IEEE nansactions on Electrical Insulation

Vol. 23 No. 6, December 1988

897

Continuum Properties from

Inter digit a1 Electrode Dielect romet r y

M a r k C. Zaretsky, Lama Mouayad

James R. Melcher,

and

Lab. E&E Systems M.I.T. Cambridge, MA

ABSTRACT
Using a modal approach, a model is derived that makes the interdigital electrode microdielectrometer developed by Senturia
and co-workers applicable to measuring continuum parameters
in a wide range of heterogeneous media. In this imposed w - IC
technique, the medium is excited at the temporal (angular) frequency w by means of an interdigital electrode structure having a spatial periodicity length X = 27r/IC and hence a dominant
wavenumber IC. Given the surface capacitance density C(w,IC)
of any linear system having property gradients perpendicular
to the plane of the electrodes, the model predicts the complex gain, taking into account the properties, geometry, and
terminal configuration of the interdigital electrode structure.
This capability can then be used with an appropriate parameter estimation strategy to determine the continuum properties
and/or geometry of the medium. Specifically illustrated, using a secant method root searching routine for the parameter
estimation, are estimations of film thickness, film permittivhickness known, and film surface conductivity with

INTRODUCTION

IMPOSED w

IC SENSING

dielectrometry illustrates how important

information concerning the physical properties and
LASSICAL

state of a medium can be obtained from the electrical

frequency response. The modal approach to dielectrometry described here not only extracts its information
from the control of the temporal frequency, but also from
the imposition of a spatial periodicity as well. In this

0018-9367/88/1200-897$1.00 @ 1988 IEEE

Zaretsky et al.: Continuum Properties from Interdigital Electrode Dielectrometry

898

imposed w - k technique, the medium is excited a t the

temporal (angular) frequency w by means of an interdigital electrode structure having a spatial periodicity
length X = 2 a / k and hence a dominant wavenumber k .

+
Y

L--q

1
t
I

Figure 1.
Representation of interdigital electrode structure and associated circuitry interacting with a
heterogeneous medium.
In Figure 1, the medium is above and the structure used to make the measurements is below. Because
the electrodes introduce higher spatial harmonics, field
quantities, such as the potential, are expressed as

where
00

b,(z)cos(k,y)

&(z,y)=

k, = 2 n r / X

n=O

representing a superposition of standing waves of dominant wavenumber k = k l in the plane of the electrodes.

The choice of the cosinusoidal representation exploits

the symmetry of the electrode structure and the presumed symmetry of the situation above (for example,
there is no motion parallel to the electrode plane).
The advantages of the w - k approach are many.
First, there is the coupling into the medium from a single surface, allowing measurements to be more readily
made noninvasive, if desired. For example, dielectric
measurements of thin films can be performed without
having to vapor-deposit a n upper metal electrode for
use with a capacitance bridge, and in many cases, can be
taken at the site of the coating process without changing
the ambient environment.
Secondly, because the fields generated by the electrodes in the medium (above in Figure 1) are quasistatic,
they tend to decay into the material exponentially with
a characteristic length that is at most a fraction of the
spatial wavelength. Consequently, the electrodes are
sensitive to dielectric materials located within a region
roughly 5 X/3 from the electrode plane. Thus, the spatial sensitivity of the device can be tailored to individual
needs.
A third advantage comes with modern microfabrication techniques that not only make micrometer-scale
electrodes possible, but make the integration with the
signal processing electronics feasible as well. Electrodes
with a very short spatial wavelength X can be deposited.
This refines the spatial resolution. For example, given
a spatial wavelength of 50 pm, thin films of 5 20 pm
(perhaps coated on the electrodes) can be distinguished
and their dielectric dispersions examined.
Having short wavelengths also improves the sensitivity t o measurements of surface conductivity Q,. Sensing conduction on the surface of an insulator without
making electrical contact with the surface requires capacitive coupling. Thus, the frequency must be high
enough to make the capacitive reactance on the order of
the resistance of the surface. Roughly, this requires that
to measure a surface conductivity U , a t the frequency
f E w / 2 a , the ratio ( u , / f ~ Xmust
) be of the order of
unity. There is a practical lower limit on the frequency.
Therefore, the smaller the wavelength A, the smaller
the surface conductivity that can be measured. Having
the electronics integrated with the electrode structure
makes it possible to use frequencies as low as 0.005 Hz
using the chip described in the next Section.
Although the emphasis here will be on dielectrometry, the imposed w-k approach can be used in systems

IEEE Transactions on Electrical Insulation

Vol. 23 No. 6, December 1088

where the electrical response reflects electromechanical

and electrochemical phenomena. This is illustrated by
earlier studies where interdigital electrodes were used
to study the electromechanics of double layers [l],electrohydrodynamic surface waves and instabilities [2] and
sensing of electrification effects in insulating fluids [3,4]

INTI MAT E S ENSI N G

contrast with remote sensing, where electromagnetic fields described by the wave equation are often used to resolve properties and states of materials
a t great distances, the imposed w - k approach exploits
fields that are quasistatic from the electromagnetic point
of view. In the electroquasistatic embodiment described
here, the dynamics result from charge migration and
diffusion, and polarization relaxation. In the magnet@
quasistatic analogue, the dynamics would involve current diffusion and magnetization hysteresis. In either
of these latter cases, the fields are quasistatic and tend
to be confined to the immediate neighborhood of the
interdigital electrode structure. Because measurements
reflect the properties and state within a short distance,
the application of such structures t o parameter estimation might be dubbed intimate sensing. Strategies exploiting the wavelength controlled resolution inherent to
having fields that are confined to the proximity of the
electrode can utilize the model developed here.

With the combination of microfabrication technology and the imposed w - k technique, a determination of electromechanical or electrochemical structures
and dynamics on a microscale becomes feasible. Fluid
boundary layers, electrochemical double layers and thin
film structures all fall in this microscale realm.
Intuitively, it makes more sense to determine the
structure of heterogeneous media by probing spatially,
rather than temporally. With the usage of the approach
presented here, a continuum model coupled with a parameter estimation routine, it becomes feasible to extract information a t a particular temporal frequency,
varying the spatial frequency of the applied potential.
This method does not have the restriction regarding the
frequency dependence of parameters such as the complex permittivity. Practically, this effort amounts to
having multiple electrode structures, each having a different spatial wavelength, monitoring the same medium,
or perhaps one array of electrodes with switchable terminal connections to produce the various spatial wavelengths.

899

MICRODIELECTROMETRY
ICRODIELECTROMETRY , as developed by Senturia
and co-workers [5,6], is a commercially available
technique [8] for measuring complex permittivity utilizing microfabrication technology to incorporate both the
sensing electrodes and associated circuitry on the same
microchip (Figure 1). A set of interdigital, planar electrodes are deposited on an insulating oxide layer along
with two FET transistors [7]. One set of electrodes is
driven by a variable frequency (0.005 to 10000 Hz) ac
voltage. The other set of electrodes is allowed t o float
by connecting it to the gate of one of the transistors.
The other transistor serves as a reference. Using the
feedback circuit shown in Figure 1, in conjunction with
the transistors, a very high impedance measurement of
the floating gate voltage can be obtained, even at very
low frequencies. The output of the device, defined as
the complex gain G, is th,e complex ratio of the floating
to driven gate voltage, (VFIVD),
and is obtained using
a correlation analysis of the two voltages.

Two specialized models have been developed for interpreting data obtained with the microdielectrometer.
One model is for its application to the monitoring of
epoxy resin cures [9,10]. This model, incorporated in
the form of a lookup table supplied with the microdielectrometer system [7], makes it possible to relate the
measured complex gain to the complex permittivity of
a uniform infinite half space. For this case, a finite difference simulation of Laplaces equation was used to determine the electric field distribution, treating the region above the electrode structure as a semi-infinite,
isotropic medium [lo]. The second model previously
developed is for the study of a very thin film (< 1 pm)
[11,12]. The thin film-oxide layer was represented as
an RC transmission line shunted by lumped capacitors.
This made it possible to represent the gate voltage as
the transmission line response t o a driving voltage.

The work reported here is an outgrowth of work being done to develop the microchip for monitoring transformer insulation [13]. There, the intent is to coat the
microchips either to: (1) passivate the chip to the adsorption of moisture on the silicon dioxide interface so
that they can be used to measure the dispersion of the
oil and oil-impregnated system or (2) to utilize their
thin region sensitivity t o monitor changes in dielectric
properties of finite thickness coatings, distinguishing between bulk dispersion and effects of heterogeneity.

Zaretsky et al.: Continuum Properties from Interdigital Electrode Dielectrometry

900

OBJECTIVES

HE first objective is to develop a general and adapt-

/ kn

able model for determining the output of the interdigital electrode dielectrometer. The second is to establish a rapport for the relationship between predicted
frequency responses and several physical situations of
immediate interest. Finally, a method for parameter
estimation utilizing the continuum model and demonstrating its practical application, will be outlined and
demonstrated.

-,G

CONTINUUM MODEL
array
spec i f icat ion

APPROACH

N Figure 2 is shown a schematic view summarizing the

structure of the continuum model developed in the

next Section. One set of inputs to the model, the lower
box on the left in the diagram, describes the electrode
array structure. Parameters such as interelectrode spacing a and the spatial wavelength X are specified here.
In addition, for the specific electrode structure considered here, interdigital electrodes deposited on an oxide
layer with ground plane underneath, the electric field
distribution below the electrodes is completely specified
by the insulating oxide layer thickness h and permittivity coz. The other input, the upper box on the left,
describes the half-space of media above the electrodes.
This description is in the form of a complex surface capacitance density C,, representing the response of the
half space to one Fourier component of a potential applied at the electrodes. As shown in the upper part of
the figure, this responseis the complex ratio of the normal displacement field D; measured at the interface to
the interfacial potential @E. The signifies a complex
quantity, the superscript indicates that the function is to
be evaluated just above the interface and the subscript
indicates the nth Fourier component. All the heterogeneity and structure of the medium is incapsulated in
C, .
A

The
box represents the step of solving the electrostatic field problem. Field quantities in the oxide
layer and in the medium are represented in terms of their
Fourier components. Boundary conditions at the electrode interface are used to match up the field solutions
for the regions above and below the electrodes. A mixed
boundary value problem occurs a t this interface, as the
potential is constrained along that part occupied by the
electrodes while conservation of charge and Gauss law

-w

Figure 2.
Schematic representation of continuum model.

combine to put a constraint on the potential and on the

jump in its normal gradient along the interelectrode part
of the surface. The solution involves discretizing the interelectrode surface by introducing a grid of unknown
voltages 5 a t k collocation points. The potential is assumed to vary linearly between these collocation points.
A system of equations is generated by maintaining a discretized conservation of charge along surface segments
bracketing these collocation points. These equations are
written in the form A.V=X where the elements of A are
the coefficients of the unknown voltages VJ and the elements of X represent the known voltage
Inverting
the matrix A yields the voltage distribution along the
interelectrode surface, and thus, the electric field distribution.

c~.

In the final box, the evaluation of the yijs is tantamount to a determination of the response with any
terminal configuration. Due to the symmetry of the
electrode structure, Y11 = E>z. All of the admittances
representing the electrode structure are determined by
,
finding the electrical terminal currents, iD and i ~ with
the floating gate electrode grounded. These currents
are obtained by integrating the current density over the

I E E E Transactions on Electrical Insulation

Vol. 23 No. 6, December 1988

901

surfaces of the electrodes. Using these admittances, the

response of the microdielectrometer with a n arbitrary
load is the complex gain, the ratio of the floating to
driven gate voltage

I
where Yl is the load capacitance of the floating gate
FET.

"D

In the medium above the electrodes and in the oxide substrate below, the potential is represented by (1)
and the fields are taken as electroquasistatic, the electric field intensity is irrotational and hence represented
by the potential Cp.

E=

(3)

-Vip

For completeness, the oxide layer will be given a complex

permittivity, E; ..

ELECTRODE STRUCTURE BOUNDARY

CONDITIONS

Figure

amount of free surface charge density o,, (6 is the unit

normal t o the interface)

shown in Figure 3. The electrodes are treated as

being one dimensional, one set driven sinusoidally with a
peak voltage VD and frequency w , the other set grounded.
It is possible to include the finite thickness of the electrodes using the techniques outlined here, but the effort
is not judged to be worthwhile a t present. For generality, an interelectrode spacing of arbitrary width a is
allowed. Thus, in a half period, the endpoint of one
electrode is a t y = yo and the beginning of the other
electrode is a t y = yo a E yk+1.

At the electrodes, the potential is constrained, whereas

along the surface between the electrodes there is only
continuity of the potential

~Deiut}

for

o Iy 5 yo

for Yk+l I
Y I

3.

Coordinate system used for electrode structure

of Figure 1.

HE electrode structure of Figure 1 is modeled as

Re

...

6 .ID!

On the interelectrode surfaces, Gauss' Law relates

the discontinuity in the normal displacement field t o the

(5)

where 0 0 implies the difference in the function within

the brackets evaluated just above and just below the interface and Vc. is the surface_divergence. The surface
electric displacement vector D, represents the possibility of polarization phenomena concentrated within the
interface. Conservation of charge accounts for the accumulation of
via the discontinuity in normal conduction current J , including conductivity within the_interface represented by the surface current density J,,

S . p - ~ + V c . ~ ao,U
+at

=o

(6)

Substituting (5) into ( 6 ) , using the complex notation introduced by (1) and assuming sinusoidal steady

(4)

+ V c . D, =

state (-

at

-+

iw) yields the following boundary condition

where complex bulk and surface permittivities have been

introduced, E * = E' - ZE" and E: = E: - Z E ; , and the conduction and displacement volume and surface current

Zaretsky et al.: Continuum Properties from Interdigital Electrode Dielectrometry

902

densities have been combined

l? in the interfacial plane),
A

(2,is the component of

A

iw~',?? = J+ wl? ; w e : E , = J,

+iw5s;

(8)

For an ohmic conductor E' and E: are constant, and

E: = ( a , / w ) , where a and 0, are constant.
6''

= ( a / w ) and

For uniform surface properties (7) yields the second

interelectrode region jump condition,

(4c), and conservation of charge, (6)' is satisfied. The

strategy taken here is to represent the interface potential
by piece-wise continuous linear segments having the collocation potential V, a t each collocation end-point yj.
Illustrated in Figure 4 are four collocation points, introduced in the interelectrode region between yo and yk+l.
An initial effort divided the region into equal length line
segments [13]. A cosinusoidal distribution has been utilized here to give finer resolution in the regions of higher
electric stress (adjacent to the electrodes). Thus, the
collocation points are a t

+ -)a2

Y j = (YO

BULK RELATIONS
s discussed earlier, the field distribution in the halfspace of material either above or below the electrodes can be represented by the appropriate surface
capacitance density. This quantity is the ratio of the
the complex Fourier amplitudes of normal electric displacement to potential, both evaluated on the ( a ) or ( b )
side of the interface.

&"(y) =

for 0

Vj(Y-Yj-l)-Vj-~(Y-Yj~
( Y j - Y j - 1)

for yj-l

&,h

&2b

42b

Here, the subscript n refers to the order of the Fourier

modes. Given the distribution of a complex permittivity E:(z) above the electrodes, with the coordinate 2:
perpendicular to the interface (including surface singularities in that d*istribution), the complex surface capacitance density CE is found by determining the Fourier
amplitude of the normal flux density at the surface (a)
caused by the nth term of the imposed potential having
the form of (1).

POTENTIALS AT COLLOCATION POINTS

The Fourier amplitudes of (1) are

The potential &" is known only a t the electrodes.

On the surface, in the interelectrode region, the potential assumes a distribution such that it is continuous,

I yj
j=1,2, . . . ,k+l

for Yk+l I Y

'i

(13)
Substituting for &"(y) into (11) yields the Fourier
series coefficients
.

I y L Yo

VD

---t
Dzh
n.
C;,b
(10)

(12)

+ 1, and the potential distribution is

for 0 5 j 5 k

k+l

- (;)cos(")

a;=--

XVD

(nnI2 (Y1 - Yo)

[ C O S ( ~ n Y O )-

COS(knY1)l

1
j=1

cos (knY j ) -

C O S ( ~ ~ Y-~ cos(knyj-1)
+I)
Yj+1 - Y j
Y j - Yj-1

The problem has been transformed t o one of solving

for k unknowns, the $ '9, with the continuity of potential
satisfied by (4c). The second continuity condition at the
interface, (9), remains to be satisfied. By dividing the
interelectrode surface into k intervals, centered on the
collocation points and satisfying conservation of-charge
(9), within every interval, k equations for the k VJ's are
obtained. The intervals are demarcated by points y; located a t the edges of the electrodes and at the midpoints
of the line segments formed by consecutive collocation
points, yJ and yj+l (Figure 4).

Yj* =

YO

for j = 1

i(Yj +Yj-l)

for 2 5 j 5 k
for j = k + I

Yk+l

(15)

IEEE Transactions on Electrical Insulation

Vol. 23 No. 6, December 1988

903

90

-[

k = 4

"I
A

"2
Figure 5.
Distribution of tangential electric field, h;(y),
along electrode-medium interface.

"3
A

v4

respective segments, extend t o the electrode as shown

in Figure 5.

Y
a
Figure 4.
Voltage distribution along discretized interelectrode surface for k=4 collocation points. Unknown voltages Vj are introduced at the points
y j . Conservation of charge is maintained for line
segments defined by the y.;

Integrating (9) over each interval] from y j to y ~ + l l

and setting the result t o zero yields k equations for k
unknowns,

With this choice of y;, the ambiguity in the value

of the tangential electric field a t the collocation points
yj is avoided. Using (3) and (13) yields

for j = 1 , 2 , . . . , I C . T h e third term, reflecting contributions due t o surface effects, integrates t o

(18)
E"(y*)
=Y
J

Y j - Yj-1

(16)

As shown in Figure 5, E;(y) is a piecewise function

that is discontinuous a t the collocation points. T h e intervals over which charge is conserved (between broken
lines in Figure 5) extend half a segment in either direction about a collocation point y j except for the first and
last intervals, which respectively have end points a t the
electrode edges. In addition, the values of the tangential
are not averfield used for the endpoints y; and
aged even though they are located a t a discontinuity. It
is more realistic t o assume that these values, for their

and using (16) yields

E:,

[(

1
Yj+1 - Y j

+ Yj -

v,+l

Yj+l

Yj

Yj-1

).;

Yj - Yj-1

(19)

Integrating t h e remaining two terms of (17), using

the cosinusoidal dependence of the normal electric field
upon o implied by (1) and (3), yields the set of equations

Zaretsky et al.: Continuum Properties from Interdigital Electrode Dielectrometry

904

" 1
-(E:&
kn
n=l

+E:,

~ik:)[sin(k,yj+,)

[ (-

sin(knyj)]

Yj+1 - Y j

Yj - Yj-1

%t1

5-1

Yj+1 -

Yj

Y j - Yj-1

] =o
~ 3 ( ij ,)

for j = 1 , 2 , . . . , k.

[ (-

+ -)

1
1
6(i-j )
Yi+l - Yi
Yi - yi-1
-6 ( i - j
1) - q - j - 1)
Yi+l - Yi
Yi - yi-1

= E,:

and
Again substituting for the electric fields, this time
using (10) (remembering continuity of the potential across
where
an interface) into (20) yields

X ( i ) = Xl(i)

X,(i) = -

"
n=l

+ XZ(2) +

A
(C:

kn(nT)'

[sin(kny;+,

(23)
(24)

X3(2)

e:)

sin(kn~i')]

Yj+l - Y j

+-+Y j - Y j - 1

G+l
Yjtl

-Yj

X,(i) = X3(i) = &;o-

Yj

Yj-1

for j = 1,2, .. ., k.

[e: e;
-

(591

(y;+l

yi')

6 ( i - 1)
Y 1 -Yo

where yj and yj* have been previously defined and 6 ( i )

is defined as unity if the argument is zero and zero for
nonzero argument. Using this representation, the set of
k equations in k unknowns can be numerically inverted
t o determine the $,;'S. A more rapidly convergent series
for (23a) and (25a) is derived in Appendix A.

The set of equations described by (21) can be expressed in matrix form, A.V=X, with the elements of A
representing the coefficients of the unknown normalized
voltages pj E (V,/Vo), and the elements of X representing the coefficient of the known drive. In order to
generate the matrix elements, (14) is substituted into
(21) and terms are rearranged such that all terms proportional to the drive are moved to the right hand side.
Any expression involving Cz also goes on the right hand
side as it can only arise from an externally applied field
and thus, is also a driven signal. This yields

the transition from the field analysis just undertaken to the general situation where the detection electrodes are not grounded but rather are free to assume
a potential consistent with th? attached circuit. With
the driven gate a t potential V, and the floating gate
grounded, Y11 and Y1z are directly related t o the electrode currents, and hence calculable from the electric
fields just determined with the floating gate grounded.

where

The electrode currents are found by integrating (6),

conservation of charge, throughout a volume enclosing

EQUIVALENT CIRCUIT PARAMETERS

HE pi network of Figure 2 is helpful for formalizing

I E E E !lkansactions on Electrical Insulation

Vol. 23 No, 6, December 1988

each electrode. Assuming a total electrode meander

length of Mcl, these currents are

COLLOCATION POINTS, FOURIER MODES

AND CONVERGENCE

+ 2JAiZ
Yk+l

[E:E,n(y

Solving- .(26) for Y11 and Y1z in terms of ;D and ;F

and incorporating ( l ) , (3), (10) and (16) yields
I

Yll = 2iwM,l

{E:,

(-

1 - 3, -

Y 1 - YO

Yk+l

Yk

(sin(knyo) - Sin(knyk+l))

+&:
Ylz = 2iwM,1

[e: - e:] + x/2 - Yk+l)}

(YO

{E:,,

U
Yk+l - Yk

-4:

-&:

[e: - e;] ( x / 2 - yk+l)}

[Cz -

where Uj
( $ / V D ) . Once (28) is programmed, it is
only the surface capacitance densities, Ct, that are altered when the medium above is changed.

DETERMINATION OF GAIN
this point it is instructive to recap the major
steps for calculating the xjsand hence, the gain.
Solution of a system of equations of the form A-U=X
with the matrix elements defined by (22)-(25) determines the Ujs and the electric field distribution. With
these C j s , the Fourier coefficients for the potential a t
the interface, &E and i ; t , using (14),and hence, the circuit parameters Y11 and Y12, using (28), are evaluated.
Finally, given a load capacitance
the gain is then
calculated according to (2).

EVERAL tradeoffs must be considered for deciding

the optimum number of collocation points and Fourier modes. Increasing the number of collocation points
improves the representation of the actual potential distribution - a t the cost of greatly increasing the computing time required to calculate each element of the matrix
A and t o compute the matrix inverse. Increasing the
number of Fourier modes used to compute the voltage
distribution increases the accuracy - of representing this
distribution with a piecewise linear function. Beyond a
certain number of modes the piecewise representation
may differ significantly from the actual one. However,
as previously mentioned, and elaborated in Appendix
A, a more rapidly convergent series is used to sum the
Fourier modes for the elements of the A matrix. This
series effectively sums up roughly 1000 modes. In this
case, it is quicker to compute either a small or a large
number of modes than to compute an intermediate number. Using the large number of modes, the number of
Fourier modes summed in (28) when computing Y11 and
Y1z can be altered. Increasing the number of terms used
here has a much smaller, though still significant, effect
on the computing time.

e:] sin(k,yk+l)

kn

905

x,

Determination of convergence requires a specification against which various results can be compared. For
the case of the microdielectrometer, where comparisons
are made with experimental results, the accuracy of the
gain is given as 0.1 dB [7]. This accuracy will be used
as a tolerance for measuring convergence. Modeling an
uncoated microchip in air, the high-frequency coupling
capacitances per unit length (normalized t o the oxide
layer permittivity) and resulting gains were computed,
using the large number of Fourier modes, for a matrix of
values for the number of collocation points and summation terms (used in (28)). It was observed that using 25
collocation points and 100 summation terms provided
sufficient convergence when compared to the maximum
values of 35 and 1000, respectively (as well as taking a
reasonable amount of computing time). For the rest of
this work, unless otherwise noted, 25 collocation points
and 100 summation terms will be standard. The specific values used for the oxide layer thickness h and load
admittance
were supplied by Micromet Instruments,
Inc., on a proprietary basis and were used t o generate
the specific gain-phase responses in this paper.

Zaretsky et al.: Continuum Properties from Interdigital Electrode Dielectrometry

906

TYPICAL RESPONSES USING

CONTINUUM MODEL

UNIFORM MEDIA
This situation, a semi-infinite half space of uniform
complex permittivity, is pictured in Figure 6a. The potential satisfies Laplace's equation and decays to zero as
x goes to infinity. Thus,

PREDICTED RESPONSES

As

a first step toward developing a scheme for identifying continuum parameters, this Section is devoted to forming insights concerning the general relationships between these parameters and the gaip-phase
response. *For all of the following examples E," = 0,
and thus C: = 0. Using the form expressed in (l),the
solution of Laplace's equation for the potential in the
insulating oxide layer below the electrodes gives

(29)

(These expressions, (29) and (30), illustrate how transfer

relations naturally fit into the spectral field description
~41).

-10

A' i

U,

L\

where the layer has a thickness h and is bounded from

below by a 'perfectly conducting' substrate.

-3

-I

log f

Figure 7.
Predicted responses for microchip (X=50 micrometers and a = X / 4 ) in uniform medium
( B I = 2 x lo-"
F/m), vary conductivity ul.

Figure

6.

Electrode structure below a ) uniform medium

or b) variable thickness homogeneous layer and
uniform medium.

For ohmic media, typical responses have the features of a linear system with one time constant - the
charge relaxation time, r, = ( E ~ / c T ~(Figure
)
7). With
the frequency high, wr, >> 1, the effect of dissipation
currents is small. The coupling between the two electrodes is purely capacitive, explaining the low gain a t

IEEE Transactions on Electrical Insulation

Vol. 23 No. 6, December 1988

are shorted together (0 d B gain and zero phase). I t is evident from the equations that the frequency enters only
with the bulk conductivity - altering the conductivity
merely shifts the frequency response without changing
the curve shapes (Figure 7). Changing the bulk permittivity alters both the breakpoint (due t o a change
in r, and the high frequency gain (due t o a change in
the capacitive coupling), as shown in Figure 8. When
the phase is plotted versus gain for the curves shown in
Figs. 7 and 8 (eliminating the frequency as a parameter), the graph reproduces the lookup table developed
for the application of the microdielectrometer t o homogeneous systems by Lee [IO].

response of varying dz while holding the complex bulk

permittivities of each region constant is illustrated by
Figure 9. In this case, the half space is taken t o be more
insulating than the layer and the surface properties of
the layer-half space interface are taken t o be zero.
For a thick layer, dz > (A/27r), the response is that
of a uniform half-space, determined solely by the properties of the layer. T h e electrodes d o not 'see' beyond

0.
-20W

- 40 ,"
V)

-60-

-80I

-I

log f
Figure 8.
Predicted responses for microchip in uniform
medium (u1 = lo-' S/m), vary permittivity c l .

A layer of thickness dz and characterized by a uniform complex bulk permittivity is placed immediately
above the electrodes (Figure 6b). Solution of Laplace's
equation in a piece-wise fashion, perhaps using transfer
relations (see Appendix B or [14]), gives

As background for inferring the layer thickness from

-3

a measurement of the gain, the effect on the frequency

-40-

VARIABLE THICKNESS LAYER

+ coth(k,dz)]-'}

90 7

the thick layer for purely geometrical reasons. As the

layer thickness is decreased, the response is still a bulk
one, but now it also reflects the fact that the fields have
begun t o penetrate into the semi-infinite region, where
the permittivity is less (accounting for the decrease in
high-frequency gain) and the conductivity is zero. At
dz O.l(A/27r), the layer is so thin that it might well be
described by a surface conductivity on the interelectrode
interface. This case is discussed in the next Section.
Further decreases in the layer thickness reduces the effective interelectrode surface conductivity. This results
in a shift of the interelectrode surface type response to
lower frequencies. Generalization of the C, t o include
multiple layers is presented in Appendix B.

SINGULAR PROPERTIES A T
SUBSTRATE-MEDIUM INTERFACE
This is again the case of a uniform half-space (Figure sa) with C: given by (30), but with a surface conductivity a t the SiOz-medium interface between the elec-

Zaretsky et al.: Continuum Properties from Interdigital Electrode Dielectrometry

908

- 40
O

a
-*OF
el = c0 F/m

- IZ0L
- 3

-I

log f

Figure 9.
Predicted responses for microchip with variable
thickness layer (EZ = 2 x lo-" F/m, uz = 2.4 x
IO-'' S/m) in air.

trodes E:,, = -i(ano/w). A typical response is shown

in Figure 10 - clearly distinguishable from a bulk response. There is a n additional characteristic time, r6e=
( c 2 / k n a 6 , ) ,representing surface charge diffusion along
the interface. Again the interplay between surface and
bulk conduction and displacement currents accounts for
the transition from low to high gain and a peak in phase.
However, the slope of the gain is steeper than -20 dB/
decade and the gain has an overshoot. The phase curve
is asymmetric and the phase peaks a t a larger angle than
previously seen. All these are characteristic of the surface charge diffusion from driven t o floating electrodes
made possible by the combination of the distributed surface conduction and the shunt capacitance between the
interface and the highly conducting substrate.
As the surface conductivity decreases (rneincreases),
the curves pass through a regime where surface and
bulk conduction are equally important to a point beyond
which a purely bulk response is observed (Figure 11).

log f

Figure 10.
Predicted response for microchip in uniform
medium (c1 = 2x10-" F/m, u1 = lo-" S/m)
with surface conductivity on interelectrode surface (uao= 1 0 - l ~ S).

The overshoot phenomena, a prediction of gains below the purely capacitive high-frequency gain, is further
evidence of the surface charge diffusion process. This
can be explained by separating the coupling between the
electrodes into two components, one through the bulk
media and one through a distributed transmission line
composed of the interelectrode surface and the insulating oxide layer. The coupling through the bulk makes
a smooth transition from purely conductive to purely
capacitive, just as in the bulk response, and therefore
involves a t most a 9 0 " phase shift. However, the part
of the signal resulting from transmission along the distributed RC transmission line comprised of the interface and the insulating oxide layer suffers larger phase
shifts and hence a contribution that tends t o cancel that
due to the 'direct' coupling. The result is a frequency
range over which the gain is smaller than that a t high
frequency and the phase shift larger than 90 * . The

I E E E !I'ransactions on Electrical Insulation

Vol. 23 No. 6, December 1088

000

-rn
U

\\

.-c
0

c3

\I

.-

-30

a
In
0

.c

-3

-I

log f

Figure 11.
Predicted responses for microchip in uniform
medium (cl = 2 x lo-" F/m, 01 = lo-" S/m),
vary surface conductivity on interelectrode surface ( o ~ ~ ) .

diminution of surface conductance coupled with a n existing large bulk capacitance 'snaps' the response from a
surface conduction coupling t o a purely capacitive coupling and accounts for the rapid changes exhibited in the
frequency response. This behavior provided the r a t i e
nale for the distributed parameter model, as developed
and implemented by Garverick [ll].

LAYER WITH SURFACE CONDUCTIVITY

In practical applications such as the reduction of

electrostatic discharges (ESD), polymer films are often
coated with conducting films. With the objective of
measuring the surface conductivity of these films without electrical contact, the frequency response of a layer
(Figure 6b) with surface conductivity attributed t o the

-3

-I

log f

Figure 12.
Predicted responses for microchip with a layer
( d 2 = 5pm, c z = ZxlO-" F/m, u2 = lo-' S/m)
having a variable surface conductivity ( 0 . 2 ) in
air.

layer-medium interface is of interest. T h e surface capacitance density is given by (31) with

having only
a n imaginary component.

&a2

In Figure 12 is shown the frequency responses for

a n electrode structure with a layer thickness dz = 0.1X
in air. By increasing the layer's surface conductivity,
aa shown, two relaxation phenomena are observed. The
higher frequency one results from the spectrum of times
associated with diffusion of surface charge along the upper surface of the layer. The lower frequency one is
from relaxation determined by the bulk properties of
the layer. If c82 << CQX, so that bulk conduction in
the layer dominates t h a t of the surface, then relaxation
through the bulk of the layer will come into play a t a
higher frequency and the relaxation due to the surface
conductivity will not be so easily observed. Altering the
layer thickness, with a n observable surface conductivity, will affect the capacitance of the layer and thus, the

g10

Zaretsky et al.: Continuum Properties from Interdigital Electrode Dielectrometry

mid-frequency gain plateau and the higher frequency

phase peak.
SENSITIVITY T O ELECTRODE STRUCTURE
PARA M ET E RS

There are four parameters describing the electrode

structure, each with its individual effect on the gainphase response. The spatial wavelength X has been discussed previously with the application of intimate sensing. Thus, this value should be consistent with the characteristic length of the medium under measurement.
Two other parameters are the insulating layer thickThey are dependent on the
ness h and permittivity,
material used (for .cox)and the fabrication method (for
h). As mentioned earlier, varying h can significantly
affect the response t o a surface conductivity a t the electrode boundary. This effect can be extended further
to include the response to thin ( 4 O.l(X/27r)) layers having either bulk or surface conductivities. With
a relatively thick, high-permittivity oxide layer, there is
no surface-like response with overshoot gains and asymmetric, large phase peaks when examining such thin layers [15]. As shown by Li, this can be altered so as to
optimize sensitivity to these type of phenomena by reducing h and

The fourth parameter is the interelectrode spacing

a. Using physical intuition, if a is decreased then the
electrode-ground plane structure approaches that of a
parallel plate capacitor. It is as if a is an electric shutter controlling the amount of electric field penetrating
into the medium. Sensitivity to changes above will be
reduced with small values of a. However, increasing
a, while improving sensitivity, decreases the net coupling between the electrodes. This reduces the floating
gate voltage and requires the electronics t o accurately
measure very small voltages with correspondingly small
signal-to-noise ratios. Thus, it appears that the (X/4)
width for a is a valid middle ground.

PARAMETER ESTIMATION

This lack of sensitivity with thick, high-permittivity

insulating layers actually extends to measurements of
not only surface but bulk properties of thick and thin
layers. Referring to (2), the gain changes with the material above due mainly to changes in Y12. The admittance Yl1 is basically capacitive and much larger than
the capacitive part of Y12. As shown by Li [15], Y12
can be thought of as representing two parallel couplings
between the floating and driven gates, one above the interface and through the medium, and the other below
the interface and through the insulating layer. With h
large, the couplings above and below are equal. Changing the permittivity of the medium above will change
the net coupling (the sum of the two couplings) by a
much smaller percentage. If the permittivity below is
much larger than above, this percentage change shrinks
even further. With h small, the coupling below is much
smaller than above. Most of the field below couples
to the ground plane, thus the contribution to Y12 from
below is greatly reduced. Now changes in the medium
above are more fully reflected in changes in Y12 and
the sensitivity is greatly enhanced. Thus, for maximum sensitivity, the oxide layer thickness and permittivity should be minimized (subject t o fabrication constraints).

GENERALAPPROACH

N the previous Section, the problem was posed as

one of determining a frequency response, given all

the pertinent complex permittivities, layer thicknesses,
and other relevant parameters describing the media. In
practice, the situation is usually reversed. The unknown
quantity is a material property such as complex permittivity (bulk or surface, distribution if inhomogeneous)
or layer thickness. What is known is one or more experimentally measured responses, perhaps a t different
temporal or spatial frequencies. A parameter estimation
scheme is required in which all a priori knowledge of the
physical situation is utilized, such as number of layers,
values of complex permittivities for each layer, and layer
thicknesses. Noise or other stochastic processes affecting the experimental data call for more sophisticated
techniques, not considered here. Thus, here the estimation routines may be viewed either as root searching or
function minimization routines [16]. The search is for
the root t o the equation
e(@)= 0

(32)

where e(0) is a set of error functions (usually the difference between one or more measured and predicted gains)
and 0 is a list of parameters to be estimated (such as
complex permittivities). Of course, the particular physical phenomenon associated with the parameters to be
estimated must make a significant contribution to the
gain, and hence, the error.

A secant method of searching was employed for

most of the parameter estimations described here (171.

I E E E Transactions on Electrical Insulation

Vol. 23 No. 6, December 1988

In this method, guesses 8; are updated by A8; using the

secant formed by the two most recent guesses,

(33)

The danger that the root does not necessarily remain bracketed by the two guesses is decreased by providing bounds on the minimum and maximum values
of the estimated parameter based upon physical const raints.
For application to microdielectrometry, acceptable
tolerances for convergence are determined by the experimental error in the response. The present version of the
device, when using a microchip sensor, has an accuracy
of 0.1 dB in gain and 0.1 in phase [7]. In the routines,
the gain tolerance is tightened to 0.05 dB. This translates to a tolerance of 0.5% in gain a t -40 dB and 0.1%
in phase a t -90 * .
The emphasis was on developing search routines
that converged and were somewhat robust, not on generating the optimal search method. Most of the validation of these search algorithms was obtained using
experimental data on semi-insulating materials. Consequently, there are areas of parameter space where the
convergence and robustness of these routines have not
been examined.

911

&,,

complex variable 8, then so are

and 6. Thus, for the
class of constitutive laws used here, it is meaningful to
apply directly one-dimensional root-finding techniques
to the estimation of a complex parameter.
Note that although one parameter is being estimated, this does not constrain the medium ,to be uniform. Using the surface capacitance density, C i , heterogeneous media can be represented as long as all but one
of the parameters necessary to characterize the medium's
dielectric properties is known.

LAYER THICKNESS
This is the case of a layer and a surrounding uniform
medium, as pictured in Figure 6b. At high frequencies, when the electrodes are purely capacitively coupled through the material, the phase will be zero and
the gain will be a function of the layer thickness and the
permittivities of the layer and the surrounding medium.
Given any three of these four quantities, it is possible
to estimate the fourth. Here the layer thickness dz is
unknown, thus 8 = dz and the search is for a purely real
root.

E I =eoF/m

- IO

t /

I x 10-10

2-20

SINGLE PARAMETER ESTIMATION

For these estimations, the error is defined as the
difference between the measured and the predicted complex gain,

where 8 iy the complex parameter to be estimated. Here,

&n and G,, are formed by evaluating the complex logarithm of the complex voltage ratio,

I "D I
where the real part is the gain in dB divided by 20 and
the imaginary part is the phase in radians (closely related to the output of the microdielectrometer). If G ,
as defined in (2), is a complex analytic function of the

.-C

5x

3x

-30

lo-'1
lo-11

2 x 10-1'

EO

-4 0
I

Figure 13.
Predicted high hequency (10 kHe) gains for microchip with variable thickness layer (&), vary
layer permittivity ( L Z ) , in air.
For the case of air as the surrounding medium, the
curves of gain versus layer thickness for various layer
permittivities are shown in Figure 13. Finding the root
of these functions is relatively straightforward. For robustness, the routine first checks whether the experimental d a t a makes sense by calculating the gain of a

912

Zaretsky et al.: Continuum Properties from Interdigital Electrode Dielectrometry

uniform medium having the lesser of the two permittivities. This gain should be less than the experimental
gain if a layer thickness is to be estimated. After passing this check the routine uses the secant method to
find the root. The search is conducted using 25 collocation points ( I C ) and 100 summation terms ( N ) . Due
to the well-behaved nature of the curves, the thickness
is uniquely estimated within 3 to 6 iterations. For a
50 pm wavelength electrode structure, the 0.1 dB gain
tolerance implies a sensitivity of better than 5% in the
thickness estimate within the range of 2 to 12 pm. The
lower bound may be determined by the experimental error in the measurement or by features of the electrode
that are not modeled such as electrode thickness. The
upper bound reflects the decrease in sensitivity due to
the exponential decay of the electrostatic fields.
COMPLEX PERMITTIVITY (BULK OR SURFACE)

AS long as $ ( e ) is an analytic function o f a complex

bulk or surface permittivity, the search is only slightly
more complicated. A complex bulk or surface permittivity is estimated using a complex data point (gain and
phase) and specifying all other relevant parameters such
as the complex bulk permittivities of the other regions,
the complex surface permittivities a t the interfaces and
the layer thicknesses. Now, in (33), 6 = E - id or
E: - ie; and .^(e),in (34), is complex.

The search routine can be started either with a user

input guess or with its own guess as to the complex permittivity. For bulk properties this guess corresponds
to that of an insulating, nonpolar medium. The first update only is calculated using the local derivative. For the
secant met hod the order of convergence is the golden
ratio of 1.618.. . [18]. The key here is t o get in the neighborhood of the root where the convergence is rapid. A
substantial reduction in the time to convergence is obtained by letting the routine start with its own guess,
and using a technique of varying the number of collocation points used for the function evaluation. This
technique performs sequential searches, beginning with
a coarse discretization of 2 collocation points and the
routines initial guess, and culminating with a search
using 25 collocation points and an initial guess from the
previous coarse estimate. Generally, anywhere from 20
t o 40 iterations are required for convergence with 2 collocation points, resulting in a requirement of only 1 to 6
subsequent iterations at 25 collocation points. It takes
roughly 12 times longer to compute a gain-phase response using 25 collocation points versus 2 collocation

(100)

-40

- 30

- 20

-I 0

I-

I20

Gain ( d B )

Figure 14.
Parameter space for complex bulk permittivity
estimation of a 5 micrometer layer in a uniform medium ( E ; = 2 x 1 0 - F/m and E : =
lo- F/m). (4 and & are in units of E F/m).

points. Thus, a substantial savings in computing time

can be realized.
A two-dimensional view of the space over which the
routine searches when estimating the complex bulk permittivity of a layer 5 p m thick surrounded by a medium
with a bulk permittivity of E: = 2 x 10-l F/m and a
loss factor of E: = 1x lo- F/m is shown in Figure 14.
This space is constructed by computing the gain-phase
response for a matrix of complex bulk permittivities of
the layer a t a constant frequency. The contours are lines
of constant E ; and E and are orthogonal because 6 is an
analytic function of E * . Reiterating, this orthogonality
justifies the computation of the complex slope used by
the one-dimensional secant search routine. As required
by (35), the gain is in dB divided by 20 and the phase in
radians multiplied by log e. This figure is analogous to
the lookup table given by Senturia, et al. [9] for the case
of a uniform medium. With the gain defined by (35) the
contours are orthogonal for the uniform medium case
too.
In Figure 14, the values of E; that make the layer
more lossy than the upper medium are concentrated
near the origin. The dot in Figure 14 is where the layer
and surrounding medium have equal complex permittivities. By contrast, in Figure 15 the upper medium
is sufficiently insulating that the point of equal complex permittivities (the dot) is near the zero phase axis.
Again, with a 5 p m thick layer, the surrounding medium
now has a bulk permittivity of E: = 2 x 1 0 - l F/m and a
loss factor of E: = 1 ~ 1 0 - ~F/m.

The kink that appears

in the left side of the plot is a result of charge diffusion
along the layer, as discussed earlier in regards to the
thin film humidity sensor.

IEEE Transactions on Electrical Insulation

Vol. 23 No. 6, December 1988

-20

- 4 0 'y

p
-60
-80

913

Simultaneous parameter estimation of more than

one complex parameter clearly requires several experimental d a t a points, varying the value of a known parameter. As pointed out above, there should be a t least
one data point representative of each physical process of
interest.

n
-100
- 120
- 40

-30

-20

-10

SUMMARY REMARKS

Gain ( d B )

Figure 15.
Parameter space for complex bulk permittivity
estimation of a 5 micrometer layer in a uniform medium (e; = 2 x lo-" F / m and e;' =
IO-' F / m ) . (e; and e:' are in units of eo F / m ) .

Given the well-behaved nature of the spaces represented by Figures 14 and 15, it is expected that estimation of unique complex permittivities is a straightforward process. The sensitivity of the estimation routine
can be observed from the two Figures. Poor estimates
occur in regions of high contour density. For example,
gain-phase d a t a adjacent to the gain axis yield poor
loss factor estimates while d a t a with gains near zero
yield poor permittivity estimates. The latter is a consequence of the high-impedance mode of operation which
utilizes the measurement of a floating voltage.
For the special case of ohmic media, E" = a / w .
Thus, contours of constant E; are also the gain-phase
trajectories obtained if the temporal excitation frequency
is varied. This may also be seen by eliminating the independent parameter of frequency from the plots in Figure 8.
For the gain-phase spaces of Figures 14 and 15,
valid d a t a must lie in the region having a lower bound
given by the E' =
curve. Upper and lower bounds
can be placed on the parameter estimates to prevent
the search from wandering too far off. These bounds are
physically motivated - no bulk permittivities less than
and no conductivities less than zero. Upper bounds
are chosen based on reasonable guesses of the order of
magnitude of the dielectric properties of the materials
to be measured.

In addition to estimating bulk parameters, this routine works just as well for estimating complex surface
parameters.

HE modal representation relating the gain-phase

response of the microdielectrometer to the continuum properties of a medium has been formulated so as
to retain generality and flexibility in its application to
various linear systems. This approach greatly extends
the usefulness of the device by providing a framework
for interpreting data obtained for heterogeneous media
in terms of absolute continuum parameters rather than
'lumped equivalents'. Typically, the expressions summarized here are programmed as a subroutine which can
then be incorporated into a parameter-estimating main
program. The main program requires the surface capacitance densities for the specific medium thought to be
under investigation. The medium might be represented
by a system of discrete layers, with the computation of
the surface capacitance densities organized as in Appendix B.
Although parameter estimation using the continuum model described here is applicable to multi-variable
systems, the key to this and other approaches is in identifying the types of d a t a that can be mapped into the
continuum properties and geometry. Illustrated here
have been one-dimensional schemes for the identification of film thickness, the complex permittivity of finite
thickness films and the surface conductivity of films. As
a practical matter, these are being used for interpretation of experimental d a t a in an on-line manner. Also
developed, though not discussed here, is the estimation
of both the thickness and permittivity of a film from two
measurements of gain.
Closely related to the techniques described here are
those under development for nondestructive evaluation
and robotic sensing [19-211. Capacitive electrode arrays
are again employed. However, a finite element model is
used to interpret experimental results.
The ability to keep track of basic continuum parameters has proven to be useful in several contexts. One
is the estimation of thickness and properties of plasmadeposited layers and of sedimenting colloidal particles in

Zaretsky et al.: Continuum Properties from Interdigital Electrode Dielectrometry

914

liquids [22]. Work will be reported elsewhere in which

these techniques have been used to investigate transformer oil, with specialized coatings used to turn the
microdielectrometer into a means for sensing either the
complex bulk permittivity of the oil [22] or the moisture content [23]. The ability t o infer from the frequency response the continuum parameters needed to
quantify surface adsorption of anti-static agents and
trace amounts of water while keeping track of changes in
the bulk properties of coatings is essential in sorting out
heterogeneous phenomena. There are, of course, limits
on how much information can be gotten from varying
the temporal frequency. However, by varying the fundamental spatial wavelength of the electrodes, a new
approach can be taken to sorting out distributions of
properties. Parameter estimation schemes for inferring
the spatial distribution of properties by exploiting measurements of gain made at variable wavelength and fixed
temporal frequency can use the numerical approach that
has been derived here and will be described elsewhere.

n= 1

Fn

sin(na) cos(np)
n2

To reduce the computational time required for this

summation a more rapidly convergent series was used.
Expression (36) can be expanded to the form of
(37)

Assuming Hn --+ Hminas n

W

00

the sum

sin(n7)

n2

Hmin
n=l

will be computed and the difference between Hn and

Hmin will be accounted for later on. From Oberhettinger [24] comes the expression

2 - 1'
=-

In [2sin(O/2)] de
(39)

n=l

ACKNOWLEDGEMENT
Using a series expansion [25]
work is being carried out in the Laboratory for
Electromagnetic and Electronic Systems at MIT,
with financial support from members of MIT's Electric
Utilities Program. The companies which are supporting the work include Allegheny Power Co., American
Electric Power Service Corp., Boston Edison Co., Empire State Electric Energy Research Corp., Northeast
Utilities Service Co., N.Y. Power Authority, Southern
California Edison Co. and Tokyo Electric Power Co.

HIS

Professor Stephen D. Senturia provided useful commentary on the intellectual and stylistic content of this
paper. Micromet Instruments, Inc. and in particular,
Huan Lee and David Day provided critical information
regarding the operation and performance characteristics
of the microdielectrometer.

APPENDIX A
RA PI D LY C 0 N V E R G EN T S ER I ES

T
of

individual terms in (23a) and (25a) that are to

be summed over the indices n have the general form
HE

t2

t4

t6

In[sin(t)] = ln[t] - - - - - 6
180 2835
-

... -

(-

l)m+122mB2mt2m

2m( 2m!)
W

= 1n[t1m=l

(40)

(- l ) m + 1 2 2 m B 2 , , p
2m( 2m!)

for t 2 < x 2 where B; are the Bernoulli numbers, and

substituting in (39) yields

n=l

.-

Due to the factorials present in the denominator in

the summation term of (41), only 5 to 10 terms need
to be computed to obtain a high degree of accuracy
(comparable to computing the first 1000 terms of the
original series). At this point, a calculation of the terms
for which Hn is significantly different from H,n;n is performed and the difference is added to the results of (41).

IEEE Transactions on Electrical Insulation

Vol. 23 N o . 6, December

APPENDIX B

1988

915

From these equations, it follows that

SURFACE CAPACITANCE DENSITY OF

MULTIPLE LAYERS

&+l)

22

HE medium may actually be comprised of a finite

number of layers, each layer described by a set of
parameters, or layers may be used to approximate what
is actually a smoothly inhomogeneous material. With
the objective of describing either of these situations,
suppose that the medium is composed of P layers, as
shown in Figure 16. The j t h layer has a thickness dj
and an upper surface designated by j . The surface per1 surface, the electrode-medium
mittivity of the P
interface, is handled differently from other continuum
properties and designated czo. There is no cZ1 as this
surface property is associated with the j = 1 surface and
is normally a t infinity. Solution of the field laws within
the layer gives the transfer relations [14],

A$)Ac;i,,

= Am
+

(Ad - Ac;i,)+ k y j

(47)

This expression can be used repeatedly, starting

from the top layer ( j = 1) and working down to the
P t h layer. BY definition, the surface capacitance density called for in evaluating the complex gain is

where DC'l) and 62") are respectively the complex

amplitudes of the n t h Fourier components of the dielectric flux density and potential, evaluated just above the
lower interface of the layer and 6;')' and &;I' are respectively these quantities evaluated just below the upper interface. In the case of a layer having a uniform
complex permittivity [14].

en ep1)

@;ti)

&(j)'

j ' t h surface

- j ' t h layer

n(j+O

Dn

f b!+l)'
G P

At the j t h interface, the potential is continuous, but

there is a complex surface permittivity &ij
and hence a
discontinuity in the dielectric flux.

&(i)

(48)

Figure 16.
Medium above electrodes represented by a multilayered structure of P homogeneous layers.

(44)

The surface capacitance density of the j t h interface

is defined in terms of the quantities evaluated just above
that interface.

Dp
=-

e(j)

6);

(46)

In the case where the first surface is a t infinity,

terms in (47) are Ier0 so that cp'
be
evaluated without
For the finite thickness layer
bounded by an infinite half space, P = 2 and (31) follows from evaluation of (47) with CA2)- 4A(1)
2 .

the

&A1).

916

Zaretsky et al.: Continuum Properties from Interdigital Electrode Dielectrornetry

REFERENCES

[l] J . C. Ziircher and J . R. Melcher, Double-Layer

Transduction At A Mercury-Electrolyte Interface
With Imposed Temporal And Spatial Periodicity,
J . of Electrostatics, Vol. 5, pp. 21-31, 1978.
[2] J . Melcher, Electrohydrodynamic Surface Waves,
in Waves On Fluid Interfaces, R. E. Meyer, ed., Academic Press, NY, pp. 167-200, 1983.
[3] J . R. Melcher, Charge Relaxation on a Moving Liquid Interface, Phys. Fluids, NO. 10, pp. 325-331,
1967.
[4] S. M. Gasworth, J . R. Melcher and M. Zahn, Induction Sensing of Electrokinetic Streaming Current, in Conf. on Interfacial Phenomena in Practical Insulating Systems, Natl Bureau of Standards,
Gaithersburg, MD, Sept. 19-20, 1983.
[5] N. F. Sheppard, Jr., D. R. Day, H. L. Lee and S. D.
Senturia, Microdielectrometry , Sensors and Actuators, Vol. 2, pp. 263-274, 1982.
[6] S. D. Senturia and S. L. Garverick, Method and Apparatus for Microdielectrornetry, U. S. Patent No.
4,423,371, Dec. 27, 1983, assigned t o MIT, licensed
to Micromet Instruments, Inc.
[7] Eumetric SYSTEM I1 Microdielectrometer manufactured by Micromet Instruments Inc., 21 Erie St.,
Suite # 22, Cambridge, MA 02139.
[8] Low Conductivity Sensor, Model S-20, available from
Micromet Instruments Inc., 21 Erie St., Suite # 22,
Cambridge, MA 02139.
[91

s. D. Senturia, N. F- ShePPard, Jr.1 H. L. Lee and D.

on Electron Devices, Vol. ED-29, No. 1, pp. 90-94,

1982.
[12] T. M. Davidson and S. D. Senturia, The Moisture Dependence of the Electrical Sheet Resistance
of Aluminum Oxide Thin Films with Application to
Integrated Moisture Sensors, in Proc. IEEE, Intl.
Reliability Physics Symp., San Diego, CA, pp. 249252, 1982.
[13] L. Mouayad, Monitoring of Transformer Oil using
Microdielectrometric Sensors, SM Thesis, EE, MIT,
Cambridge, MA, Feb., 1985.
[14] J. R. Melcher, Continuum Electromechanics, MIT
Press, MA, p. 2. 33, 1981.
[15] P. Li, Low Frequency, Millimeter Wavelength, Interdigital Dielectrometry of Insulating Media in a
Tbansformer Environment, SM Thesis, EE, MIT,
Cambridge, MA, May, 1987.
[16] L. Ljung, System Identification: Theory for the
User, Prentice-Hall, Inc., NJ, 1987.
1171 T. R. Cuthbert, Jr., Optimization Using Personal
Computers, Wiley, NY, 1987.
[18] W. H. Press, B. P. Flannery, S. A. Teukolsky and W.
T. Vetterling, Numerical Recipes, The Art of Scientific Computing, Cambridge University Press, NY,
1986.
[19] B. A. Auld, J. Kenney and T. Lookabaugh, Electromagnetic Sensor Arrays - Theoretical Studies ,
Review of Progress in Quantitative Nondestructive
Evaluation, Vol. 5A, (Ed. by D. 0. Thompson and
D. E. Chimenti), Plenum Press, NY, pp. 681-690,
1986.

[lo] H. L. Lee, Optimization of a Resin Cure Sensor, EE

Thesis, MIT, Cambridge, MA, Aug., 1982.

[20] P. R. Heyliger, J . C. Moulder, P. J . Shull, M. Gimple, and B. A. Auld, Numerical Modeling of Capacitive Array Sensors Using the Finite Element
Method, Review of Progress in Quantitative Nondestructive Evaluation, Vol. 7A, (Ed. by D. 0.
Thompson and D. E. Chimenti), Plenum Press, NY,
pp. 501-508, 1987.

[ll] S. L. Garverick and S. D. Senturia, An MOS Device for AC Measurement of Surface Impedance with
Application to Moisture Monitoring, IEEE Trans.

[21] M. Gimple and B. A. Auld, Position and Sample Feature Sensing with Capacitive Array Probes,
Review of Progress in Quantitative Nondestructive

R. Day, In-situ Measurement of the Properties of

Curing Systems with Microdielectrometry, J. Adhesion, Vol. 15, pp. 69-90, 1982.

I E E E Transactions on Electrical Insulation

Vol. 23 No. 6, December 1988

Evaluation, Vol. 7A, (Ed. by D.0. Thompson and

D. E . Chimenti), Plenum Press, NY, pp. 509-515,
1987.
[22] M. C. Zaretsky, P. Li and J. R. Melcher, Estimation of Thickness, Complex Bulk Permittivity and
Surface Conductivity Using Interdigital Dielectrometry, IEEE Intl. Symp. on Electrical Insulation,
Cambridge, MA, pps. 162-166, 1988.
[23] M. C. Zaretsky, J. R. Melcher and C. M. Cooke,
Moisture Sensing in Transformer Oil Using Thin
Film Microdielectrometry, IEEE Intl. Symp. on
Electrical Insulation, Cambridge, MA, pps. 12-17,
1988.
I241 F. Oberhettinger, Fourier Ezpansions, A Collection
of Formulas, Academic Press, NY, p. 9, formula 1.
25, 1973.
[25] W. H. Beyer, CRC Standard Math Tables, CRC
Press, FL, p. 350, 1981.

Manuscript was received on 14 Mar 1987, in final form 5 Sep

1988.

917