Journal of Physics D: Applied Physics

Related content
- An improved iteration method for
A direct method for the calculation of the edge calculating the axial potential distribution in
slit-aperture electrostatic lenses
capacitance of thick electrodes M J Wallington

- Capacitance corrections for disc

electrodes on sheet dielectrics
To cite this article: H El Kamchouchi and A A Zaky 1975 J. Phys. D: Appl. Phys. 8 1365 H J Wintle and D G W Goad

- Formulae for the potential and current

distribution in a thin crystal with coplanar
electrodes
C E Wilsdon
View the article online for updates and enhancements.

Recent citations
- Rajat Vishnoi et al

- 2-D Threshold Voltage Model for the

Double-Gate p-n-p-n TFET With Localized
Charges
Dawit Burusie Abdi and Mamidala
Jagadesh Kumar

- A full-range dual material gate tunnel field

effect transistor drain current model
considering both source and drain
depletion region band-to-band tunneling
Pratyush Pandey et al
J. Phys. D: Appl. Phys., Vol. 8, 1975. Printed in Great Britain. 01975

A direct method for the calculation of the edge capacitance of

thick electrodes

H El Kamchouchi and A A Zaky

Electrical Engineering Department, Alexandria University, Egypt

Received 24 February 1975, in final form 1 May 1975

Abstract. A new method is presented for the evaluation of the edge-capacitance correc-
tion for electrodes of finite thickness. The corrections are divided into three separate
components: electrode thickness, the lower edge of the electrodes, and the back of the
electrodes. Simple correction formulae are derived for the first two components and a
set of curves is provided which enables the direct evaluation of the third component from
an equally simple formula. Numerical results show excellent agreement with those
obtained by more tedious graphical methods.

1. Introduction

The edge capacitance of a circular electrode capacitor is usually calculated either from
Kirchhoff’s formula (1877) or from empirical ones (ASTM 1965). In all these formulae the
thickness of the electrodes is assumed negligibly small compared with the electrodes
separation. Recently, however, Lynch (1973) proposed a graphical method for the
evaluation of the edge capacitance taking into consideration the electrode thickness.
In this method the edge correction is separated into components at various distances
from the edge. These components are determined by drawing the equipotentials and
flux lines for three different regions : the lower corner, the vertical edge (thickness), and
the back of the electrodes. The method of drawing the flux lines and equipotentials is
essentially a relaxation method. For the vertical edge of the electrode the spacings of the
flux lines, except near the lower corner, were assumed to form a geometrical progression.
This assumption becomes increasingly inaccurate when the electrode thickness is of the
same order as the electrodes separation. Moreover, in the method proposed by Lynch
for finding the capacitance to the back of the electrodes, it is necessary to draw the
equipotentials and the flux lines for each diameter, each thickness and each separation.
The cases for which the presented edge-correction formulae are applicable are given in
the conclusions.
The present paper gives a simple and direct method for evaluating the edge capaci-
tance for thick disc electrodes without recourse to tedious field plotting. By using suitable
conformal transformation the edge capacitance for thick unguarded electrodes is attrib-
uted to three definite regions. For each region the capacitance is given by a simple
formula valid for any thickness and any separation. A set of curves is given by means of
which the back electrode capacitance can be directly calculated (without the need for
solving the transcendental equation resulting from the conformal transformation used)
over a wide range of electrode thickness and separation.

1365
1366 H El Kanzchouciii and A A Z a k y

The only assumptions used in the present paper are the same as those made by Lynch
and Kirchhoff: namely (i) the capacitance to the back of the electrodes is obtained by
replacing each electrode by a rectangular one having the same area as the real one, such
that two of its parallel sides are together equal to the perimeter of the original electrode,
and (ii) the total edge capacitance is the capacitance per unit length of a semi-infinite
plane multiplied by the actual electrode perimeter.
Numerical results obtained by the method given here are in agreement with those
obtained by Lynch's method and verified experimentally by Blanco White (I 973).
z - plane Iy
' 3l7

w- plane
(I
Figure 1. The geometrical configuration in the z-plane and its transformation in the
w-plane.

2. The transformation

The Schwartz-Christoffel transformation which transforms the edge of the thick plate
capacitor into the real axis in the w-plane (figure 1) is (Binns and Lawrenson 1963,
Durand 1966)

where
R= [(w+ l ) / ( w + ~ ) ] ~ ' ~

It should be pointed out that the value of a given by Durand (1966) is incorrect.
A second transformation is now introduced to transform the w-plane into the complex
potential X-plane such that the region U 2 0 is transformed into the entire real axis, and
the region U < 0 is transformed into an infinite line parallel to the real axis. Hence, the
upper electrode of the capacitor is transformed into an infinite plate as shown in figure 2.
This transformation is
x=4+j#=(V1/n) In w (5)
 Calculation of the edge capacitance of thick electrodes 1367

$ x - plane

~/////////////~/////////////////~
1:

where $ is the flux function (stream function) and $ is the potential function. The curves
$ = 01 and $ = p, where ac and p are constants, represent respectively flux or stream lines
and equipotential lines.
In the X-plane the lines of force are parallel to the $-axis and the equipotentials are
parallel to the $-axis. The point z=jg which corresponds to w = - 1 is transformed into
x=(Vl/n-)In(-l)=jV1 (6)
and the point z = j (gi-p) which corresponds to w = - a is transformed into
x=(Vl/x) In ( - a ) = ( V ~ / n )In (a>+jVl. (7)

3. Edge-capacitance calculations

The electric flux per unit length passing between any two points 1 and 2 lying on the
flux lines $ = $1 and $ = 4 2 is ~ ( 4 -
2 $1). The capacitance per metre length of a curvilinear
square bounded by two equipotentials $1 and $2 and flux lines $1 and $2 is given by
(Weber 1965)

In our problem $2 is taken equal to the potential VI of the upper electrode and $1 is
taken as the zero potential plane midway between the capacitor electrodes.
The total edge capacitance to the neutral plane is given by contributions from three
regions (figure 3) :
(i) the region bounded by the vertical side of the edge, the real axis and the flux lines
c#q and 4 2 . From equations (6) and (7) it is clear that 41= O and 4 2 =(Vl/n) In (a).

tI
-XZ
$0

0
I
Figure 3. Division of edge capacitance into three components.
1368 H El Kamchouclzi and A A Zaky

Hence the contribution to the edge capacitance from this region is

C I = ( ~ OIn
/ ~(a)
) (F m-1). (9)
(ii) The region bounded by the back of the electrode which extends from 4=4z to
I$=c$~and the real axis. To determine 4 3 , let z= -XI +j ( g + p ) which corresponds to
w = -u1. x1 is equal to half the smaller side of the equivalent rectangle as defined in
the introduction. In this case the complex potential by equation (5) is
x = (VI/?.) +
In (- ul) = (V1j.r)In ( a ) j Vi
and
4 3 = (V1i.r) In (~1).

Thus the capacitance due to this region is

CZE EO/^) In (ul/a) ( F m-1).
The relationship between XI and u1 is obtained by substituting z= -xl+j ( g + p ) and
w = -u1 in equations (1) and (2) and then equating the real parts of equation (1) to
give the transcendental equation

where
R = (;-;)l’z.
~

For any given value of XI, electrode thickness p and separation g, this equation can be
solved for u1 and the capacitance determined from equation (IO). The set of curves given
in figure 4 (obtained with the aid of a digital computer) give the relationship between
XI and ul/a for four electrode separations and thicknesses varying from 1 to 10 mm.
(iii) The region bounded by the lower surface of the upper electrode which extends
from C$ = $1 to 4 1 ’ and the real axis. The flux $1‘ is determined such that the capacitance
between $1’ and a flux line $0 (corresponding to z= -xz) is equal to the geometrical
capacitance. The location of $0 is at a distance sufficiently far from the edge such that
the field can be considered uniform. The capacitance between 40and 41 is
C’= -EOC#O/VI=(EO/T)I~
(lluz) (12)
where u2 <a. The geometrical capacitance between the point corresponding to $0 and
the edge is
Cg= w z / g (F m-1). (13)
Due to the concentration of the flux lines near the edge it is evident that C’>Cg.
The difference between C’ and Cgrepresents C3, the contribution to the edge capacitance
in this region. This difference is obtained as follows.
In the w-plane, the region - 1 < U < 0 corresponds to - 00 < x < 0 and this in turn corres-
ponds to an infinite capacitance. For a finite value of x= - X Z where x2 is large, there
corresponds a point w= - uz which is near to point C in the w-plane (figure 2). It follows
that uz is very small compared to unity and compared to a. Hence equation (2) gives
R = [(I - uz)/(a- U Z ) ] ~2/ (1
~ -

Substituting z = -xz+jg and the above value of R in equation (l), and noting that R is
 Calculation of the edge capacitance of thick electrodes 1369

I I I I J I
3 0 5 IO 15 20 0 5 IO I5 20

Figure 4. Relationship between xa and ua/a for 1-10" electrode thicknesses and
separations (2g) of (U) 0.5 mm, (b) 1.0 mm, (c) 1 5 mm and ( d )2.0 mm.

small we obtain after equating the real parts

+
u2 = 1 -tanh2 [(rx2/2g) (1 - U Z ) ~ / ~ ]
and since u~ is much smaller than unity
u2 = 1 -tanh2 0
1370 H El Kamchouchi and A A Zaky

where
0 = (nxz/2g)4- 1.
For large values of 0 we may write
tanh 0 = 1 - 2 [exp (- 20) - cxp (-48) + exp (-SO)-. . .]
and
tanh2 Or 1- 4 exp (-20) + 8 exp (-48);
thus
uzr4exp(-20)-8 exp(-40). (14)
Therefore equation (12) gives
C’=~oxz/g+(~0/~){2-1n4-ln
[1-2exp(-2O>]}=Cg+C3.
Since xp is very large compared to the electrode separation the exponential term is
very small and we may write
C=0.613 EO/^ (F m-1). (15)
This is equivalent to the capacitance between the flux lines $1 and 41’. The contribution
to the edge capacitance in this third region is almost independent of the electrode separa-
tion and thickness.
The total edge capacitance is thus half the sum given by equations (9), (10) and (15)
multiplied by the perimeter.
As a practical numerical application of the above results we have considered the
50 mm diameter Hartshorn-and-Ward electrodes for which Lynch obtained the edge-
capacitance correction graphically for 5 mm thick electrodes at various separations. The
results of these calculations for three electrode separations are shown in table 1 in which
Lynch’s results and those obtained by Kirchhoff’s formula are included for comparison.

Table 1. Edge capacitance corrections (pF)

Separation in mm (2g)
0.5 1.0 2.0
___________._~__

Analytical 2.172 1.883 1.595

Graphical (Lynch) 2.160 1.840 1.530
Kirchhoff 1.820 1.520 1.230

A typical calculation for a 1 mm gap is as follows. Using g=O*5, p = 5 , and from

equations (3) and (4), we find that a= 482. For 50 mm diameter electrodes, x2 = 12.5 mm
and the corresponding value uz/a= 5,625 is obtained from figure 4(b). From equations
(9), (10) and (15) we find that the total edge correction is
Ce= 17*4+4-86+1*72=23*98p F in-1.
The actual edge correction is half this capacitance to neutral times the perimeter
Cte = 0.5 (23.98) (0.157) = 1,883pF.

4. Conclusions

A new method of general applicability for evaluating edge-capacitance corrections for

electrodes of finite thickness has been presented. The method is based on conformal
 Calculation of the edge capacitance of thick electrodes 1371

transformations and, with reasonable assumptions, it has been shown that the edge-
capacitance correction can be obtained directly from three mathematical formulae
without recourse to field plotting. Two of these formulae are very simple, whilst the use
of the third has been rendered simple by provision of a set of curves which cover a range
of electrodes thicknesses and separations.
The results can easily be extended to the case of one disc and one infinite plane, and
to shapes other than circular discs as well as to the following arrangements: discs com-
pletely immersed in liquid; discs containing a solid specimen of the same size as, or
smaller than the discs, of thickness equal to the separation of the electrodes. It is also
approximately valid for the following arrangements : discs containing solid or liquid
specimens of much greater diameter than the discs, of thickness equal to or less than the
separation of the electrodes; discs containing a solid specimen of the same size as, or
smaller than, the discs but of thickness less than the separation of the discs.
It should be pointed out, however, that the results given in the paper do not apply
as such to an arrangement in which the disc electrodes contain a specimen slightly larger
than the discs.

References

ASTM 1965 ASTM Standards D 150-65 T

Binns K J and Lawrenson P J 1963 Electric and Magnetic Field Problems (London: Pergamon)
Blanco White J R 1973 Proc. IEE 120 939-40
Durand E 1966 Electrostatique vol2 (Paris: Masson) pp 304-5
Kirchhoff G 1871 Monatsberichte der Koniglich Preussischen Akademie der Wissenschaften zu Berlin
144-62
Lynch A C 1973 Proc. IEE 120 934-8
Weber E 1965 Electromagnetic Theory (New York: Dover)