Reactive Fluxes from Atmospheric Pressure Plasmas for Deactivation of
Bacteria on Rough Surfaces and Suspended in Air*
Natalia Yu. Babaeva and Mark J. Kushner
University of Michigan, EEC S Department, 1301 Beal Ave, Ann Arbor, MI 48109 USA
Abstract: Many mechanisms have been proposed for deactivating bacteria using
atmospheric pressure plasmas (APPs). Ultimately, the bacteria are killed by a
combination of reactive fluxes of ROS (reactive oxygen containing species),
energetic ions, UV light, and electric fields in the context of electroporation. To
date, the relative contributions to bacterial deactivation of these fluxes and
electric fields are poorly understood. In this paper, results from a numerical
investigation are used to discuss the scaling of reactive fluxes produced in APPs
sustained in dry and humid air, and their interaction with bacteria suspended in
air and on rough surfaces. Two plasma sources are considered - dielectric-barrier
discharges and remote plasma jets.
Keywords: Bacteria, Dielectric Barrier Discharge, Plasma Jet
1. Introduction
Low temperature non-equilibrium atmospheric
pressure plasmas for air and surface cleaning
technologies have proven to be robust and safe
enough for use indoors. This is partly due to high
bactericidal effectiveness of plasmas and partly to
their ability to penetrate into narrow and confined
spaces, small cracks and microscopic openings.
Electron temperatures in APPs are typically a few
eV, sufficient to initialize many chemical reactions
to modify DNA, proteins and cell membranes. The
gas temperatures of non-equilibrium plasmas are
close to room temperature and so the process is
compatible with temperature sensitive materials.
Bacteria deactivation, disinfection, decontamination,
sterilization with plasmas are rapid. Experiments
have shown that a few seconds to a few minutes are
necessary to deactivate many different types of
microorganisms such as viruses and bacteria on
surfaces and in aqueous solutions [1,2].
The most likely plasma inactivation
mechanisms that produce lethal effects on
microorganisms are: (1) production of reactive
atoms, molecules and radicals, in both the ground
and excited states, that modify membrane chemistry
and permeability of cells, (2) production of ionic
species near and even on living tissues, (3)
bombardment of the cell membranes by energetic
ions, (4) UV irradiation resulting of dissociation
and/or cross linking of DNA strands, and (5) large
electric fields resulting in membrane disruption
either due to electroporation or charge accumulated
on the cell membrane. In addition, factors such as
the type of plasma source being used and the
geometry of the surface being sterilized, in addition
to the type of bacteria, determine the effectiveness of
the sterilization process.
Two approaches are being pursued in the
use of non-thermal atmospheric pressure plasmas in
medicine. In the first, the plasma is produced
remotely, and its afterglow is delivered in a plume to
the biological tissue. The sterilizing or therapeutic
effects are likely produced by relatively long-lived
neutral species and radicals as most of the charged
particles do not survive outside the plasma
generation region. Usually operated heavily diluted
with helium to avoid plasma instabilities, the
discharge is doped with a few percent of molecular
gases such as O2. These resulting oxygen-containing
active species may play a role via de-excitation and
subsequent energy transfer onto the microorganism’s
surface. Different devices exist for these treatments,
from plasma needles [1] to plasma jets [3]. Potential
applications of the remote plasma sources are
surface sterilization in a more targeted way than is
possible with large volume plasmas.
In the second approach, plasmas are
generated in direct contact with living tissue. When
dielectric barrier discharges (DBDs) are used for this
purpose, the plasma device typically contains the
powered electrode while the tissue is the counter
electrode [4]. The direct method fundamentally
differs from the indirect technique in two respects.
First, plasmas propagate and touch the biological
surface, providing the possibility of charging the
surface and delivering energetic ions. The second is
the magnitude of the electric field produced at the
surface – many orders of magnitude larger in the
direct method.
These two general categories of plasma
sources (indirect and direct) imply different
compositions of the plasma species and activation
energy delivered to the surface, particularly when
surfaces are rough and nonplanar. In this paper, we
discuss results from a numerical investigation of
direct and remote plasma treatment of biological
surfaces. The fluxes of active species delivered to
the surface in each case are compared.
2. Description of the Model
The physical processes and algorithms used
in the model are discussed in detail in Ref. [5-7] and
so will be only briefly reviewed here. The model,
nonPDPSIM, is a multi-fluid hydrodynamics
simulation in which transport equations for all
charged and neutral species and Poisson’s equation
are integrated as a function of time. Updates of the
charged particle densities and electric potential are
followed by an implicit update of the electron
temperature, Te, by solving the electron energy
conservation equation.
The electron transport
coefficients and rate coefficients for bulk electrons
as a function of average electron energy are obtained
by solving Boltzmann’s equation for the electron
energy distribution.
Poisson's equation was solved throughout
the entire the computational domain. Continuity
equations for gas phase charged and neutral particles
are only solved in the plasma region. Conservation
equations for surface and volume charges are solved
on and inside all non-metallic materials. Radiation
transport is addressed using a line-of-sight
propagator. UV radiation is generated by high lying
excited states that are produced largely in the high
E/N in the avalanche front. The UV radiation is
absorbed with and without producing ionization in
the gas and it flux recorded when striking surfaces.
An unstructured mesh having a dynamic range of
>1000 is used to enable both the DBD or plasma jet
reactor scale and the surface roughness of the flat
dielectric (or a floating particle) to be resolved in a
single mesh. IEADs to surfaces are computed using
the Plasma Chemistry Monte Carlo Module
(PCMCM), as described in Ref. [6].
3. Shadowing of Activation Energy on
Rough Surfaces: Direct Treatment by
DBDs
The plasma in DBDs typically consists of
filaments having area densities of tens to hundreds
per cm2. These filaments treat surfaces uniformly
due to the spreading of the discharge along the
surface, the spatial averaging of many discharges
and the diffusion of gas phase radicals. DBD
filaments or streamers are narrow conductive plasma
channels that are initiated near the anode for a
positive polarity and the cathode for a negative, and
propagate towards the opposite electrode that is
usually covered by dielectric.
Evolution of a negative filament is shown in
Fig. 1. A filament is sustained in a 1 mm gap
between top and the bottom electrodes. The powered
electrode is biased to -15 kV. The top powered
electrode is covered by a dielectric (ε/ε0 = 3). The
bottom grounded electrode is covered with a
dielectric polymer 127 µm thick with ε/ε0 = 16. The
polymer surface has roughness where bacteria
reside.
Figure 1. Electron density (flood, 2 decades log scale) and
potential (lines, linear scale, kV) as a DBD filament approaches
the dielectric polymer with a rough surface. The electric
potential is compressed in front of the filament resulting in an
enhanced electric field.
The gas mixture used in the simulations is
atmospheric pressure humid air N2/O2/H2O =
79/20/1 at 300 K. As shown in Fig. 1, the electron
densities in the filament channel reach 2 ×1015 cm-3
when the filament touches the rough surface of the
polymer. Upon intersection with the polymer, the
filament charges the surface, producing lateral
electric fields which result in the spreading of the
filament over the surface. When the filament strikes
the surface, much of this voltage is then transferred
to the resulting sheath at the surface. As a result, the
electric field in the sheath can exceeds 400 kV/cm
(1600 Td).
Local structure of the surface roughness
seeded with bacteria is shown in Fig. 2 where the
electron density, electric field and negative charges
are plotted near and around the nooks-and-crannies
of the surface. APPs have Debye lengths of 0.5 to 1
µm and are able to penetrate into the roughness
having dimensions of a few µm. The electric field is
enhanced near the convex edges and depleted near
the concave valleys due to the charging of
protruding structures.
charging by electrons of points with larges view
angles concentrates ion fluxes to those points.
Photons are simply line-of-sight shadowed from the
site of emission of the photon. On the other hand,
neutral radicals penetrate indiscriminately into the
roughness.
This disparity between directional
(photons), quasi-directional (charging) and nondirectional (neutral) fluxes may result in different
types of bacteria being more or less resistant to
plasma treatment on difference surfaces based on
characteristics in addition to roughness.
For
example, a dielectric surface will distribute charge
and electric fields different than conductive surfaces.
Figure 3. Integrated fluxes of photons and ions along the portion
of surface with roughness. Locations of bacteria are indicated
by numbers 1 to 5 corresponding to numbering in Fig. 2.
4. Activation Energy on Rough Surfaces:
Indirect Treatment by Plasma Jet
Figure 2. Close-up of the rough surface of the polymer for t=0.5
ns (corresponding to the right frame in Fig. 1). Electron
density, electric field and negative charge density are shown
near and around the shallow roughness.
Note that electric field as high as 700 kV/cm
can be induced by the approaching DBD filament
near the shallow roughness. The electric field in
combination with high surface charges (1013 – 1016
cm-3) can lead to the disruption of bacteria
membranes as noted, for example in Ref. [8]. These
electric fields are non-uniformly distributed along
the surface due to shadowing of directional fluxes of
charged particles. The charging and shadowing of
surfaces are particularly problematic in plasma
sterilization of rough surfaces. Bacteria in nooksand-crannies will receive fluxes in part proportional
to their view angle to the plasma as shown in Fig. 3.
Ratios of integrated fluxes of photons and ions vary
significantly across short distances treating
individual bacteria differently. For example, initial
A homogeneous glow discharge can be
sustained between two parallel plates at dimensions
less than 1 mm, with an effluent emitted into
ambient air [3]. Simulations results for a plasma-jet
operated at 13.56 MHz in are shown in Fig. 4. The
nozzle between two parallel plates supplies a
He/O2/H2O=95/4.5/0.5 mixture at 1000 sccm. The
surrounding gas is dry air, allowed to flow along the
outside of the discharge electrodes with a net rate of
50 sccm. The surface being treated is about 2 cm
away having roughness shown in the insert. In
indirect plasma treatment, the lack of significant
fluxes of ions and photons onto the surface is
compensated by the abundance of radicals and ozone
that contribute to bacteria deactivation.
For
example, the OH radicals density reaches 1.1x1012
cm-3. Gas compression due to the jet impinging onto
the surface produces slightly higher OH density in
the roughness of the surface (two bottom frames).
Remote processing provides more uniformity but at
the cost of lack of activation energy from ions and
photons.
5. Bacteria Suspended in Air
When a filament interacts with a suspended
bacteria, the bacteria charges, producing a sheath
around its body. The electric field in the sheath can
be 100’s kV/cm. Even with mean free paths of a
micron (or less), an ion can be accelerated in this
field to energies as high as tens of eV. For example,
we simulated the intersection of a positive streamer
in humid air with a dielectric bead 45 µm in radius.
(Although this is much larger than a bacterium, the
trends we observe here extend to smaller bacterium.)
The resulting distributions of O2+ ion energies
incident onto the top of the particle facing the
streamer are shown in Fig. 5. Ion energies of a few
eV are incident on the particle. The same
phenomena will occur on a smaller bacteria, perhaps
tempered by whatever intrinsic conductivity the
bacteria may have.
*
This work was supported by Department of
Energy Office of Fusion Energy Science
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
M. Laroussi, D. A. Mendis and M. Rosenberg,
New J. Phys. 5, 41 (2003).
G. E. Morfill, M. G. Kong and J. L.
Zimmermann, New J. Phys. 11, 115011 (2009).
T. Gans, D. O’Connell, V Schulz-von der
Gathen, J. Waskoenig, Plasma Sources Sci
Technol. 19, 045018 (2010).
G. Fridman, M. Peddinghaus, H. Ayan, A.
Fridman, M. Balasubramanian, A. Gutsol, A.
Brooks, and G. Friedman, Plasma Chem. Plasma
Process. 26, 425 (2006).
N. Yu. Babaeva and Mark J. Kushner, J. Phys.
D: Appl. Phys. 43, 185206 (2010) .
N.Yu. Babaeva and Mark J. Kushner, J. Phys. D:
Appl. Phys. 41, 062004 (2008).
A. N. Bhoj and M. J. Kushner, J. Phys. D. 40,
6953 (2007).
D. A. Mendis, M. Rosenberg, and F. Azam,
IEEE Trans. Plasma Sci. 28, 1304 (2000).
Figure 4. Plasma-jet device operated in an RF mode (13.56
MHz) in He/O2/H2O=95/4.5/0.5 mixture injected in atmospheric
pressure air with sccm=1000.
Figure 5. Ion energy and angle distributions on a dielectric bead
45 µm in radius resulting from the intersection of a positive
corona filament. Dielectric constant of the bacteria ε/ε0 =80.