3862

J. Phys. Chem. B 2003, 107, 3862-3870

Phenomenological Theory of Low-Voltage Electroporation. Electric Field Calculations

Istva n P. Suga r,*, James Lindesay, and Robert E. Schmukler
Departments of Biomathematical Sciences and Physiology/Biophysics, Mount Sinai School of Medicine,
New York, New York 10029, Computational Physics Laboratory, Howard UniVersity, Washington, DC 20059,
Stanford Linear Accelerator Center, Stanford UniVersity, Stanford, California 94309, Pore2 Bioengineering,
19212 Orbit DriVe, Gaithersburg, Maryland 20879, and Drexel UniVersity, Philadelphia, PennsylVania 19104
ReceiVed: October 31, 2002; In Final Form: February 6, 2003

In common electroporators, cells can be transfected with foreign genes by applying a 150-700 V pulse on
the cell suspension. Because of Joule heating, the cell survival rate is 10-20% in these elecroporators. In a
recently developed electroporator, termed the low-voltage electroporator (LVEP), cells are partially embedded
in the pores of a micropore filter. In LVEP, cells can be transfected by applying 25 V or less under normal
physiological conditions at room temperature. The large increase in current density in the filter pores, produced
by the reduction of current shunt pathways around each embedded cell, amplifies 1000-fold the local electric
field across the filter and results in a high-enough transmembrane voltage for cell electroporation. The Joule
heat generated in the filter pore is quickly dissipated toward the bulk solution on each side of the filter, and
thus cell survival in the low-voltage electroporator is very high, about 98%, while the transfection efficiency
for embedded cells is above 90%. In this paper, the phenomenological theory of LVEP is developed. The
transmembrane voltage is calculated along the membrane of the cell for three different cell geometries. The
cell is either fully, partially, or not embedded in the filter pore. By means of the calculated transmembrane
voltage, the distribution of electropores along the cell membrane is estimated. In agreement with the
experimental results, cells partially embedded in the filter pore can be electroporated by as low as 1.8-3.5
V of applied voltage. In the case of 25 V applied voltage, 90% of the cell surface can be electroporated if the
cell penetrates further than half of the length of the filter pore.

1. Introduction
Biological membranes are known to become transiently more
permeable by the action of short electric field pulses1-4 when
the threshold value of the transmembrane voltage, about 0.5-1
V, is exceeded. (The transmembrane voltage is defined by the
potential difference between the inner and outer surfaces of the
cell membrane.) This phenomenon is called electroporation or
electropermeabilization, and it can be used to transfect cells with
foreign genes.5 Electroporation of biological cells is commonly
carried out in a cell suspension using a parallel plate capacitor
chamber.6 The field between the plates is essentially homogeneous because the cell density is low. The voltage required for
electroporation varies from 150 to 700 V across a 0.2 cm gap
of physiologic solution (0.15 M NaCl). The applied voltage
depends on factors such as the spacing between the capacitor
plates, the cell type, and solution temperature. The field strengths
needed for suspension electroporation normally vary between
750 and 2000 V/cm. The resulting current produced by these
fields in the low-resistivity physiologic solution is in the range
of 25-100 A. Substantial Joule heating, electrode products, and
solution electrolysis are byproducts produced by these fields in
cell suspension,7 and thus the cell survival rate is low. For
COS-7 cells, the survival rate in suspension experiments varies
from 10% (ref 8) to 20% (ref 9). These survival rates are in
agreement with the rates quoted by commercial companies for
their systems (personal communications with BTX Corp., Life
Technologies, Inc., and Savant/E-C Apparatus, Inc.)

Mount Sinai School of Medicine.

Howard University and Stanford University.
Pore2 Bioengineering and Drexel University.

Recently, an alternative to cell suspension electroporation

(SEP), the method of low-voltage electroporation (LVEP), was
introduced.10-16 A schematic of the low-voltage electroporator
(LVEP) is shown in Figure 1.17 The vertical chamber consists
of two mirror-image halves. The inside diameter of the
cylindrical chamber is 1 cm, and cylindrical porous carbon
electrodes enclose the upper and lower ends of the chamber.
The carbon electrodes apply the input signal and are separated
by 2 cm. This produces a cylindrical measurement volume with
dimension of 1 cm in diameter and 2 cm in length. A
polycarbonate Nuclepore filter (its plane aligned perpendicular
to the symmetry axis of the chamber) is sealed into the center
of the chamber, and the cells are then embedded in the filter
pores (see enlargement in Figure 1) by using a hydrostatic
pressure of 25-30 mmHg. In LVEP, as low as 2-25 V of
applied voltage is sufficient to induce electroporation because
40% of the applied voltage drops in the 13 m long micropores
of the filter.17 The average field across the entire chamber for
10 V input is less than 5 V/cm, while the average field across
the filter with cells is about 3000 V/cm. Thus, the field in a
LVEP is highly inhomogeneous, amplified about 1000 times
in conjuntion with the increase in current density through the
filter pores. However, the current produced in this system is
only 25-50 mA. The bulk temperature increase caused by a
90 ms pulse of 10 V is less than 0.003 C, and the local Joule
heating generated in the filter pore is dissipated in less than 0.3
ms (ref 16). Because of the negligibly small Joule heating, the
cell survival rate is about 98% (refs 12 and 16).
The development of the phenomenological theory of SEP
started 30 years ago. The transmembrane voltage around a

10.1021/jp022343k CCC: $25.00 2003 American Chemical Society

Published on Web 03/26/2003

Theory of Low-Voltage Electroporation

J. Phys. Chem. B, Vol. 107, No. 16, 2003 3863

We notice that the transmembrane voltage, V(,t), can be
calculated by means of eq 1 not only at subcritical pulses but
also at supracritical pulses if the cell membrane is assumed to
be unporated. In reality, pore formation takes place where the
absolute value of this calculated transmembrane voltage exceeds
the critical voltage, Vcr. In the case of supracritical electric
pulses, the phenomenological theory of SEP has been developed
by Kinoshita and co-workers. They assumed that the probability
of pore formation is directly proportional to |V(,t)| - Vcr, where
V(,t) is defined by eq 1. Thus, at any time t after the application
of the supracritical pulse, the excess specific conductivity in
the porated region of the membrane, m(,t), is

|V(,t)| - Vcr
m(,t) ) m(o,t)
)
|V(o,t)| - Vcr
|V()|f(t) - Vcr
m(o,t)
(2)
|V(o)|f(t) - Vcr

Figure 1. Schematic of the low-voltage electroporator (LVEP). Shaded

areas mark the cylindrical carbon electrodes at the top and the bottom
of the vertical chamber. The chamber is divided by a micropore filter.
The plane of the filter aligned perpendicular to the symmetry axis of
the cylindrical chamber is marked by a heavy solid line. The inset shows
a magnified part of the filter with cells partially embedded in the
micropores. Note that in the figure the chamber is stretched along its
symmetry axis. In reality the chambers inner length and diameter is 2
and 1 cm, respectively.

spherical cell placed into a constant, subcritical electric field,

V(), was determined by solving the Laplace equation.18,19 The
field is subcritical as long as the absolute value of the
transmembrane voltage is below the critical value, Vcr 0.5-1
V. When the field is switched on at time t ) 0, the steady-state
transmembrane voltage, V(), is attained after the charging of
the membrane. In this case, the solution can be separated to the
steady-state and transient part, f(t), as follows:

V(,t) ) [V()][f(t)] ) [1.5REo cos()][1 - e-t/]

(1)

where R is the radius of the cell, Eo is the field strength far

from the cell, is the angle (azimuthal angle) between the
direction of Eo and the vector directed from the center of the
cell to the considered membrane segment, and is the
membranes charging time constant. According to eq 1, the
absolute value of V() is maximal at the poles of the cell, while
it is zero at the equator.
In the case of supracritical fields, when electroporation takes
place, however, there is no closed form solution of the Laplace
equation. The azimuthal dependence of the transmembrane
voltage, Vexp(,t), was measured on a spherical sea urchin egg
stained with voltage-sensitive fluorescent dye at different time
points after the application of a supracritical electric field.20-22
These measurements showed that (i) those regions of the cell
membrane that would experience supracritical transmembrane
voltage appear to be porated within less than 1 s and (ii) the
transmembrane voltage remains symmetrical around the z-axis
(the axis going through the poles of the egg), although it
decreases significantly within a certain range around the pole.

where at ) o |V()| assumes its global maximum. By using

the above function for the excess membrane specific conductivity, Hibino et al.21,22 solved the Laplace equation numerically.
The solution was in accordance with the measured transmembrane voltage, Vexp(,t). At any given time, t, the excess specific
conductivity of the membrane at the pole, m(0,t), was the
only adjusted parameter of the theory of SEP. The analysis of
the experimental data revealed that m(0,t) gradually increased
as long as the electric field was on. After switching off the field,
the decrease of m(0,t) could be described by two exponentials
with time constants of 7 and 500 s. We note that the above
theory is more complicated in the case of an asymmetric
electroporation model.22
In this paper, after defining the geometrical and material
parameters of the system in the Model section, we present the
solutions of the Laplace equation in the Results section for
different lengths of cell penetration into the filter pore. In the
Discussion section, the calculated electric field is compared with
the field around a single spherical cell in cell suspension, and
the importance of the current density amplification (CDA) is
discussed. The distribution of the electropores along the
membrane is calculated for different cell geometries. The
calculated minimal applied voltage needed to induce electroporation is compared with the available experimental result, and
the efficiency of electroporation is defined and calculated for
different cell geometries.
2. Model
2.1. Geometry of the Model of Low-Voltage Electroporator. The LVEP can be modeled by N electrically identical,
parallel units, where N is the total number of the filter pores.
One unit consists of a filter pore and its surrounding. The pore
is cylindrical (pore length is 13 m and pore radius is 1 m),
and its symmetry axis is perpendicular to the surface of the
filter. The unit itself is assumed to be a cylinder too; its
symmetry axis (z axis) coincides with the symmetry axis of the
pore, while its cross-sectional area, 254.5 m2/unit, is equal to
the average filter area per filter pore. The cross section of a
unit along the z axis is shown in Figure 2a. The gray area marks
the filter pore and the bulk regions on both sides of the filter,
while white areas represent the filter around the pore. The double
solid line shows the cell membrane. The vertical, z, axis is the
symmetry axis of the unit, while the horizontal axis measures
the radial distance, r, from the symmetry axis. The unit contains
one cell of surface area 137.3 m2, which is the average surface

3864 J. Phys. Chem. B, Vol. 107, No. 16, 2003

Figure 2. The geometry of a unit of the low-voltage electroporator:

(A) schematic of cell partially embedded in the filter pore; (B)
transmission electron micrograph of human erythrocyte in the filter
pore (the finger length is about 8 m); (C) schematic of cell fully
embedded in the filter pore; (D) schematic of cell out of the filter.
White area represents the filter, while gray area marks the bulk solution
regions on both sides of the filter and the filter pore. Double solid line
shows the cell membrane.

area of an erythrocyte.23 In Figure 2a,c,d, the cell is partially

embedded in the filter pore, fully embedded in the filter pore,
and outside the pore, respectively. In each case, the center or
symmetry axis of the cell coincides with the z axis of the unit.
Outside the filter pore, the cell is assumed to be spherical (Figure
2d). The geometry and location of the cell can be given by its
radius (r2 ) 3.305 m) and the coordinate of its center, z2. When
the cell is fully embedded in the filter pore its geometry is
assumed to be two truncated spheres connected with a cylindrical tube (Figure 2c). In this case, the geometry and location of
the cell can be described by the center and radius of the lower
truncated sphere (z1 and r1), the outer radius of the tube (rt )
0.9 m), and the center and radius of the upper truncated sphere
(z2 and r2). In the case of partially embedded cells, the same
parameters define the location and geometry of the cell with
the restriction that one of the truncated spheres is a hemisphere
(at the tip of the finger) of radius rt. The part of the cell that is
penetrated into the filter pore is called the finger of the cell.
The geometry of the cell partially and fully embedded in the
filter pore has been confirmed by direct observation.10,14,17 Figure
2b shows the transmission electron micrograph of a human
erythrocyte partially embedded in a filter pore. When physi-

Sugar et al.
ologic solution is in the extracellular space, the finger length
of the embedded erythrocyte cell is about 8 m (Figure 2b).
The flaccidity of the cell and thus the finger length can be
modified by changing the salt concentration of the extracellular
space.
The geometry of the model system agrees almost completely
with the geometry of the LVEP. There are only three aspects
in the geometry of the model that differ from the experimental
geometry: (1) the membrane thickness of the cell in the model
is 0.1 m, while in reality the thickness of the cell membrane
is about 0.01 m (ref 23); (2) the thickness of the narrow passage
between the finger surface and the filter pore wall is 0.1 m,
while in reality it is estimated to be 0.01 m (see Appendix 1
and ref 17); (3) the thickness of the bulk region on each side of
the filter is 13 m, while in reality it is 1 cm. In the case of this
model geometry, we are able to obtain reliable numerical
solutions of the partial differential equation of the electric
potential. The effect of the deviations 1 and 2 on the transmembrane voltage is negligibly small (see Appendix 2), while
deviation 3 can be easily corrected. It was shown by Schmukler17
that 40% of the applied voltage drops in the filter. Thus in our
model calculations, the voltage applied very close to the filter
surfaces represents 40% of the voltage applied to the capacitor
plates of the LVEP chamber. For example, if 10 V is applied
to the LVEP unit, the corresponding voltage applied to the LVEP
chamber is 25 V.
2.2. Laplace Equation of the Model of Low-Voltage
Electroporator. In this section, the partial differential equation
of the electric potential of a unit of the LVEP is described.
2.2.1. Boundary Conditions. In every calculation, the potential
applied at z ) 26 m, the top of the cylindrical unit, is u(r,26)
) 10 V, while the applied potential at z ) -13 m, the bottom
of the cylindrical unit, is u(r,-13) ) 0 V. A Neumann-type
boundary condition was utilized at every other boundary (the
wall of the filter pore, the top and bottom surfaces of the filter,
the borders to the neighbor units, and the symmetry axis of the
unit) because the normal component of the current to each of
these boundaries is zero:

-n(fru) ) nj ) 0

(3)

where n is the normal vector to the surface of the boundary, r

is the radial distance from the symmetry axis (z-axis), f is the
electric conductivity of the extra- and intracellular space, and j
is the current density.
2.2.2. Laplace Equation in Inhomogeneous Medium. The
steady-state electric potential in an axially symmetric unit of
the LVEP can be determined by solving the following Laplace
equation:

u
r
+ r
)0
r
r
z
z

( ) ( )

(4)

2.2.3. The Matching Conditions. The electric conductivity,

) (r,z), is piecewise continuous and is discontinuous on the
outer and inner surfaces of the cell membrane. In our calculations, the same conductivity, f, is taken in the extra- and
intracellular regions, while the conductivity of the cell membrane
is m. The conductivity ratio of the extra- or intracellular space
(0.15 M NaCl) to the human erythrocyte membrane at 25 C is
f/m ) 2.3 104 (ref 23), while the conductivity of the filter
is assumed to be zero. The matching conditions on the
membrane surface of normal vector n are

uf ) um

(5)

Theory of Low-Voltage Electroporation

uf
um
) m
n
n

J. Phys. Chem. B, Vol. 107, No. 16, 2003 3865

(6)

2.2.4. Numerical Solution of the Laplace Equation. The

numerical solution of the Laplace equation is obtained by using
the partial differential equation (PDE) toolbox of the Matlab
program (The Math Works, Inc.). This program package is
capable of calculating the electric potential, u, at every (r,z)
point of our model system, that is, capable of solving a 3D
Laplace equation when the system possesses axial symmetry.
The program uses the finite element method to solve PDEs. It
approximates the two-dimensional, (r,z), computational domain
with a union of triangles. The triangles form a mesh. The
triangular mesh is automatically generated and can be further
refined. Before solving the PDE, to get fine meshes everywhere
in the membrane, the program refines the original mesh twice.
To solve the Laplace equation, the default parameters of the
program are utilized.
3. Results
The electric field in a unit of the LVEP was calculated in the
case of different cell positions. The cell position is characterized
by zmin the z-coordinate of the bottom of the cell. Table 1 lists
the geometrical parameters of the cell at each calculated cell
position.
In Figures 3-5, the contour plots of the calculated potential,
u, are shown at three different cell positions. Because of the
LVEP units axial symmetry, the calculated potential is symmetric. Thus in Figures 3-5, it is sufficient to show only half
of the LVEP unit. In the figures, the consecutive contour lines
are 0.5 V apart from each other. To make the contour lines
more visible in the membrane and in the narrow passage, the
figures are stretched in the direction of the horizontal axis, and
thus the shape of the cell is distorted. These plots show that the
strongest electric field in the cell membrane is at r ) 0 and z
) zmin, that is, at the bottom of the cell. Note that there is another
local maximum of the density of the contour lines in the
membrane at the top of the cell, that is, at r ) 0 and z ) zmax;
however, the respective electric field strength is lower than the
field strength at the bottom of the cell membrane.
The transmembrane voltage (the potential at the inner
membrane surface minus that at the outer membrane surface)
has been calculated along the cell membrane. In Figure 6a-c,
the transmembrane voltage, V(z), is plotted against the z
coordinate of the membrane segment for cases when the cell is
out of the filter and partially and totally embedded in the filter
pore, respectively.
4. Discussion
4.1. Transmembrane Voltage. Figure 6a-c shows that, at
every considered cell position, the transmembrane potential is
maximal at zmin, the bottom of the cell. If the cell is fully
embedded in the pore, the transmembrane voltage is essentially
constant along the cell membrane protruding at the bottom of
the filter pore. The transmembrane voltage then linearly
decreases along the tubular section of cell. The transmembrane
voltage changes sign at the point where the increasing potential
along the outer membrane surface becomes equal with the
potential inside the cell. The transmembrane voltage stops
decreasing and becomes constant along the cell membrane
protruding at the top of the filter pore. When the cell is partially
embedded in the filter pore, the transmembrane voltage changes
similarly along the tubular and protruding section of the cell.
The change of the transmembrane voltage along the half-

TABLE 1: Geometrical Parameters of the Cell at Different

Cell Positionsa
zmin
(m)
14.822
14.322
13.822
13.322
12.822
11.1
10.1
9.1
8.1
7.1
6.1
5.1
4.1
3.1
2.1
1.1
0.1
-0.9
-1.203 71
-1.517 33
-1.835 13
-2.146 26
-2.465 76
-2.790 84

z1 b
(m)

12
11
10
9
8
7
6
5
4
3
2
1
0
-0.26
-0.48
-0.68
-0.86
-1.04
-1.22

r1 b
(m)

z2 c
(m)

r2 c
(m)

cell position

0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.943 71
1.037 33
1.155 13
1.286 26
1.425 76
1.570 84

18.127
17.627
17.127
16.627
16.127
16.03
15.96
15.89
15.81
15.73
15.65
15.57
15.48
15.40
15.29
15.20
15.10
15.00
14.884
14.819
14.763
14.673
14.563
14.417

3.305
3.305
3.305
3.305
3.305
3.2092
3.1382
3.0657
2.9916
2.9153
2.8372
2.7566
2.6738
2.5886
2.5002
2.4085
2.3133
2.2141
2.184 57
2.133 15
2.065 23
1.985 95
1.886 83
1.767 84

ouside
ouside
ouside
ouside
ouside
partially embedded
partially embedded
partially embedded
partially embedded
partially embedded
partially embedded
partially embedded
partially embedded
partially embedded
partially embedded
partially embedded
partially embedded
partially embedded
fully embedded
fully embedded
fully embedded
fully embedded
fully embedded
fully embedded

a The surface area of the cell, S ) 137.3 m2, is related to the z and
i
ri parameters as follows:

S ) (h12 + rt2) + 2rt[(z2 + r2 - h2) (z1 - r1 + h1)] + (h22 + rt2) (8)

where the first and third terms are the surface area of the truncated
sphere on the bottom and top of the cell, respectively, while the second
term is the surface area of the connecting tube of radius rt. The height
of the ith truncated sphere is hi ) ri + xri2 - rt2, where i ) 1 or 2.b z1
and r1 are the centers z coordinate and the radius of the truncated
sphere on the bottom of the cell. c z2 and r2 are the centers z coordinate
and the radius of the truncated sphere on the top of the cell.

spherical section of the finger is shown in Figure 6d, where the

relative transmembrane voltage, V()/V(0), is plotted against
the azimuthal angle, (the angle between the z-axis and the
vector directed from the center of the hemisphere to the
considered membrane segment; ) 0 at the bottom of the cell).
Each curve belongs to a different cell position. When the halfspherical section of the finger protrudes at the bottom of the
filter pore (zmin ) -0.9 m), the relative transmembrane voltage
is practically independent of the azimuthal angle (top curve in
Figure 6d). However, when the half spherical section is within
the filter pore, the relative transmembrane voltage decreases with
increasing azimuthal angle, and the decrease becomes steeper
with decreasing finger length. It is important to mention,
however, that none of the angular dependences are as steep as
the cos() function (dashed line in Figure 6d), which is the
angular dependence of the relative transmembrane voltage of a
spherical cell placed in a homogeneous electric field and in an
electrically homogeneous medium (see eq 1). The above result
suggests that the effective membrane area for electroporation
increases with increasing finger length and in the case of long
fingers pores can form practically anywhere in the half-spherical
section of the finger if V(0) is larger than the critical transmembrane voltage. In the case of 10 V applied to the LVEP
unit, that is, Vapp ) 25 V applied to the capacitor plates of the
LVEP chamber, the transmembrane voltage at the bottom of
the finger is larger than the critical voltage at every finger length
(see Figure 6b). For comparison, we note that if there is no
filter in our electroporator the same applied voltage results in

3866 J. Phys. Chem. B, Vol. 107, No. 16, 2003

Figure 3. Calculated electric potential when the cell is out of the

filter: (A) solution for the entire unit; (B) solution at the bottom of the
cell. The contour lines are 0.5 V apart from each other. The cell position
is zmin ) 13.322. The voltage applied to the capacitor plates of the
LVEP chamber is Vapp ) 25 V.

Figure 4. Calculated electric potential when the cell is partially

embedded in the filter pore: (A) solution for the entire unit; (B) solution
at the bottom of the cell. The contour lines are 0.5 V apart from each
other. The cell position is zmin ) 5.1. The voltage applied to the capacitor
plates of the LVEP chamber is Vapp ) 25 V.

only V(0) 1.5Vappr2/L ) 6.2 mV (where r2 ) 3.3 m is the

radius of the cell and L ) 2 cm is the spacing between the
capacitor plates) transmembrane voltage at the poles of the
spherical cell, eq 1. It is the current density amplification (CDA)
in the filter pore that produces about a 1000-fold increase of
the transmembrane voltage relative to the cell suspension
electroporation. CDA estimated by the ratio of the surface area
of the filter per pore (254.5 m2) and the cross-sectional area

Sugar et al.

Figure 5. Calculated electric potential when the cell is fully embedded

in the filter pore: (A) solution for the entire unit; (B) solution at the
bottom of the cell. The contour lines are 0.5 V apart from each other.
The cell position is zmin ) -2.146 26. The voltage applied to the
capacitor plates of the LVEP chamber is Vapp ) 25 V.

of a narrow passage (0.0753 m2, see Appendix 1) is about

3400. Note, that the actual CDA is smaller because part of the
electric current flows through the cell membrane (Appendix 2).
The finding that a transmembrane voltage change occurs
along the finger of a filter-embedded cell seems somewhat
counterintuitive, on the basis of the case of a spherical cell in
suspension. The cell membrane of the finger, except for the
hemisphere at the end of the finger, is parallel to the direction
of the electric field. Our initial assessment, based on angular
dependence of the transmembrane voltage along the spherical
cell, was that the transmembrane voltage change along the finger
length should therefore be zero. However, the finding that the
transmembrane voltage changes along the finger can be
explained by using concepts from spatial amplification.24 The
differences in boundary conditions between a cell embedded
in a pore and a cell in suspension explains this finding. Spatial
amplification is defined as the amplification of the electric field
across the cell membrane for a cell in suspension at low
frequencies when the cell membrane becomes nonconductive.
Essentially in spatial amplification, the electric field for the
conductive path through the cell integrates along the length of
the cell parallel to the field direction. The conduction path
through the cell differs from the external conduction pathway
because of the presence of cell membranes at each end of the
cell. Whereas the voltage drop along the cell in the external
medium is linear, uniform and very small, the voltage drop
through the cell is not uniform. Nearly the entire voltage drop
in the conduction pathway through the cell occurs across the
cell membranes at the ends of the cell. This is because, in
comparison, there is a negligible voltage drop in the intracellular
solution of high conductivity or low resistance. The voltage drop
in the external medium is very small, so the potential external
to the membrane is essentially constant. The electric field
strength across the cell membranes at either end of the cell is
amplified by (1/2)(cell length/membrane thickness). Thus, the
transmembrane voltage change for a suspended cell is maximal
at the two opposite cell ends. In comparison, for a cell embedded

Theory of Low-Voltage Electroporation

J. Phys. Chem. B, Vol. 107, No. 16, 2003 3867

Figure 7. Distribution of electropores along the cell membrane. Spatial

distribution of electropores, p(z,t)/a(t), is calculated by eq 7 for three
different cell geometries. The solid line represents the cell out of the
filter pore, zmin ) 12.8 m; the dashed line represents the cell partially
embedded in the filter pore, zmin ) 2.1 m; the dotted line represents
the cell fully embedded in the filter pore, zmin ) -2.79 m. The voltage
applied to the capacitor plates of the LVEP chamber is Vapp ) 25 V.
The critical voltage is Vcr ) 1 V.

in a pore, the boundary conditions are reversed with respect to

a cell in suspension. In this case, the voltage drop in the
extracellular space along the finger in the pore is also linear
and uniform, but in contrast to a suspended cell, this voltage
drop is significant and not small. The significant external voltage
drop results from the high resistance of the narrow passage
around the finger in a pore. In this case, at all frequencies, a
substantial voltage drop exists in the external conduction
pathway. The conduction pathway through the cell is also
different compared to the cell suspensions. For an embedded
cell, there is a relatively small voltage drop across the membrane
of the cell protruding out of the filter pore. This is because the
capacitance of the protruding section is about 10 times larger
than the capacitance of the hemisphere at the tip of the finger.
A very small voltage drop also occurs inside the cell. This means
that the situation is different from the situation for a cell
suspension in that the voltage drop through the cell before the
tip of the finger is small, while the voltage drop in the external
pathway is large. This difference produces a significant transmembrane voltage change along the finger that would not occur
in cell suspensions. The specialized geometry of a cell embedded
in an insulating filter is such that the transmembrane voltage
along a cell membrane perpendicular to the filter surface can
be nonzero and it can change in response to an applied electric
field.
4.2. Distribution of Electropores. Electroporation takes place
where the absolute transmembrane voltage of the unporated cell,
|V(z,t)|, is higher than the critical voltage, Vcr. After the charging
of the membrane, the temporal and spatial distribution of the
electropores in the cell membrane can be given by the following
expression:20-22
Figure 6. Calculated transmembrane voltage at different cell positions.
In Panels A, B, and C, the transmembrane voltage is plotted against
the z coordinate of the membrane segment. In panel A, the cell is out
of the filter. The five different cell positions are listed in Table 1. In
panel B, the cell is partially embedded in the filter pore. The 13 different
cell positions are listed in Table 1. In panel C, the cell is fully embedded
in the filter pore. The six different cell positions are listed in Table 1.
Panel D shows the angular dependence of the relative transmembrane
voltage, V()/V(0), in the half-spherical section of the finger of a
partially embedded cell. The solid lines from top to bottom belong to
cells of decreasing finger length (see labels). The 13 different cell
positions are listed in Table 1. The dashed line represents angular
dependence of the relative transmembrane voltage in the case of a
spherical cell placed into a homogeneous field (see eq 1). The voltage
applied to the capacitor plates of the LVEP chamber is Vapp ) 25 V.

|V(z)| - Vcr
if |V(z)| > Vcr
p(z,t) )
Vappl
0
otherwise
a(t)

(7)

where p(z,t) dz is the probability of finding a porated region in

the membrane segment from z to z + dz and the proportionality
factor a(t) gradually increases until the supracritical electric field
is on. The spatial distribution of the electropores can be
characterized by p(z,t)/a(t). By using the transmembrane voltage
curves, V(z), in Figure 6a-c, we have calculated the spatial
distribution of the electropores for three different cell geometries
(Figure 7). The electropore density is constant along the

3868 J. Phys. Chem. B, Vol. 107, No. 16, 2003

Figure 8. The potential at the inner surface of the membrane plotted

against the z coordinate of the membrane segment. The curves belong
to six fully and 13 partially embedded cell positions listed in Table 1.
The z coordinate of the leftmost point of each curve is zmin, characterizing the cell position. The voltage applied to the capacitor plates of
the LVEP chamber is Vapp ) 25 V.

protruding sections of the cell membrane and linearly decreases

toward the inside of the filter pore (see dashed and dotted lines
in Figure 7). The bottom of the nonembedded cell is electroporated, but then the electropore density sharply drops to zero
(solid line in Figure 7). By using Figure 6d, one can also
calculate the azimuthal dependence of the pore density in the
half-spherical section of the cell finger. The pore density is
almost constant in the case of long cell fingers, while for shorter
cell fingers the electropore density decreases with increasing
azimuthal angle, and the decrease becomes steeper with
decreasing finger length.
4.3. Electric Field and Potential. The transmembrane voltage
and consequently the electric field strength is highest at the
bottom of the cell finger. In the case of 25 V applied to the
capacitor plates of the chamber, at the bottom of the finger the
through-membrane electric field strength changes from 4.5 to
7 105 V/cm, while the finger length increases from 2 to 10
m. The more embedded the cell is into the filter pore, the lower
the potential within the cell becomes. The potential is constant
within the protruding section(s) of the cell, while it slightly
changes within the finger of the cell (Figure 8). Thus the field
strength is negligible in the protruding section(s), and it is less
than 460ez V/cm in the finger. Because of the CDA, the strongest
current density of the LVEP can be found in the narrow passage
of the filter pore, and similarly in the extracellular space, the
field strength is strongest in the narrow passage because the
current density is directly proportional to the field strength
(Ohms law). The field strength in the narrow passage close to
the membrane surface can be estimated by means of the steepest
slope of the transmembrane voltage curves in Figure 6b,c
(Appendix 2). When the cell is fully embedded in the filter pore,
the field strength in the narrow passage at the membrane surface
is 7230ez V/cm. In the case of partially embedded cells, the
electric field in the narrow passage at the membrane surface
increases with decreasing finger length from 7230ez to 23100ez
V/cm. When the geometry of the LVEP unit approaches the
real geometry of LVEP, that is, the thickness of the membrane
and narrow passage are simultaneously decreased, the relative
increase of the electric field strength is significant only inside
the cell finger (see Appendix 2). The current density is
proportional to the field strength (Ohms law), and the Joule
heating is proportional to the square of the current density. Thus,
the Joule heating during the pulse is highest at the narrow
passage and negligible in the bulk regions. According to the
calculations16 after three pulses each of amplitude 10 V and

Sugar et al.

Figure 9. Minimal applied voltage of electroporation. Solid line

represents the calculated minimal applied voltage vs cell position. zmin
values between the vertical dash-dotted lines refer to positions of the
cell embedded partially into the filter pore. Open circle and open square
marks minimal applied voltage calculated at 0.5 and 1 V critical voltage
of electroporation, respectively. The minimal applied voltage observed
in LVEP for human erythrocyte is marked by X.

duration 30 ms, the Joule heat accumulated in the narrow

passage dissipates quickly, within 0.3 ms, toward the bulk
regions without causing permanent cell damage. Our experimental results show that after applying the above characterized
three pulses only about 2% of the cells die. For comparison,
we mention that in SEP during electroporation the Joule heating
takes place everywhere in the extracellular space and thus the
heat dissipation after the pulse is very slow. The heat transfer
through the slowly resealing electropores warms the intracellular
space causing eventually permanent cell damage. This explains
that in LVEP cells survive even 20 kV/cm local electric field
strength, while in SEP 3000 V/cm with pulse duration in the
millisecond range is the upper limit of cell survival.7
4.4. Minimal Applied Voltage. Because the transmembrane
voltage is directly proportional to the applied voltage, one can
calculate (from the data in Figure 6) the applied voltage needed
to get 0.5-1 V transmembrane voltage at the bottom of the
cell. This is the minimal applied voltage needed to electroporate
the cell at least at the bottom of the cell. Figure 9 shows the
minimal voltage applied on the capacitor plates of the LVEP
chamber as a function of the cell position, zmin. Open circle
and open square marks minimal applied voltage calculated at
0.5 and 1 V critical voltage, respectively. The observed minimal
applied voltage for human erythrocyte, marked by X in Figure
9, is within the range of the calculated values. In Figure 9, cell
positions between the two vertical dash-dotted lines refer to
cells partially embedded in the filter pore. The calculated
minimal applied voltage increases steeply with increasing cellto-filter distance. This theoretical result is similar to the
experimental data of Yang et al.9 They reported a procedure
for in situ electroporation of cells grown on microporous
membranes of polyethylene terephthalate or polyester (but not
pushed into the filter pores) and induced electroporation from
as low as 70 V applied voltage. It is important to note that if
there is no filter in our LVEP the minimal applied voltage is
4030 V!
4.5. Efficiency of Electroporation. Finally, we point out that
the efficiency of electroporation is much higher for fully and
partially embedded cells than for cells out of the filter. The
efficiency of the electroporation can be characterized by the
proportion of the surface of the cell membrane where the

Theory of Low-Voltage Electroporation

J. Phys. Chem. B, Vol. 107, No. 16, 2003 3869

Figure 10. The efficiency of electroporation. The efficiency of the

electroporation (i.e., the proportion of the membrane surface area where
the critical transmembrane voltage, Vcr ) 1 V, is exceeded) is plotted
against zmin. The curves belong to the following voltages applied to
the capacitor plates of the LVEP chamber: 10 V (dash-dotted line);
12.5 V (dotted line); 25 V (solid line); 75 V (dashed line). zmin values
between the vertical dash-dotted lines refer to positions of the cell
embedded partially into the filter pore. The total surface area of the
erythrocyte cell is 137.3 m2.

transmembrane voltage exceeds the critical value, Vcr. By using

Figure 6a-c, one can get the z coordinates of the membrane
segments (for a given cell position), where the transmembrane
voltage is above the critical voltage. Then one can calculate
the surface area of the cell membrane belonging to these z
coordinates. The efficiency of the electroporation is this area
to the total surface area of the cell. In Figure 10, the calculated
efficiency of the electroporation is plotted against the cell
position. The curves are calculated at different applied voltages.
The efficiency of the electroporation is high, 70-98%, when
the transmembrane voltage is above the critical value at both
the bottom and the top of the cell. With decreasing finger length
(i.e., with increasing zmin), the transmembrane voltage decreases
at the top of the cell, and when it becomes less than the critical
voltage, the efficiency of the electroporation drops considerably.
Then the efficiency decreases linearly with decreasing finger
length until zero efficiency is attained. At a given cell position,
higher applied voltage produces the higher efficiency, and at
higher applied voltage, the drop of the efficiency takes place at
shorter finger length.
5. Conclusions
In a LVEP, cells are embedded in the pores of a micropore
filter. The narrow conductive passages in the filter pores result
in a highly inhomogeneous electric field in the electroporator.
At as low as 2 V of applied voltage, the field strength becomes
1000-4000 V/cm in each micropore and the transmembrane
voltage exceeds the critical voltage of cell electroporation at
the tip of the finger, that is, at the bottom of the cell penetrating
into the filter pore. The LVEP is ideal for cell transfection with
foreign genes. The Joule heat accumulated mainly in the filter
pores quickly dissipates toward the bulk solutions of the LVEP
chamber before the interior of the embedded cells would warm.
Thus the cell survival rate is very high, about 98%. At 25 V
applied to the capacitor plates of the LVEP chamber, the
transmembrane voltage is higher than the critical value at 8790% of the cell surface if the cell penetrates further than half
length of the filter pore. Because a large percentage of the cell
surface can be electroporated, the observed transfection efficiency for the embedded cells is higher than 90%.
Acknowledgment. The authors thank Professor Herman
Schwan for his helpful criticism and valuable comments. Dr.

Figure 11. Calculated transmembrane voltage and potential along the

membrane at different membrane and narrow passage thicknesses.
Narrow passage thicknesses are 0.125 m (dotted line), 0.1 m (solid
line), 0.075 m (dashed line), and 0.05 m (dash-dotted line). At every
calculation, the membrane thickness is taken to be equal with the
thickness of the narrow passage. zmin ) 5.1 m. In panel A, the
transmembrane voltage is plotted against the z coordinate of the
membrane segment. In panel B, the potential at the inner membrane
surface is plotted against the z coordinate of the membrane segment.
The voltage applied to the capacitor plates of the LVEP chamber is
Vapp ) 25 V.

Sugar acknowledges Mrs. Gardners generous support. This

work was supported by Pfizer Inc. and by a NIH Grant (Grant
2 R44 HGO1589-02A1).
Appendix 1
Estimation of the Thickness of the Narrow Passage. The
total measured resistance of the electrically parallel narrow
passages of the filter pores, that is, the leak resistance, RL, is
200 and the number of micropores in the filter of radius 0.5
cm is NP ) 3.3 105.17 Thus the average resistance of one
narrow passage is RP ) RLNP ) 6.6 107 . The average
cross-sectional area of a narrow passage is AP ) FfLP/RP )
0.0753 m2, where the resistivity of the physiologic solution
(0.15 M NaCl) is Ff ) 7.1 105 m and the average length
of a narrow passage is LP ) 7 m. The thickness of the narrow
passage is expected to be tP ) ro - ri ) ro - xro2-AP/ )
0.012 m, where ro (1 m) and ri are the outer and inner radii
of the narrow passage, respectively.
Appendix 2
On the Deviations of the Models Geometry from the
Electroporators Geometry. In our model, both the membrane
and the narrow passage thickness are 10 times larger than the
observed values. To investigate the effects of these geometrical
parameters on the calculated transmembrane voltages, we
simultaneously decreased the thickness of the membrane and
the narrow passage first by 25% and then by 50%. The obtained

3870 J. Phys. Chem. B, Vol. 107, No. 16, 2003

Sugar et al.

TABLE 2: Electric Field in the Narrow Passage and in the

Cell Finger
membrane
thickness (m)

narrow passage
thickness (m)

Ef (V/cm)

Ep (V/cm)

0.125
0.1
0.075
0.05

0.125
0.1
0.075
0.05

220.4
278.3
361.1
531.9

11 584
11 641
11 725
11 896

transmembrane voltage curves, in Figure 11a, do not show

significant deviations from the result obtained in the case of
the original model geometry (see solid line in Figure 11a). This
is the case because the simultaneous decrease of these two
geometrical parameters similarly increases the electric field
strength on both sides of the membrane. On one hand, upon
narrowing the passage, the current density and the field strength
increase in the passage. On the other hand, upon decreasing
the membrane thickness, the membrane resistivity decreases and
more current flows into the cell finger, that is, the field strength
increases in the finger. In Table 2, the calculated field strengths
are listed at different thicknesses of the narrow passage and
cell membrane. The field strength inside the finger, Ef, is
calculated from the steepest slope of the inner potential curve
in Figure 11b. The field strength in the narrow passage, Ep, is
calculated from the following relationship: Ep ) -dV/dz + Ef,
where dV/dz is the steepest slope of the transmembrane voltage
curve in Figure 11a. Note that in the narrow passage the electric
field strength increases only by 2% when the thickness of the
narrow passage and cell membrane are simultaneously reduced
by 50%.
References and Notes
(1) Stampfli, R. An. Acad. Bras. Cienc. 1958, 30, 57-63.
(2) Sale, A. J. H.; Hamilton, W. A. Biochim. Biophys. Acta 1968, 163,
37-43.

(3) Neumann, E.; Rosenheck, K. J. Membr. Biol. 1972, 10, 279-290.

(4) Zimmermann, U.; Schulz, J.; Pilwat, G. Biophys. J. 1973, 13, 10051013.
(5) Neumann, E.; Schaefer-Ridder, M.; Wang, Y.; Hofschneider, P.
H. EMBO J. 1982, 1, 841-845.
(6) Neumann, E.; Sowers, A. E.; Jordan, C. A. In Electroporation and
Electrofusion in Cell Biology; Neumann, E., Sowers, A. E., Jordan, C. A.,
Eds.; Plenum Press: New York and London, 1989.
(7) Friedrich, U.; Stachowicz, N.; Simm, A.; Fuhr, G.; Lucas, K.;
Zimmermann, U. Bioelectrochem. Bioenerg. 1998, 47, 103-111.
(8) Eidsath, A.; Button, D.; Schmukler, R. E. Unpublished work, 1995.
(9) Yang, T. A.; Heiser, W. C.; Sedivy, J. M. Nucleic Acids Res. 1995,
15, 2803-2810.
(10) Schmukler, R. E. Methods of Electroporation and Electrofusion.
U.S. Patent No. 5,173,158, 1992.
(11) Schmukler, R. E. Apparatus for Electroporation and Electrofusion.
U.S. Patent No. 5,283,194, 1994.
(12) Schmukler, R. E. Mater. Res. Soc. Proc. 1996, 411, 45-56.
(13) Huang, Y.; Rubinsky, B. Biomed. MicrodeVices 1999, 2, 145150.
(14) Schmukler, R. E. Proc. XI Int. Conf. Electr. Bio-Impedance 2001,
233-236.
(15) Sugar, I. P.; Schmukler, R. E. Proc. XI Int. Conf. Electr.
Bio-Impedance 2001, 149-152.
(16) Schmukler, R. E.; Eidsath, A.; Button, D.; Lindesay, J.; Khizder,
K.; Ahmed, A.; Sugar, I. P. Unpublished work, 2003.
(17) Schmukler, R. E. A new technique for measurement of isolated
cell impedance. Eng. Sc. D. Thesis, Columbia University, New York, 1981.
(18) Cole, K. S. Membranes, Ions and Impulses; University of California
Press: Berkeley, CA, 1972.
(19) Schwan, H. P. In Electroporation and Electrofusion in Cell Biology;
Neumann, E., Sowers, A. E., Jordan, C. A., Eds.; Plenum Press: New York
and London, 1989; pp 3-22.
(20) Kinosita, K., Jr.; Ashikawa, I.; Saita, N.; Yoshimura, H.; Itoh, H.;
Nagayama, K.; Ikegami, A. Biophys. J. 1988, 53, 1015-1019.
(21) Hibino, M.; Shigemori, M.; Itoh, H.; Nagayama, K.; Kinosita, K.,
Jr. Biophys. J. 1991, 59, 209-220.
(22) Hibino, M.; Itoh, H.; Kinosita, K., Jr. Biophys. J. 1993, 64, 17891800.
(23) Bordi, F.; Cametti, C.; Misasi, R.; De Persio, R. Eur. Biophys. J.
1997, 26, 215-225.
(24) Takashima, S.; Schwan, H. P. J. Membr. Biol. 1974, 17, 51-68.