, O2 was injected at a 0.5-L/min flow rate in a continuous
mode during each treatment period, maintaining a gas bubbling
throughout the reactor. In the second place, an intermittent
0.5-L/min flow of O2 was admitted for 15-s periods every
30 s until the end of the treatment period. The 0.5-L/min O2
gas intake flow optimized the production of ozone (O3 ) under
the experimental conditions.
The plasma treatment of B. subtilis bacteria demanded a sus-
pension prepared by placing 15 mL of distilled water containing
103 , 105 , and 107 cells/mL. The experiments were carried out
during periods dependent on each concentration and in three
modes, namely, without gas flow (null mode), continuous, and
modulated modes, maintaining the discharge parameters as
mentioned before.
Immediately after plasma treatment, the suspension volume
samples (water containing bacteria) were removed from the
reactor and transferred to sterile tubes after which the micro-
biological analysis was carried out.
D. Microbiological Analysis
Three Petri dish (90-mm diameter) sets per sample of 103
E. coli cells/mL (undiluted, at a 103 cells/mL concentration)
in LB agar were compounded by inoculating 100 μL in each
RODRÍGUEZ-MÉNDEZ et al.: GAS FLOW EFFECT ON BACTERIA INACTIVATION IN WATER 149

dish using a spread-plate technique. Samples were taken from

both untreated and treated samples. Aliquots of 103 cells/mL,
from 104 , 106 , and 108 E. coli cells/mL concentrations, were
obtained from an untreated sample by standard dilution assay.
In order to determine its initial concentration, 100 μL of the
dilution was spread onto Petri dishes containing agar medium.
With respect to treated samples, dilutions from 1 : 10 to 1 : 105
were spread in triplicate directly onto the surface of LB agar,
and the cultivation method was applied. Once labeled, the
inoculated dishes were incubated overnight at 37 ◦ C for 24 h.
The culture of B. subtilis bacteria was carried out on three
Petri dishes per sample containing trypticase soy agar (TSA)
and compounded by inoculating aliquots of 100 μL that were
taken from both the untreated and the treated bacterial sus-
pension. The bacteria were cultured directly on the plate when
the concentration was 103 cells/mL. For 105 and 107 cells/mL, Fig. 2. Typical voltage and current waveforms of a PDBD in water.
samples of 103 cells/mL were obtained from a series of succes-
sive dilutions and spread on TSA. The inoculated dishes, once x are mol per liter (mol/L). The step change in xi , after the
labeled, were incubated for 24-h periods at 37 ◦ C. completion of Rj denoted by dij = rij − lij , is given by
After incubation, the number of colonies from the initial ⎧
⎨ −lij if Xi is a reactant in Rj
cell concentration and the treated suspension was estimated
dij = +rij if Xi is a product in Rj (2)
as the number of surviving cells, assuming that every viable ⎩
0 otherwise.
bacteria results in the formation of a colony [10]. Colony-
forming units (CFUs) were enumerated, and concentrations We can thus write for the reaction rate vj of Rj
per unit volume (CFU/mL) were determined by direct colony
counts. The bacterial inactivation effect versus the total plasma 
N
l
treatment time (kinetics) was determined and presented in v j = kj xiij (t) (3)
i=1
semilogarithmic scale. Later, the “D” value (decimal value or
decimal reduction), which is the interval of time required to and for the system of ODEs describing how each concentration
provide one decimal logarithm or 90% reduction in the initial changes over time
viable bacterial population, was determined.
dxi (t) 
M
= dij vj (x). (4)
dt j=1
E. Description of the Chemical Kinetics Model
To describe the chemical processes during the high-voltage
discharge in water, a chemical kinetics model was developed.
III. R ESULTS
The model was based on the construction of a set of rate
equations represented as ordinary differential equations (ODEs) A. Electrical Measurements of the PDBD
with concentrations of chemical species as variables. We con-
The PDBD voltage and current waveforms are shown in
sider that chemical reactions are the only cause of concentration
Fig. 2. The current waveform presents peaks, which indicate
changes, and diffusion, thermal shift, etc., were neglected for
streamer formation in the reactor, which is limited by the
simplicity. Therefore, the model becomes 0-D and time depen-
dielectric avoiding arcing and maintaining the typical behavior
dent. To set up a general mathematical framework, we consider
of a PDBD [9].
a chemical reaction network or pathway involving N species.
The energy per pulse dissipated in the reactor was determined
To apply the law of mass action, the framework is decomposed
from the time integral of the voltage and current waveforms,
into M elementary reactions, where an elementary reaction will
measured across the electrodes of the reactor (using the Waves-
be written as [11]
tar program from Tektronix), resulting in ∼100 mJ/pulse. It
seems that the PDBD is more energy efficient compared with

N
kj 
N
Rj : lij Xi → rij Xi , j = 1, . . . , M (1) other electrohydraulic discharge systems where PCDs consume
i=1 i=1 energy in the 1-J/pulse range. Likewise, pulsed arc discharges
require energy in the range of 1 kJ/pulse and larger [12].
where Xi is the ith chemical species, Rj is the jth elementary
reaction, lij and rij are the stoichiometric coefficients, and
B. E. coli Inactivation
kj is the rate coefficient for Rj . If a species Xi is not a
reactant in reaction Rj , lij = 0. Similarly, rij = 0 if Xi is not Initial experiments of E. coli inactivation were carried out
a product in Rj . It is common to use the notation xi for the comparing the three gas flow modes: without gas, continuous,
molar concentration [Xi ], where the units of the concentrations and modulated O2 gas flow at 103 , 104 , 106 , and 108 CFU/mL
150
Fig. 3. E. coli inactivation by PDBD treatment with the following initial
conditions: (a) 108 , (d) 106 , (g) 104 , and (j) 103 CFU/mL without gas flow;
(b) 108 , (e) 106 , (h) 104 , and (k) 103 CFU/mL in continuous oxygen flow; and
(c) 108 , (f) 106 , (i) 104 , and (l) 103 CFU/mL in modulated flow.
concentrations. Results are shown in Fig. 3. Without gas flow, it
took around 240 s of treatment time to achieve a 1.4-log CFU/
mL reduction at 103 CFU/mL concentrations. A 1.7-log re-
duction in continuous gas flow was achieved, and a 2.6-log
reduction was achieved when a modulated gas flow was applied,
both at 120 s. At concentrations of 104 CFU/mL, it took around
300 s to achieve a 2-log order reduction without gas flow.
These are extended up to > 3-log when the treatment period is
300 s with continuous O2 flow. In contrast, modulated gas flow
produced a greater than 3-log inactivation in measured E. coli
CFU/mL concentration, within 240 s. At 106 CFU/mL, the re-
duction is < 1-log without gas until 420 s, 3.4-log with 360 s of
continuous O2 injection, or 4-log when the O2 was modulated
for 300 s. Finally, if concentrations as high as 108 CFU/mL are
chosen, 3-log reductions demand 600-s periods with modulated
gas flow; otherwise, the times extend much longer.
C. B. subtilis Inactivation
Experiments of B. subtilis inactivation were performed at
103 , 105 , and 107 CFU/mL for different treatment periods under
the conditions without gas flow, in continuous flow, and in
modulated O2 flow. The quantitative results are given in Fig. 4.
The survival curves in this figure indicate that a 2.8-log
reduction was achieved in about 360-s exposure treatment
time at concentrations of 103 CFU/mL without gas flow. A
3.3-log reduction was determined in continuous O2 flow at
120 s, whereas a 3-log reduction was achieved in only 30 s
with modulated gas flow. At concentrations of 105 CFU/mL,
reductions are achieved on the order of 2-log at 600 s in null
flow mode and 3.9-log at 360 s when the gas is injected in
continuous mode, and a 4.8-log reduction was obtained when
a modulated flow of gas was provided. Finally, at 107 CFU/mL,
around 1-log reduction demanded 1200-s periods without O2
gas flow, whereas a 3-log reduction in continuous mode in the
same time and a 4-log reduction in modulated gas flow at 600 s
are achieved.
IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 41, NO. 1, JANUARY 2013
Fig. 4. Inactivation kinetics curve of the B. subtilis by PDBD at (a) 107 ,
(d) 105 , and (g) 103 CFU/mL without gas flow; (b) 107 , (e) 105 , and
(h) 103 CFU/mL in continuous oxygen flow; and (c) 107 , (f) 105 , and
(i) 103 CFU/mL in modulated gas flow.
TABLE I
D-VALUES FOR E. coli AND B. subtilis I NACTIVATION
Based on the data of Figs. 3 and 4, preliminary results
indicate that O2 increased the efficacy of the process and that
the modulation action is much more effective and raises the
elimination efficacy. D-values, given in seconds for E. coli and
B. subtilis, are listed on Table I.
D. Chemical Analysis of the Discharge
The electric discharges produce physical phenomena and
reactive chemical species acting synergistically, thereby ac-
celerating the elimination rate of microorganisms. In particu-
lar, breaking down in water can generate powerful oxidation
species such as the hydroxyl radical (OH) with a high potential
to inactivate the bacteria [13], [14]. The main chemical pro-
cesses concerning the atoms of O and H and the molecules of
O2 , O3 , OH, H2 O, HO2 , and H2 O2 are listed in Table II, where
we have identified some of them by applying OES.
The emission spectrum from chemical species determined by
OES is shown in Fig. 5, displaying the emission bands from the
discharge with and without O2 gas flow.
Fig. 5(a) shows the respective emission spectra, mainly from
hydroxyl radical in the 306.3-nm band (A2 Σ+ → X2 Π, 0-0),
the band of the N2 second positive system (337.1 nm, C3 Πu →
B3 Πg , 0-0), the N+2 first negative system (358.2, B Σu →
2 +
2 +
X Σg , 1-0), the N2 second positive system (380.4 nm,
C3 Πu → B3 Πg , 0-2), the Hα (656.3 nm, 3d2 D → 2p2 P o ), and
RODRÍGUEZ-MÉNDEZ et al.: GAS FLOW EFFECT ON BACTERIA INACTIVATION IN WATER
TABLE II
S UMMARY OF THE M AIN C HEMICAL R EACTIONS IN WATER
Fig. 5. Typical emission spectrum in PDBD in water: (a) Without oxygen gas
flow and (b) with oxygen gas flow into the reactor.
O (777.4 nm, 3p5 P → 3s5 S o ), produced by the PDBD in H2 O
without O2 flow. It should be observed that the spectra are
dominated by N2 transitions.
When O2 gas was bubbled, the emission intensities
[see Fig. 5(b)] of the Hγ (434 nm, 5d2 D → 2p2 P o ), Hβ
(486.1, 4d2 D → 2p2 P o ), Hα (656.3 nm, 3d2 D → 2p2 P o ),
O (777.4 nm, 3p5 P → 3s5 S o ), and O (844.6 nm, 3p3 P →
3s3 S o ) are increased considerably compared to the case with-
out gas. A strong emission is observed in the positive oxygen
151
Fig. 6. Numerical simulation of the main chemical species generated by
PDBD in water.
4 −
molecule ion (O+ 2 ) first negative system (529.5, b Σg → a Πu ,
4
2-0) and the OH radical in the band of 613.7 nm (X2 Π, 5-0),
presenting a significant difference compared to Fig. 5(a).
Ozone concentration, in parts per million (ppm), was ob-
tained from all experiments. Concentration was saturated be-
tween 140 and 155 ppm in continuous oxygen flow, whereas
with modulated gas, saturation did not exist, but it presented an
oscillatory behavior between 30 and 120 ppm. The ozone level
measured in the process without gas flow is < 10 ppm.
For the second part of our chemical analysis, the simplified
chemical model considers a system of 15 ODEs, and all equa-
tions were solved using the ODE solver in MATLAB program.
First, the rate constant values of the reactions were obtained
mainly from [15]–[17]; we assume quasi-stationary concentra-
tions of the species initially present in the reactor (H2 O and
O2 ) and electron impact reactions. The initial concentrations of
water and oxygen were 0.83325 and 0.0186 mol/L, respectively,
and the concentrations of all other species are assumed to be
equal to zero at the reactor. The electron density was taken on
the order of 1014 cm−3 [18], and the reactor temperature was
fixed at room temperature, 300 K.
Results from the mathematical model obtained through sim-
ulations are shown in Fig. 6 where the concentrations (mol/L)
are plotted versus time. During the initial voltage pulse period,
electron impact reactions take place and produce radicals.
These radicals then react with other species in the system to
initiate the chemistry in water.
Under the condition that O2 gas is not bubbled (Fig. 7), an ef-
fect in the OH radical, H2 O2 , HO2 , and O3 species is observed
with respect to when oxygen is bubbled in continuous mode
(see Fig. 8). The initial concentrations used in the simulations in
these last figures were taken from simulation data from Fig. 6.
We can notice that, by adding or removing the oxygen supply in
the last interval of the simulation, we obtain very different kinds
of profile charts. The formation of OH, O, H, and O+ 2 radicals
in this system was proven by the emission spectroscopy, but
from the viewpoint of application of electrical discharges for
water treatment, OH radical and H2 O2 are the most important
species [19].
152
Fig. 7. Kinetic simulations without O2 gas in the reactor gas inlet and its
effect on chemical species concentrations.
Fig. 8. Dependence of the O2 gas injected in the reactor on chemical species
concentrations.
IV. D ISCUSSION
In order to understand the difference between the spectra
shown in Fig. 5 and numerical simulations shown in Figs. 7
and 8 and its effect in bacterial inactivation, we will discuss the
production processes of O, H, and OH radical under the condi-
tion of added O2 . The O2 can be dissociated through reactions
(R1)–(R5), whereas considerable proportions of O(1 D) and O
atoms can be produced [13]. The main pathway leading to OH
radical formation by discharge is thought to be the result of
the direct dissociation of water molecules by electron impact,
leading to the generation of O(1 D), H, and OH, (R7)–(R9).
However, it has also been reported that the generation of OH
radicals increases along with the concentration of O2 molecules
through the reactions (R11) and (R14) [14]. However, there
are other reactions responsible for the OH decay as well as
the recombination reactions such as reactions (R16)–(R21).
Ozone (O3 ) is the primary species produced by an electrical
discharge occurring in either air or oxygen, through reactions
(R22)–(R24). It is the only active species that can diffuse into
the liquid phase and react with aqueous contaminants that are
IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 41, NO. 1, JANUARY 2013
susceptible to degradation by direct attack [19]. Thus, O3 can
react in an aqueous medium both directly, e.g., as molecular
ozone, or indirectly, via radical intermediates formed during
ozone decomposition in the aqueous media. OH radical is the
most important species formed during ozone decomposition
through reactions (R25)–(R26). H2 O2 is a strongly oxidative
compound and is produced by the recombination of OH radical
(R21) and by the recombination of hydroperoxyl (HO2 ) radical
through reaction (R27).
Atomic oxygen and OH radicals are reported to be one of the
important contributors to bacterial inactivation, are suggested
to play an important role to cause the physical destruction
of bacteria, and lead to irreversible cytoplasmic membrane
damage by oxidation [20]. Ozone germicidal effects are caused
by its interference with cellular respiration. The unsaturated
fatty acids of the cell membrane are susceptible to attacks by
OH radical; the presence of this radical can therefore compro-
mise the function of the membrane lipids. Protein molecules
are susceptible to oxidation by atomic oxygen or metastable
oxygen molecules [21]. We estimate that the pure discharge
survival curve reflects inactivation by intrinsic photodesorption
and rupture of the cytoplasmatic membrane due to the elec-
trostatic force and/or UV radiation, mainly since not enough
ROS from the gaseous phase are available. In contrast, when
adding O2 , which partially dissociates to yield O, O(1 D), O+ 2,
O3 , and then both electrostatic force and ROS contribute to
the bacterial inactivation. The results for O2 modulated can be
interpreted as ROS and radicals being formed during the O2
gas-on events, which react with the environment during and as
soon as the flow is suppressed (residual inactivation ability) and
where one or more active species interact with the cytoplasmic
cell membrane suppressing consequentially the bacterial activ-
ity. The residual inactivation ability might be contributed by
long-life active species, such as H2 O2 and O3 . By contrast, a
saturation of O2 and O3 molecules and O atoms produced by
the added O2 in continuous mode can react together and with
OH radicals through reactions (R28)–(R30) [13] producing O2
and maintaining a relative abundance of hydroperoxyl radical,
but it is a much less strong oxidizer than the OH radical, O3 , or
H2 O2 . Thus, although a vast number of OH radicals are formed
with H and HO2 reacting with O3 , (R25)–(R26), a saturation of
O and O3 could be counterproductive.
V. C ONCLUSION
A PDBD technique in water has been carried out success-
fully, proving that its application is determinant for Gram-
positive and Gram-negative bacteria inactivation in water even
at high concentrations. Discharges in water were efficient in the
formation of chemically active species, particularly OH radicals
and ozone, where the optical emission spectra of chemical
species in water, with and without oxygen flow, have been
observed successfully. It appears that, at atmospheric pressure,
the principal agent of bacterial inactivation is the chemical
species. When O2 gas was bubbling, the effectiveness for bac-
terial inactivation was higher because of the formation of active
species based in oxygen (ROS). The enhanced effectiveness
of the modulated oxygen intake discharges can be attributed
RODRÍGUEZ-MÉNDEZ et al.: GAS FLOW EFFECT ON BACTERIA INACTIVATION IN WATER
to the prevention of gas saturation as well as to an improved
ROS–water reaction leading to the production of more free
radicals through different chemical pathways. By contrast, sat-
urated oxygen and ozone may react with the free OH radicals,
which counterbalance the bactericidal activity. This finding is
highly relevant for future applications of electrical discharges in
bacterial inactivation in water. Despite its simplicity, the results
from the model simulation help to understand the principles of
the process such as chemical species in water and its influence
on the bacterial inactivation. Future experiments and simula-
tions will focus on evaluating all contributors, thus bringing
forth more understanding on the mechanisms of bacteria inac-
tivation in water.
ACKNOWLEDGMENT
The authors would like to thank I. A. Rojas O., M. T. Torres
M., P. Angeles E., and I. Contreras V. for their technical support.
R EFERENCES
[1] H. Ayan, D. Staack, G. Fridman, A. Gutsol, Y. Mukhin, A. Starikovskii,
A. Fridman, and G. Friedman, “Application of nanosecond-pulsed dielec-
tric barrier discharge for biomedical treatment of topographically non-
uniform surfaces,” J. Phys. D, Appl. Phys., vol. 42, no. 12, pp. 125202-1–
125202-5, Jun. 2009.
[2] A. Pokryvailo, Y. Yankelevich, M. Wolf, E. Abramzon, S. Wald, and
A. Welleman, “A high-power pulsed corona source for pollution control
applications,” IEEE Trans. Plasma Sci., vol. 32, no. 5, pp. 2045–2054,
Oct. 2004.
[3] M. Sato, “Environmental and biotechnological applications of high-
voltage pulsed discharges in water,” Plasma Sources Sci. Technol., vol. 17,
no. 2, pp. 024021-1–024021-7, May 2008.
[4] M. Laroussi, “Low temperature plasma-based sterilization: Overview and
state-of-the-art,” Plasma Process. Polym., vol. 2, no. 5, pp. 391–400,
Jun. 2005.
[5] C. W. Chen, H. M. Lee, and M. B. Chang, “Influence of pH on inactivation
of aquatic microorganism with a gas-liquid pulsed electrical discharge,”
J. Electrostat., vol. 67, no. 4, pp. 703–708, Jul. 2009.
[6] B. Gweon, D. B. Kim, S. Y. Moon, and W. Choe, “Escherichia coli
deactivation study controlling the atmospheric pressure plasma discharge
conditions,” Current Appl. Phys., vol. 9, no. 3, pp. 625–628, May 2009.
[7] M. M. Guivan, T. Kamikozawa, H. Kurokawa, H. Motomura,
K. Kadowaki, and M. Jinno, “Comparative inactivation of Bacillus subtilis
spores using a DBD-driven xenon iodide excilamp and a conventional
mercury lamp,” IEEE Trans. Plasma Sci., vol. 38, no. 8, pp. 1972–1977,
Aug. 2010.
[8] H. Ayan, “Uniform dielectric barrier discharge with nanosecond pulse
excitation for biomedical applications,” Ph.D. dissertation, Drexel Univ.,
Philadelphia, PA, 2009.
[9] B. G. Rodríguez-Méndez, R. López-Callejas, R. Peña-Eguiluz,
A. Mercado-Cabrera, R. Valencia Alvarado, S. R. Barocio, A. de
la Piedad-Beneitez, J. S. Benítez-Read, and J. O. Pacheco-Sotelo,
“Instrumentation for pulsed corona discharge generation applied to
water,” IEEE Trans. Plasma Sci., vol. 36, no. 1, pp. 185–191, Feb. 2008.
[10] P. S. Lee, Practical Handbook of Microbiology, Quantitation of Microor-
ganisms., 2nd ed. Boca Raton, FL: CRC Press, 2009.
[11] M. Ullah, H. Schmidt, K.-H. Cho, and O. Wolkenhauer, “Determinis-
tic modelling and stochastic simulation of biochemical pathways using
MATLAB,” Proc. Inst. Elect. Eng.—Syst. Biol., vol. 153, no. 2, pp. 53–60,
Mar. 2006.
[12] J. S. Chang, “Thermal plasma solid waste and water treatments: A critical
review,” Int. J. Plasma Environ. Sci. Technol., vol. 3, no. 2, pp. 67–84,
Sep. 2009.
[13] F. Liu, W. Wang, S. Wang, W. Zheng, and Y. Wang, “Diagnosis of OH
radical by optical emission spectroscopy in a wire-plate bi-directional
pulsed corona discharge,” J. Electrostat., vol. 65, no. 7, pp. 445–451,
Jun. 2007.
[14] R. Ono and T. Oda, “Measurement of gas temperature and OH density
in the afterglow of pulsed positive corona discharge,” J. Phys. D, Appl.
Phys., vol. 41, no. 3, pp. 035204-1–035204-11, Feb. 2008.
153
[15] A. A. Joshi, B. R. Locke, P. Arce, and W. C. Finney, “Formation of
hydroxyl radicals, hydrogen peroxide and aqueous electrons by pulsed
streamer corona discharge in aqueous solution,” J. Hazard. Mater., vol. 41,
no. 1, pp. 3–30, Apr. 1995.
[16] D. R. Grymonpré, A. K. Sharma, W. C. Finney, and B. R. Locke, “The role
of Fenton’s reaction in aqueous phase pulsed streamer corona reactors,”
Chem. Eng. J., vol. 82, no. 1–3, pp. 189–207, Mar. 2001.
[17] J. Zhang, J. Chen, and X. Li, “Remove of phenolic compounds in water by
low-temperature plasma: A review of current research,” J. Water Resour.
Protect., vol. 1, no. 2, pp. 99–109, Aug. 2009.
[18] B. Eliasson and U. Kogelschatz, “Modeling and applications of silent
discharge plasmas,” IEEE Trans. Plasma Sci., vol. 19, no. 2, pp. 309–323,
Apr. 1991.
[19] P. Lukes, A. T. Appleton, and B. R. Locke, “Hydrogen peroxide and ozone
formation in hybrid gas-liquid electrical discharge reactors,” IEEE Trans.
Ind. Appl., vol. 40, no. 1, pp. 60–67, Jan./Feb. 2004.
[20] R. Pompl, F. Jamitzky, T. Shimizu, B. Steffes, W. Bunk, H.-U. Schmidt,
M. Georgi, K. Ramrath, W. Stolz, R. W. Stark, T. Urayama, S. Fujii, and
G. E. Morfill, “The effect of low-temperature plasma on bacteria as ob-
served by repeated AFM imaging,” New J. Phys., vol. 11, pp. 115023-1–
115023-11, Nov. 2009.
[21] M. Laroussi, “Low-temperature plasmas for medicine?” IEEE Trans.
Plasma Sci., vol. 37, no. 6, pp. 714–725, Jun. 2009.
Benjamín Gonzalo Rodríguez-Méndez (M’12)
received the Ph.D. degree from the Instituto
Tecnológico de Toluca, Toluca, México, in 2008.
He is currently a Researcher with the Plasmas
Physics Laboratory, Instituto Nacional de Investi-
gaciones Nucleares, México, México, where he is
working on power electronics for the electrical dis-
charge generation and plasma applications. His re-
search interests include the modeling and electronic
instrumentation in the plasma field.
Alma Neli Hernández-Arias was born in 1984. She
received the Eng. degree in chemical engineering
from the Instituto Tecnológico de Toluca, Toluca,
México, in 2008, where she is currently working
toward the Ph.D. degree.
She is currently with the Instituto Nacional de
Investigaciones Nucleares, México, México. She has
been involved in research on the decontamination of
water and pulsed power applications to the biological
field. Her current research interests include biologi-
cal effects of electric discharges.
Régulo López-Callejas received the Ph.D. degree
in electronics engineering from the Instituto Tec-
nológico de Toluca, Toluca, México, in 2002.
Since 1979, he has been a Full-Time Researcher
in plasma physics instrumentation with the Plasmas
Physics Laboratory, Instituto Nacional de Investi-
gaciones Nucleares, México, México. His current
research interests include plasma applications, new
plasma technologies, and electronic instrumentation.
154
Raúl Valencia-Alvarado received the Ph.D. degree
in nuclear science from the Universidad Autónoma
del Estado de México, Toluca, México, in 2006.
He is currently a Researcher with the Plasmas
Physics Laboratory, Instituto Nacional de Investi-
gaciones Nucleares, México, México, where he is
focusing on the plasma characterization and plasma
processing of materials.
Antonio Mercado-Cabrera received the Ph.D. de-
gree in plasma physics and engineering from Paul
Sabatier University, Toulouse, France, in 2003.
He is currently a Full-Time Researcher with the
Plasmas Physics Laboratory, Instituto Nacional de
Investigaciones Nucleares, México, México, where
he is working on chemical kinetics, computational
fluid dynamics for air pollution control, and new
technologies for waste treatment.
Rosendo Peña-Eguiluz (S’99–M’03) received the
Ph.D. degree in electrical engineering from the In-
stitut National Polytechnique de Toulouse, Toulouse,
France, in 2002.
He is currently a Researcher with the Plasmas
Physics Laboratory, Instituto Nacional de Investi-
gaciones Nucleares, México, México. His research
interests include the modeling and control of power
static converters, plasma applications, and applied
digital control.
IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 41, NO. 1, JANUARY 2013
Samuel Roberto Barocio-Delgado received the
Ph.D. degree in physics from the University of
Manchester, Manchester, U.K., in 1997.
He was a Full-Time Researcher in the areas of
modeling and data interpretation with the Plasmas
Physics Laboratory, Instituto Nacional de Investiga-
ciones Nucleares, México, México, since 1981. He
passed away on 2012.
Arturo Eduardo Muñoz-Castro received the M.Sc.
degree in materials science from the Universidad
Autónoma del Estado de México, Toluca, México,
in 2008.
He is currently a Researcher with the Plasmas
Physics Laboratory, Instituto Nacional de Investi-
gaciones Nucleares, México, México, where he is
working on plasma physics, plasmas applied to sur-
faces, and biomaterials.
Aníbal de la Piedad-Beneitez received the M.Sc.
degree from the Faculty of Sciences, Universidad
Nacional Autónoma de México (UNAM), México,
México, in 1979.
For several years, he was with the Institute of
Physics, UNAM, where he worked on radiation
physics. Since 1988, he has been with the Instituto
Tecnológico de Toluca, Toluca, México. He special-
izes in plasma applications.