JoumalofFoodEngineering 25 (1995) 261-281

Elsevier Science Limited

Printed in Great Britain
0260.8774/95/$9.50
ELSEVIER
0260-8774(94)00030-l

Engineering Aspects of Pulsed Electric Field

Pasteurization

Qinghua Zhang,” Gustav0 V. Barbosa-C%ovas”* & Barry G.

Swansonb

“Departmentof Biological Systems Engineering, hDepartment of Food Science and

Human Nutrition, Washington State University, Pullman, WA 99164-6120, USA

(Received 24 May 1993; revised version received 18 April 1994;

accepted 10 June 1994)

ABSTRACT

Pulsed electricfields (PEF) is a promising technologyfor the non-thermal

pasteurization of foods and a sound complement or replacement to
traditional thermal pasteurization, which inactivatesbacteria and other
microorganisms harmful to humans, but also degrades color, flavor,
texture and nutrients. Foods can be pasteurized withpulsed electric fields
at ambient or r@igerated temperatures for a short treatment time of
seconds or less and the fresh-like quality offood is preserved.
Although successful in laboratory tests, applying pulsed electric fields
to food pasteurization on a large scale presents many unresolved
engineering problems. This paper reviews current literatureand discusses
design considerations for high voltage PEF pasteurization equipment.
Examples of PEF microbial investigationare included.

NOTATION

A Electrode area (cm’)

c Effective capacitance of food in the treatment chamber (p F )
co Energy storage capacitor (p F )
CP Specific heat of fluid food inside the treatment chamber (kJ/kg “C)
d The gap between two parallel plate electrodes (cm)
Actual peak electric field strength (kV/cm)
& Critical peak electric field strength (kV/cm)
Frequency of repetitive pulses (Hz)
$ Volumetric flow rate of food (cm3/s)

*To whom correspondence should be addressed.

261
 262 Q. Zhang, G. I/. Barbosa-Ccinovas,B. G. Swanson

Z Current (kA )
k Microorganism constant (kV/cm)
L Inductance of pulse forming unit (p H)
Number of pulses, dimensionless
; Continuous power rating of PEF generator (W )
Q Heat input to the chamber (J/cm3)
R Effective resistance of food in the treatment chamber (B )
Rs Charging current limiting resistor (Q )
s Microbial survival rate, dimensionless
t Treatment time (s)
tc Critical treatment time (s)
V Volume of treatment chamber (cm’)
V Peak discharge voltage (kV )
v, Initial charged voltage over the energy storage capacitor (kV)

EO The permittivity of free space, 8.84 X 10 - 8 (p F/cm)

E, Relative permittivity of the food material, dimensionless
P Resistivity of the food (Qcm)
Pf Density of fluid food inside the treatment chamber (g/cm3)
u Conductivity of the food (Siemens/m)
z Pulse width (pus)

INTRODUCTION

Minimally processed, fresh-like food products are becoming common terms in

the food industry in the 1990s. As a consequence, the non-thermal pasteuriza-
tion of foods is receiving increased attention (Mertens & Knorr, 1992). Pulsed
electric fields, oscillating magnetic field pulses, microwave induced electro-
magnetic fields, electron ionizing radiation, intense light pulses, and high
pressure treatments are promising non-thermal pasteurization techniques, and
may be combined to give the total bacteria reduction desired as in the ‘hurdle
concept’, applying gentle individual steps to gradually reduce microbial counts
(Mertens & Knorr, 1992).
Pulsed electric field (PEF ) treatment reduces the activity of bacteria and
other microorganisms. Numerous researchers have reported the effective
bactericidal activity of PEF, including the following: Hamilton and Sale (1967);
Sale and Hamilton (1967, 1968); Hiilsheger and Niemann ( 1980); Hiilseger et
al. (1981, 1983); Jacob et al. (1981); Hofmann (1984); DUM and Pearlman
( 1987); Mizuno and Hori ( 1988) Gupta and Murray ( 1988); Matsumoto et al.
( 1991); Jayaram et al. ( 1992); Grahl et al. ( 1992); Castro et al. (1993);
Pothakamury et al. ( 1994); and Zhang et al. ( 1994).
One of the explanations for the PEF bactericidal effect is the dielectric
rupture theory (Zimmermamr et al., 1976; Zimmermann, 1986). An applied
electric field induces an electric potential across the membrane of a living cell
suspended in the electric field. The electric potential, in turn, causes an electro-
static charge separation in the cell membrane based on the dipole nature of
the membrane molecules (Bryant & Wolfe, 1987). When the transmembrane
 Engineering aspects of pulsed electric field pasteurization 263

potential exceeds a critical value of approximately 1 V, the repulsion between

charge-carrying molecules initiates the formation of pores in weak areas of the
membrane (Zimmermamr et al., 1976; Benz et al., 1979; Benz & Zimmermamr,
1980; Hofmann & Evans, 1986; Zimmermann, 1986; Chang & Reese, 1990). A
lethal effect to living cells is observed when the transmembrane electric
potential exceeds the critical value by a wide margin.
Known as electroporation or electrofusion, the pore-forming feature of PEF
is primarily used in biotechnology where foreign DNA material is diffused into
the living cell to modify the cellular characteristics (Sowers, 1985, 1986;
Somkuti & Steinberg, 1988; Tsong, 1991). The lethal effect of PEF is related to
the length of time the field is applied. Inactivation becomes substantial after a
number of pulses, which indicates combined thermal kinetics. Dunn and Pearl-
man (1987) suggested an optimum fluid temperature range of 60-75X.
Jayaram et al. ( 1992) also reported increased PEF microbial inactivation at ele-
vated temperatures. Other parameters that may affect the microbial inactivation
are microbial growth stages, ionic strength and species in the suspension fluid,
pH and hydrostatic pressure.
Although PEF pasteurization of foods is observed in laboratory tests, the
engineering aspects for scaling up to industrial operation remain unresolved. To
effectively utilize PEF technology, the following concerns must be addressed:
1. critical, optimum and maximum electric field strengths applicable to the
inactivation of bacteria in food materials;
2. effect of Joule heating and cooling considerations;
3. arcing and dielectric strength of food materials;
4. flow rate of foods and power requirement for PEF equipment;
5. safety of PEF equipment operation.
In this paper, related literature is reviewed, design considerations for the PEF
pasteurization equipment are analyzed, and some experimental results are
discussed.

PULSED ELECTRIC FIELDS AND PASTEURIZATION

Applying PEF to food materials on an industrial scale requires knowledge of

effective electric field strengths necessary to inactivate microorganisms, as well
as the electrode field enhancement, dielectric breakdown phenomena, and
dielectric and electrical properties of food materials.

Generation of pulsed electric fields

Liquid food materials are, in general, electrical conductors because of the

presence of large concentrations of ions as electrical charge carriers. To
generate a high voltage pulsed electrical field of several kV/cm within the foods,
a large flux of electrical current must flow through a piece of food in a treatment
chamber for a very short period of time (i.e. microseconds). The time between
pulses is much longer than the pulse width. Therefore, the generation of pulses
involves slow charging and fast discharging of an electrical energy storage device
such as a capacitor. Figure 1 illustrates a simplified circuit of an exponential
264 Q. Zhang, G. V. Barbosa-Cdnovas, B. G. Swanson

Discharge Switch
ChargingResistor

DC Power
SUPPlY

Fig. 1. A simplified circuit for exponential decay pulse generation.

0
0 10 20 30 40 50
Time (KS)

Fig. 2. Exponential decay pulse (solid line) and square pulse (dotted line) with the
same peak voltage V, = 26 kV and equivalent energy (740 J). The pulse width for the
exponential decay pulse is 11 ,us, and for the square pulse it is 5.6 ,BS.

decay pulse generator with R being the effective resistance of food in the treat-
ment chamber, R, the charging current limiting resistor, and C, the energy
storage capacitor.
Many devices may be used as the discharge switch, including a mercury
ignitron spark-gap, a gas spark-gap, a thyratron, a series of SCRs, a magnetic
switch or a mechanical rotary switch. Two types of high voltage power supplies
may be used to charge the capacitor, such as an ordinary direct current (DC)
power supply or a capacitor charging power supply. An ordinary DC power
supply transforms alternating current (AC) from an utility line (60 Hz) into high
voltage AC, then rectifies to a high voltage DC. The capacitor charging power
supply uses high frequency AC input (in 100 kHz range) and provides command
charging. Command charging power supplies can provide higher repetitive rates
(up to kHz) than ordinary DC power supplies since a smaller R, value may be
used.
An exponential decay pulse generated by the circuit illustrated in Fig. 1 is
plotted in Fig. 2 (solid line) along with a square wave pulse (dashed line). Both
 Engineering aspects ofpulsed electricfieldpasteurization 265

Fig. 3. Layout of a square generator using a pulse forming network of three

capacitor-inductor units.

pulses have the same peak voltage and equivalent energy. The pulse width is
defined as the time needed to decrease the voltage to 37% of its peak value.
Exponential decay pulses have a long tail with a low electric field, during
which excess heat is generated in the food without bactericidal effect. On the
other hand, square pulse maintains a peak voltage for a longer time than the
exponential decay pulse. Since both waveforms effectively inactivate micro-
organisms, square wave pulses will save energy and require less cooling effort.
Exponential decay and square wave pulses are two major waveforms in
pulsed electric field technology. Exponential decay pulses are easier to obtain
and change. Generation of square pulses is complex and usually involves a pulse
forming network (PFN) as illustrated in Fig. 3. The square waveform is realized
when both the treatment chamber and the PFN have matching impedance,
which is difficult to realize in practice. The maximum voltage delivered to the
chamber is half that of the charged volume on the capacitors.
It is well known that by using the two PFNs one can superimpose voltage
waves and obtain the full charged voltage as the peak discharge voltage
(Valencia, 1987). A PFN using saturable inductors (with a ferromagnetic core
instead of an air core), can compress the pulse width and increase the peak
current flowing through the treatment chamber producing bell-shaped pulses
(Chu & Valencia, 1987). This type of magnetic compression pulse generator is
very attractive since AC power supply can be used instead of DC high voltage
power supply, and the discharge switch may be eliminated (Hofmann, 1984).

Electric pulse field strength and pasteurization requirements

Average electric field strength is defined as the electric potential difference for
two given points in space divided by the distance between them:

(1)
266 Q. Zhang, G. V. Barbosa-C&ovas, B. G. Swanson

In the case of a charged capacitor (C,) discharging through a resistor (R), the
voltage across the food in the treatment chamber decreases exponentially with a
pulse width
r= RC, (2)
The treatment time is defined as t= n. z, where n is the number of pulses and r
is the pulse width.
In microorganisms, the structural change due to a potential difference of
about 1 V gives rise to an irreversible loss of the membrane function (Sale &
Hamilton, 1967; 1968). Mizuno and Hori (1988) observed microscopically
fragmented yeast cells in samples treated with 175 pulses at 20 kV/cm electric
field strength and 160 ,us pulse duration. The cell membrane is fractured at the
lipid-protein junction or protein-protein junction (Castro et al., 1993)
described as the dielectric breakdown of the cell membrane.
The critical electric field strength, E,, above which the microbial survival rate
decreases linearly with respect to the applied electric field in a semi-log plot,
ranges from 3.8 to 14 kV/cm for various microorganisms (Castro et al., 1993). A
critical field strength greater than 30 kV/cm was reported for some spore-
forming bacteria (Hamilton & Sale, 1967). The critical treatment time, t,,
beyond which the microbial survival rate decreases linearly with respect to
treatment time in a log-log plot, ranges from 2 to 80 ps (Hiilsheger et al., 1983).
Hiilsheger et al. (1983) also reported a critical transmembrane voltage ranging
from 0.26 to 2.63 V depending on the species, size, and shape of the micro-
organisms.
The microbial survival rate is related to the field strength and treatment time
by the equation
$=( t/tc)-(fi-wk
(3)
where k is a microorganism constant, E and E, are the actual and critical peak
electric field strengths, respectively, and t and t, are the actual and critical
treatment times, respectively (Hiilsheger et al., 198 1).
To maintain a non-thermal operation, the energy input to the food being
treated must be restricted. The energy density (Q) for exponential decay pulses
is approximated as

v;c,n v;t
Q=,v=-2Rv
where C, is the capacitance of the energy storage capacitor, V, is the initial
charge voltage, n is the number of pulses applied, t is the treatment time, R is the
effective resistance, and v is the volume of treatment chamber.
For square pulses, energy input is approximated as

vzrn Pzn v2t

Q=-=__
V Rv -Rv

where V, Z, t are the voltage, current and pulse width of the square pulses,
respectively.
 Engineering aspects of pulsed electricjield pasteurization 267

To achieve a microbial survival rate of 10m4, Grahl et al. (1992) reported

optimum field strength and energy density for the inactivation of E. coli, where
E= 25 kV/cm and Q= 297 kJ/L.

Electric and dielectric properties of some fluid foods

Fluid foods are primarily composed of water and nutrients such as proteins,
vitamins, triglycerides and minerals. When subjected to a pulsed electric field,
polarization of dipole molecules and bulk movement of charge carriers, such as
ions, induce a capacitive current and a resistive current (Riley & Watson, 1987;
Braithwaite & Weaver, 1990) (Fig. 4(a)). The circuit model for dielectric polari-
zation is a capacitor (Fig. 4(b)) and the circuit model for charge carrying conduc-
tion is a resistor (Fig. 4(c)). The combined circuit is shown in Fig. 4(d) as a
parallel resistor-capacitor (RC) circuit.
Assuming the food material has homogeneous dielectric and electrical
properties, the effective capacitance and effective resistance can be calculated as

and

+d=Ed (7)
aA A

Anode Cathode
+ V i
Voltage = V

3
V
i i
‘c i,
V
f

T i
R R
C
i
C

Area=A

a Dipole molecules
4 Negative charge carrier
Q+ Positive charge carrier

(a) (b) (4 (4

Fig. 4. Fluid food subject to a dynamic voltage. (a) Polarization and electronic
currents. (b) Dielectric circuit model. (c) Electronic circuit model. (d) Combined circuit
model for homogeneous fluid food.
268 Q. Zhang, G. V. Barbosa-Ca’novas, B. G. Swanson

where q, is the permittivity of free space, E, is the relative permittivity, i.e.

dielectric constant of the food material, A is the electrode area, d is the gap
between the two parallel electrodes, u is the conductivity of the food, and p is
the resistivity of the food.
Dielectric and electrical properties of food materials with respect to micro-
wave heating are summarized by Kent (1987). Theoretical modeling of
dielectric and electrical properties of biological materials is reported by Pethig
(1979).
In general, the dielectric constant increases with increasing water content
(E, = 80 for pure water) and decreases with increasing temperature. Skim milk
with 86% water content, for example, exhibits dielectric constant values of
E, = 635, 60.0, and 55.0 at temperatures of 0, 30 and 50°C respectively. Skim
milk with 91% water content exhibits an E, = 70.1, 66.3, and 60.7 at tempera-
tures of 0,30 and 50°C respectively (Kent, 1987).
Conductivity of fluid foods is affected by temperature. Conductivity of whole
milk with 20% total solids is O-55,0*90, 1.25, and 1.66 (S/m) at temperatures of
10, 30, 50, and 70°C respectively. Conductivity of skim milk with 20% total
solids is 0.76, 1.25, l-77, and 2.35 (S/m) at temperatures of 10, 30, 50, and
70°C respectively (Kent, 1987). Conductivity of 2% fat milk assayed with high
voltage pulses is plotted in Fig. 5. Calculated conductivity from Dunn and
Pearlman (1987) for milk is in the range of 0.32 to 0.46 (S/m) for temperatures
from 17 to 43°C.

Dielectric breakdown and prevention

The intent of food pasteurization with PEF is to induce the dielectric break-
down of the cell membrane, not the dielectric breakdown of the fluid food. The
dielectric breakdown of fluid food, termed spark-over, is not desired in PEF
pasteurization.

0.7
l
go
0.6-

10 15 20 25 30 35
Temperature.(C)

Fig. 5. Conductivity of 2% milk measured with pulsed electric fields in 5- 15 kV range.

 Engineering aspects of pulsed electricfield pasteurization 269

When the applied field strength E becomes equal to the dielectric strength EB
of the food, dielectric breakdown of the food inside the treatment chamber
between the electrodes takes place, observed as a spark. The passage of a spark
through a liquid is characterized by (Krasuchi, 1968):

l a large electrical current flow in a narrow channel;

0 a bright luminous spark,
0 the evolution of bubbles of gas;
l the formation of pits on the electrodes; and
l an impulsive pressure through the liquid with an accompanying explosive
sound.

Dielectric breakdown in pure liquids is attributed to several factors (Lewis,

1968), including:

0 electron emission from the cathode surface;

0 collision and ionization of originally neutral molecules; and
6 localized energy input and evaporation of liquid.

In general, vapors exhibit a much lower dielectric strength than do pure

liquids. Dielectric breakdown of pure liquids is also stochastic, affected by the
roughness of the electrode. The smoother the cathode surface, the smaller the
probability of dielectric breakdown (Lewis, 1968).
Dielectric breakdown is attributed to the presence of impurities (Gallagher,
1975). Impurities in liquid, gas or solid, can substantially enhance a local electric
field due to differences in dielectric properties. It is more frequent to have
dielectric breakdown at the gas-liquid or at the liquid-solid interface than in
homogeneous liquids (Krasuchi, 1968).
Dielectric breakdown tests with apple juice and milk conducted at
Washington State University show that local dielectric breakdown takes place
when a gas bubble is present inside the chamber above some electric field level.
Local dielectric breakdown will increase the size of the bubbles already present
and successive applications of PEF wiIl result.in a spark. One possible source
of gas bubbles is dissolved air. When the bulk temperature of the fluid food
increases, the solubility of the gas in the liquid decreases, resulting in gas bubble
formation. A second source for gas bubble formation is local heating which
promotes partial vaporization. An electric field strength of 30-40 kV/cm and
pulse width beyond 20 ,us may result in gas bubble formation after one pulse.
To reduce the probability of dielectric breakdown in foods, the following is
suggested:

l using a smooth electrode surface to minimize electron emission;

0 employing round electrode edges to prevent field enhancement;
l designing the treatment chamber to provide uniform electric field strength
so that the actual applied field strength does not exceed the dielectric
strength of the fluid foods under the test conditions;
0 degassing prior to treatment to eliminate gas bubble formation; and
l pressurizing the fluid food within the treatment chamber to prevent gas
bubble formation.
270 Q. Zhang, G. V. Barbosa-Ca’novas, B. G. Swanson

TREATMENT CHAMBER DESIGN

Treatment chambers are designed to hold the food material during PEF applica-
tion and to house the discharging electrodes. Design of the treatment chamber is
critical to the development of PEF pasteurization technology.
A treatment chamber consists of two electrodes held in position by insulating
material that also form an enclosure containing food materials. Parallel plates,
parallel wires, concentric cylinders, and a rod-plate are the possible electrode
configurations discussed by Hofmann ( 1989). Parallel plates produce uniform
electric field strength distribution in a large usable area and are the most practi-
cal choice. Concentric cylinders, on the other hand, provide smooth and
uniform product flow, and are attractive in industrial applications.

Static treatment chambers

Uniform electric field strength can be achieved by parallel plate-electrodes with

a gap sufficiently smaller than the electrode surface dimension. Disk-shaped,
round-edged electrodes can minimize electric field enhancement and reduce the
possibility of dielectric breakdown of the fluid foods.
Materials selected to construct the treatment chamber need to be washable
and autoclavable. Polysulfone and stainless steel are the recommended insula-
tion and electrode materials, respectively (Hofmamr, 1984) Bushnell et al.
(1993) suggested the use of electro-chemically inert materials, such as gold,
platinum, carbon and metal oxides, to construct the electrodes or electrode
surfaces.
Designing the treatment chamber to facilitate sample filling and removal adds
to the complexity of the chamber design. Since a gas bubble is a potential trigger
of dielectric breakdown, the filling port needs to facilitate the complete expelling
of air during filling.
An acoustic pressure pulse has been observed by the authors while PEF is
applied. A pressure buffering device is recommended for a sealed static treat-
ment chamber. A completely sealed treatment chamber is dangerous. When the
test fluid experiences a spark, high pressure develops rapidly and the chamber
may break apart. A pressure release device must be included in the treatment
chamber design to ensure safety of the operation.

Sale and Hamilton (196 7) static treatment chamber

The treatment chamber designed by Sale and Hamilton (1967) contains two
carbon electrodes backed with brass blocks hollowed out for coolant flow. A
U-shaped polyethylene spacer, about 3 mm in thickness, is placed between the
electrodes to form the chamber, as illustrated in Fig. 6. The three pieces are
clamped together. Changing the thickness of the spacer changes the electrode
gap. The open top configuration limits the maximum electric field strength to
25 kV/cm due to the dielectric strength of air. A maximum of 10 kV pseudo-
square wave pulses with 2-20 ,us pulse width were tested. An E. coli survival
rate of less than 1% was obtained with 10 pulses of 20 kV/cm peak field and 20
,us pulse width (Sale & Hamilton, 1967).
 Engineering aspects ofpulsed electric field pasteurization 271

Air Carbon Electrode

t
Coolant
3

Spacer Polyethylene Spacer

Fig. 6. Static PEF chamber designed by Sale and Hamilton ( 1967). Left: cut-away view
showing the alignment of three parts. Right: U-shaped spacer and coolant connection.

Plexiglas Spacer

Fluid Food SS Electrode

Fig. 7. Cut-away cross-section of the static treatment chamber designed by Dunn and
Pearlman (1987). All three parts are axi-symmetric.

Dunn and Pearlman (1987) static treatment chamber

Dunn and Pearlman ( 1987) designed a circular parallel stainless steel electrode
chamber, as presented in Fig. 7, with a gap of 5 mm and an effective electrode
area of 20 cm’. The electrode spacer was made from acrylic Plexiglas. The
filling port was on the upper electrode. Applied peak field strength ranged from
5 kV/cm to 25 kV/cm. At an electric field strength greater than 30 kV/cm,
sparks were observed (Dunn & Pearlman, 1987).

Grahl et al. (1992) static treatment chamber

The treatment chamber designed by Grahl et al. (1992) is similar to the

treatment chamber by Sale and Hamilton ( 1967). Two rectangular carbon-brass
electrodes were separated by a rectangular Plexiglas frame with a thickness of
05 or 1.2 cm. The effective area of the electrode was 50 cm2 and the maximum
voltage ranged from 5 to 15 kV (30 kV/cm maximum field strength). No cooling
was provided for the electrodes. An open top or filling port construction was
not discussed.
272 Q. Zhang, G. V Barbosa-Cdnovas, B. G. Swanson

Washington State University static treatment chamber

The static treatment chamber designed at Washington State University (WSU) is

presented in Fig. 8. Two round-edged, disk-shaped, stainless steel electrodes are
polished to mirror surfaces. Polysulfone or Plexiglas is used as the insulation
material. The effective electrode area is 27 cm2. The gap between electrodes can
be selected as O-95 cm or O-51 cm. Electric field strengths up to 70 kV/cm were
tested. Cooling of the chamber is provided by circulating water at pre-selected
temperatures through jackets built into the electrodes (not shown).

Continuous flow treatment chambers

A continuous flow chamber can borrow most of its design from the static
chamber design. To maintain a desired flow rate of the food, PEF will be applied
repetitively. The energy input into the chamber requires serious consideration.
When PEF is applied at a high repetitive rate, cooling of the electrode may not
be sufficient to maintain a low temperature operation. Intermediate cooling
between two successive treatment chambers may become necessary.

Dunn and Pearlman (1987) continuous flow treatment chamber

A continuous flow through chamber in which the fluid flow effectively ‘switches’
on and off the electric field (Fig. 9) was outlined by Dunn and Pearlman ( 1987).
A high voltage DC power supply provides high voltage DC to the chamber
continuously without the use of a discharge switch. The fluid switching con-
figuration is likely to be lower in cost than those utilizing pulse generators. At a
high pulse repetitive rate and large scale of operation, the command charging
power supply and high speed electrical switch are the major costs of the pulse
generator. The disadvantage of the fluid switching configuration is the lack of
flexibility. Until the engineering aspects of flow through pasteurization systems
have been addressed, the fluid switching system cannot be implemented.

To high voltage puker

Fig. 8. Schematic drawing of WSU static treatment chamber. Electrodes are

positioned horizontally at work.
 Engineering aspects of pulsed electric field pasteurization 273

To high voltage DC power

in Plow out
>

E4 Position or time

(b)

Fig. 9. A continuous flow treatment chamber sketched by Dunn and Pearlman (1987).
(a) Chamber; (b) electric field versus time or position as fluid switching.

Matsumoto et al. (199 1) converged electric field type treatment chamber

In Japan, Matsumoto et al. ( 199 1) devised a converged electric field type treat-
ment chamber, as illustrated in Fig. 10. Sample liquid was continuously
introduced into the vessel through the hole of the disc electrode. An insulating
plate (Teflon, 1 cm thickness) with small holes was placed between the parallel
disc electrodes to concentrate the electric field. Only the fluid inside the holes of
the insulating plate is subjected to the high voltage pulsed electric field
treatment. The current density at the electrode-liquid interface is held low to
minimize electrolysis and reduce gas bubble formation. Matsumoto et al. ( 199 1)
reported three to six log cycles of microbial inactivation, in S. cerevisiue, E. coli
and B. subtilis, using 30 to 40 kV/cm electric field treatments. A phosphate
buffer was used as the suspension medium.

WSU continuous flow treatment chambers

Parallel plate treatment chamber

Figure 11 illustrates the schematic of a small scale continuous flow chamber that
has been designed, constructed and tested by the WSU PEF research group.
This flow through chamber was modified from the WSU static treatment
chamber by adding baffled flow channels inside the chamber. The designed
operating conditions are as follows: chamber volume 20 or 8 cm3; electrode gap
0.95 or 0.5 1 cm; PEF strength 35 or 70 kV/cm; pulse width 2-15 PCS;pulse
repetition rate of 1 Hz; and food flow rate of 1200 or 600 cm3/min. Cooling of
the chamber is provided by circulating water at a selected temperature through
jackets built in the two stainless steel electrodes.
274 Q. Zhang, G. I/ Barbosa-&novas, B. G. Swanson

Insulating plate
with holes
(teflon)

Electrode
Rz>> 2xR,

(a) Chamber Structure (b) Equivalent Circuit

Fig. 10. Structure and equivalent circuit of converged electric field type treatment
chamber (Matsumoto et al., 199 1).

Coaxial treatment chambers

Coaxial configurations can be easily manufactured and give well-defined electric
field distributions (Hofmann, 1989). The field strength between coaxial
electrodes is defined as

where r is the radius at which electric field is measured. R, and R, are the radius
of the inner and outer electrode surfaces, respectively. The electric field in
coaxial chambers is not uniform and depends on the location. The uniformity of
the electric field is improved when (R, - R,) 4 R,.
Although the electric field is not uniform, coaxial treatment chambers
received much attention due to their uniformity in fluid flow and simplicity in
chamber structure. There are many coaxial treatment chamber designs
including those by Boulart (1983), Hofmamr (1984), Kadyshev et al. (1990),
Masuda ( 199 l), Bushnell et al. ( 1993), and Qin et al. ( 1994). The utilization or
elimination of local electric field enhancement is of major concern in the design
of coaxial treatment chambers. Traditional concentric cylinder electrodes were
used by Boulart (1983) and Hofmarm ( 1984) where electric field enhancement
at the ends of cylinder was not considered. Kadyshev et al. (1990) proposed the
use of a star-shaped inner electrode with two membranes to separate the fluid
foods from the electrodes. Masuda ( 1991) used a non-smooth inner electrode
surface to increase the turbulence of the fluid flow and enhance the electric field.
Bushnell et al. (1993) employed protruded inner electrode surfaces in their
designs to reduce local electric field enhancement.
 Engineering aspects of pulsed electricJield pasteurization 275

To high voltage pulser

Electrode

Electrode

Fluid Food In Fluid Food Out

Fig. 11. Schematic drawing of WSU flow through treatment chamber. Fluid inside the
chamber is baffled to avoid dead spots.

A coaxial treatment chamber developed at WSU (Qin ef al., 1994) used a

protruded outer electrode surface to enhance the electric field within the
treatment zone and reduce the field intensity in the rest of the chamber.
Electrode configuration was obtained by optimizing the electrode design with a
numerical electric field computation. The gap between the coaxial electrodes is
adjustable in the range of 2-6 mm by selecting the diameters of the inner
electrode. Cooling jackets were built into both electrodes to maintain a low
treatment temperature. Figure 12 illustrates the conceptual design of the WSU
coaxial treatment chamber. The design parameters for this chamber are: electric
field intensity 80 kV/cm, flow rate of 1 to 2 L/min, and pulse repetition rate of
10 Hz.

OTHER ENGINEERING ASPECTS

To properly design PEF equipment for food pasteurization, there are several
aspects to be considered, such as peak voltage and current, pulse polarity,
energy input into the chamber, fluid flow rate, pulse repetition rate, and power
requirements for the pulse generator.

Peak voltage and current

Electric field strength is determined by the voltage across the electrodes and the
distance between the electrodes, as defined in eqn (1). Increasing the gap, d, will
require higher voltage to obtain the desired electric field strength.
For a desired decay time constant, t, the capacitance of the energy storage
capacitor, C,, can be calculated using the conductivity of the fluid food, u, the
276 Q. Zhzng, G. V. Barbosa-Ccinovas, B. G. Swanson

I
---- Insulating
1 i 1 Foodout mamial -
I /

Outer electrode
/- connectingto
ground

Food in
2 .

L Inner electrode
Insulating [ connectingto
material high voltage

Fig. 12. Conceptual design of the WSU coaxial treatment chamber. Cooling jackets
are not illustrated.

gap between the electrodes, d, and the area of the electrode surface, A, using
eqns(l),(2)and(6)as

The peak current is a function of capacitance, inductance and resistance in the

discharge circuit. Inductance in the circuit delays and reduces the peak current.
Nevertheless, the maximum peak current is defined as

V VuA
ZMAX=-=- (10)
R d

The actual peak current will be smaller than ZMAXbut close to the calculated
value.

Energy input and temperature change

According to eqns (4) and (5), pulsed electric charge is discharged into the fluid
food in the form of Joule heating. The maximum temperature increase (AT ) can
 Engineering aspects of pulsed electricfieldpasteurization 277

be calculated using a basic thermodynamic equation where no cooling is

provided, or where the pulse repetition rate is very large

AT=-
Q
(11)
PfG

where pf and C, are the density and specific heat of fluid food inside the treat-
ment chamber, respectively. For the optimum setting derived by Grahl et al.
(1992), 27 kV/cm and 297 kJ/L, maximum temperature change can be
calculated as 7 1°C.
If food entered the treatment chamber at room temperature (25”C), the
food would leave the chamber at 96°C when no cooling was provided. To
maintain a low temperature during PEF pasteurization, extensive cooling and/
or low pulse repetition rate is required. In this study, the authors suggest that
high electric field short time pulses be used to reduce the amount of heat input
to the food.

Pulse polarity

Electrolytes, protein and living cells exhibit a net electric charge, and may
migrate to the surface of charged electrode(s), forming a shielding layer when
successive monopolar pulses are applied (Pethig, 1979). This shielding layer will
alter the otherwise uniform electric field. Inside the shielding layer, the electric
field will increase. Tests performed with the WSU static treatment chamber
demonstrated a layer of protein deposit on the anode electrode.
Although more expensive to generate, bipolar pulses are more attractive than
monopolar pulses. Bipolar pulses minimize polar deposit of charged molecules.
When bipolar pulses are used, the shielding layers will not develop, thus making
the pulsed electric field treatment uniform. Therefore, bipolar pulses are
potentially more efficient compared to monopolar pulses in achieving the same
microbial inactivation.

Pulse generator considerations

Pulse repetition rate, continuous power rating, maximum voltage and peak
current are the key parameters in specifying a PEF generator. Optimal configu-
ration should be used to minimize the continuous power rating of the treatment
unit. For a desired number of pulse treatments, n, treatment chamber volume, z),
volumetric flow rate of fluid food, F, and pulse repetition rate, f (in Hz) can be
calculated as

f=F (12)

Continuous power rating, P, for exponential decay pulses, is related to other

parameters as

,~fc,v’_f~v’
(13)
2 2R
278 Q. Zhang, G. V Barbosa-Cknovas, B. G. Swanson

where C, is the energy storage capacitor, T/ is the peak voltage, R is the effective
resistance of fluid food inside the treatment chamber, and z is the decay time
constant.
For square waveform pulses, P is defined as

p-f TV2 (14)

R

where z is the pulse width.

MICROBIAL INACTIVATION TESTS

Extensive microbial inactivation tests have been conducted to validate the

concept of using pulsed electric fields as a non-thermal food pasteurization
process (Pothakamury et al., 1994; Zhang et al., 1994). We present two exam-
ples of PEF microbial inactivation here: one conducted with the WSU static
treatment chamber (Fig. 7); and the other conducted with the WSU continuous
treatment chamber (Fig. 9).
Figures 13 and 14 illustrate the Escherichiu coli (ATCC 11229) survival
fraction affected by the number of pulses applied to the static and continuous
treatment chambers, respectively. Three to four log cycles of reduction in E. co&
viability were demonstrated.
E.coli was inoculated into skim milk with an initial viable count of 80 x 10’
&/ml. The viability of the E. coli before and after PEF treatments was assayed
by counting colony forming units (cfu). Each viability was calculated as the mean
cfu of four replicated plate counts. The survival fraction was determined as the
fraction between the viabilities of the treated and untreated samples. Skim milk
samples were maintained at 15 f 0*2”C during PEF treatment. Peak electric field
intensity of 40 000 V/cm was selected. Energy densities were 12 J/ml pulse and
15 J/ml pulse for the static and continuous chambers, respectively.

1.OOE+OO

\y
1.00E-01 -
\

1 10 loo
Number of pulsesat 40 kV/cm

PEF inactivation of E. coii in skim milk using the WSU static treatment
chamber.
 Engineering aspects ofpulsed electricfield pasteurization 279

Number of pulses at 40 kV/cm

Fig. 14. PEF inactivation of E. cofi in skim milk using the WSU continuous treatment
chamber.

At the current settings, it is estimated that a three stage PEF treatment system
is required to pasteurize milk. Each stage provides 16 pulse treatments and
accomplishes two to three log cycles of microbial reductions. This system
delivers a total energy density of 600 kJ/ml to the milk. As a reference, energy
density for the heating portion of the high-temperature short-time milk pasteuri-
zation process is about 300 J/ml.
It is important to point out that milk is one of the most electrically conductive
fluid foods. To achieve a given field intensity, the energy requirement is inversely
proportional to the electrical conductivity of the food. Therefore, the energy
density for PEF pasteurization of other fluid foods will, most likely, be lower
than that of milk. Further investigation is recommended to improve the efficacy
of PEF pasteurization process.

CONCLUSIONS

Applying pulsed electric field technology to the pasteurization of food materials

poses a new challenge to food and electrical engineers. Critical electric field
strength, treatment times, and desired inactivation of particular microorganisms
are the basic design parameters.
The design of both static and flow-through treatment chambers must achieve
the pasteurization requirement, yet minimize the probability of dielectric break-
down of the foods. Energy input and temperature increase must be considered
in the design. Electrode cooling is effective in preventing gas bubble formation
during the application of PEF, but electrode cooling may not be sufficient to
dissipate all the heat input to the chamber when highly repetitive pulses are
applied. Intermediate cooling between successive treatment chambers is
suggested.
Microbial inactivation tests, performed at Washington State University with
static and continuous treatment chambers, demonstrate the feasibility of using
PEF for food pasteurization.
280 Q. Zhang, G. V. Barbosa-Canovas, B. G. Swanson

ACKNOWLEDGEMENTS

The funding for the WSU PEF project is by the US Army Natick Research
Development and Engineering Center, Natick, MA and the Bonneville Power
Administration, Department of Energy, Walla WalIa, WA.

REFERENCES

Benz, R & Zimmermann, U. (1980). Pulse-length dependence of the electrical break-

down in lipid bilayer membranes. Biochim. Biophys. Acta, 597,637-42.
Benz, R., Beckers, F. & Zimmermann, U. (1979). Reversible electrical breakdown in
lipid bilayer membranes: A charge-pulse relaxation study. J. Membrane Biol., 48,
181-204.
Boulart, J. (1983). Process of Protecting Fluid Product and Installation for This Process
to Work. French Patent 25 13,087.
Braithwaite, N. & Weaver, G. (1990). Electronic Materials. Butterworth Scientific Ltd,
London.
Bryant, G. & Wolfe, J. (1987). Electromechanical stress produced in the plasma
membranes of suspended cells by applied electrical fields. J. Membrane Biol., 96,
129-39.
Bushnell, A. H., Dunn, J. E., Clark, R. W. & Pearlman, J. S. (1993). High Pulsed Voltage
Systems for Extending the Shelf Life of Pumpable Food Products. US Patent
5,235,905.
Castro, A. J., Barbosa-Canovas, G. V. & Swanson, B. G. (1993). Microbial inactivation
of foods by pulsed electric fields. J. Food Proc. Pres., 17,47-73.
Chang, D. C. & Reese, T. S. ( 1990). Changes in membrane structure induced by electro-
poration as revealed by rapid-freezing electron microscopy. Biophys. J., 58, l-l 2.
Chu, E. Y. & Valencia, V. I. (1987). Saturable inductor switch and pulse compression
power supply employing the switch. US Patent 4,707,619.
Dunn, J. E. & Pearlman, J. S. (1987). Methods and apparatus for extending the shelf life
of fluid food products. US Patent 4,695,472.
Gallagher, T. J. (1975). Simple Dielectric Liquids, Mobility Conduction and Breakdown.
Oxford University Press, London.
Grahl, T., Sitzmann, W. & Mkkl, H. (1992). Killing of microorganisms in fluid media by
high-voltage pulses. DECHEMA Biotechnol. Conference Series, 5B, 675-8.
Gupta, R. P. & Murray, W. (1988). Pulsed high electric field sterilization. IEEE Pulsed
Power Conf. Record, pp. 58-64.
Hamilton, W. A. & Sale, A. J. H. (1967). Effects of high electric fields on microorganisms
II. Mechanism of action of the lethal effect. Biochim. Biophys. Acta, 148,789-800.
Hofmann, G. A. ( 1989). Cells in electric fields, physical and practical electronic aspects
of electro-cell fusion and electroporation. In Electroporation and Electrofusion in
Cell Biology, ed. E. Neuman. A. E. Sowers SCC. A. Jordan. Plenum Press, NY.
Hofmann, G. A. & Evans, E. A. (1986). Electronic genetics - physical and biological
aspects of cellular electromanipulation. IEEE Eng. Med. Biol., 5,6-25.
Hofmann, G. A. (1984). Microflora reduction in liquids with pulsed electric fields.
internal Technical Report, Biotechnologies and Experimental Research, Inc., San
Diego.
Hiilsheger, H., Potel, J. & Niemarm, E. G. (1983). Electric field effects on bacteria and
yeast cells. Radiat. Environ. Biophys., 22, 149-62.
Hiilsheger, H., Potel, J. & Niemann, E. G. (1981). Killing of bacteria with electric pulses
of high field strength. Radiat. Environ. Biophys., 20,53-65.
Hiilsheger, H. & Niemann, E. G. ( 1980). Lethal effects of high-voltage pulses on E. coli
K12. Radiat. Environ. Biophys., 18,281-g.
 Engineering aspects of pulsed electric field pasteurization 281

Jacob, H. E., Forster, W. & Berg, H. ( 1981). Microbiological implications of electric field
effects II. Inactivation of yeast cells and repair of their cell envelope. 2. Allg. Mikro-
biol., 21,225-33.
Jayaram, S., Castle, G. S. P. & Margaritis, A. (1992). Kinetics of sterilization of Lacto-
bacillus brevis by the application of high voltage pulses. Biotech. Bioeng., 40 (1 l),
1412-20.
Kadyshev, G. G., Kvanin, Yu. V., Fjodorov, F. A. & Tshekov, A. V. ( 1990). Method for
Sterilizing Food and a Device for its Implications. USSR Patent SU 1611320 Al.
Kent, M. (1987). Electrical and Dielectric Properties of Food Materials. Science and
Technology Publishers, England.
Krasuchi, Z. (1968). Breakdown of commerical liquid and liquid-solid dielectrics. In
High Voltage Technology, ed. L. L. Alston. Oxford University Press, London.
Lewis, T. J. (1968). The electric conduction and strength of pure liquids. In High Voltage
Technology ed. L. L. Alston. Oxford University Press, London.
Masuda, S. (1991). Destruction of Yeast in Fluid Foods. Japanese Patent 3-98565.
Matsumoto, Y., Satake, T., Shioji, N. & Sakuma, A. (1991). Inactivation of micro-
organisms by pulsed high voltage applications. Conf Rec. of IEEE Industrial Applica-
tions Society Annual Meeting, pp. 652-9.
Mertens, B. & Knorr, D. (1992). Developments of nonthermal processes for food preser-
vation Food Technol., 46 (S), 124-33.
Mizuno, A. & Hori, Y. (1988). Destruction of living cells by pulsed high-voltage applica-
tions. Trans. IEEE Ind. Appl., 24 (3), 387-94.
Pethig, R. (1979). Dielectric and Electronic Properties of Biological Materials. John Wiley
and Sons, NY.
Pothakamury, U. R., Monsalve-Gonzalez, A., Barbosa-Canovas, G. V. & Swanson, B. G.
( 1994). Inactivation of Escherichia coli and Staphylococcus aureus in model foods by
pulsed electric field technology. Food Preservation 2ooO Proceedings, Natick, MA.
Qin, B.-L., Zhang Q., Barbosa-Cbnovas, G. V. & Swanson, B. G. (1994). Pulsed electric
field pasteurization with a coaxial treatment chamber. CIFST/ICSTA 37th Annual
Conference, P103, May 1994, Vancouver BC, Canada.
Riley, T. & Watson, A. (1987). Polarography and Other Voltammetric Methods. John
Wiley and Sons, NY.
Sale, A. J. H. & Hamilton, W. A. (1968). Effects of high electric fields on microorganisms
III. Lysis of erythrocytes and protoplasts. Biochim. Biophys. Acta, 163,37-43.
Sale, A. J. H. & Hamilton, W. A. (1967). Effects of high electric fields on microorganisms
I. Killing of bacteria and yeasts. Biochim. Biophys. Acta, 148,781-g.
Somkuti, G. A. & Steinberg, D. H. (1988). Genetic transformation of Streptococcus
thermophilus by electroporation. Biochimie, 70,579-85.
Sower, A. E. (1986). A long-lived fusogenic state is induced in erythrocyte ghosts by
electric pulses. J. Cell Biol., 102, 1358-62.
Sower, A. E. (1985). Movement of a fluorescent lipid label from a labeled erythrocyte
membrane to an unlabeled erythrocyte membrane following electric-field-induced
fusion. Biophys. J., 47,519-25.
Tsong, T. Y. (1991). Electroporation of cell membranes. Biophys. J., 60,297-306.
Valencia, V. I. (1987). Symmetrically Charge Pulse-Forming Circuit. US Patent
4,684,820.
Zhang, Q., Monsalve-Gonzalez, A., Barbosa-Cannovas, G. V. & Swanson, B. G. ( 1994).
Inactivation of E. coli and 5. cerevisiae by pulsed electric fields under controlled
temperature conditions. Trans. ASAE, 32 (2), 581-7.
Zirnmermann, U. (1986). Electrical breakdown, electropermeabilization and electro-
fusion. Rev. Phys. Biochem. Pharmacol., 105,176-256.
Zimmermann, U., Pilwat, G., Beckers, F. & Riemann, F. (1976). Effect of external
electrical fields on cell membranes. Bioelectrochem. Bioenerg., 3,58-83.