A Broadband Multilayer CAD Tool

Utilizing the Method of Continuants

By
Douglas H. Colvin
2016
INTRODUCTION
A novel method for the analysis of multilayered isotropic homogeneous media is
presented. The media may be for optical or microwave applications such as antireflective coatings, radomes, and fiber Bragg gratings.
It is desired to find the power reflection coefficient (R) and transmission
coefficient (T) given any combination of layers, frequencies, or angle of
incidence. Pederson and Kuh [1] use the method of continuants for circuit
analysis; this method has also been used to evaluate lumped circuit transmission
line models [2].
An algorithm based on cascaded transmission line equivalents is employed. The
method of continuants and its algorithm allow the formulation to accurately
determine R and T at any frequency or wavelength. In addition, the effects of
optical wavelengths incident upon metallic media resulting in negative-real
permittivity (e.g., metamaterials) may be readily evaluated.
The use of continuants allows for extremely fast and efficient computation without
sacrificing accuracy. Methods of circuit synthesis with the aid of continuants are
presently under investigation by the author.

TRANSMISSION LINE EQUIVALENT NETWORK

Multilayered media may be modeled in terms of cascaded transmission line
sections whose z-parameters and characteristic impedances depend on layer
thicknesses, frequencies (or wavelengths), and angles of incidence. Figure1
illustrates the multilayered medium and its corresponding transmission line
network.
1

1
z1,a

2
z1,b

z2,a

z2,b

zn,a

zn,b

zN,a

zN,b
t

i
z1,c

z2,c

zn,c

zN,c

Vi

Figure 1. Multilayered Medium

The transmission line impedance values for z are given by:

z n,a z n,b z11(n) z12(n) jn tan ( f n n n )

z n ,c z12 ( n ) j n csc(2f n n n )
where n

n
intrinsic impedance of the n' th section
n

and n n' th section length

(1)
The circuit shown in Figure 1 may be cast into a simple network characterized in
terms of cascaded inverted L-sections as shown in Figure 2. The even
subscripts correspond to admittances and the odd subscripts correspond to
impedances.

Vt

v0
i

v2

x1
i1

v4 x5

x3
i3

i5

x2

zin

x2N-1 v2N x2N+1 vt= v2N+2

x2n-1 v2n x2n+1

i2N-1

x4

t= v2N+2

i2N+1

x2n

x2N

Vi
Figure 2. Multilayered Medium as Cascaded Inverted L-Sections
The values for the impedances and admittances in Figure 1 are defined by

x n ( odd ) z n 1 z n 1
2

,b

for 3 n(odd) 2N 1
,a

and
x n ( even )

(2)

for 2 n(even) 2N 2

z n ,c

where x 2N 2 t
The method of continuants must now be exploited in order to continue the
multilayered model.
METHOD OF CONTINUANTS
The circuit must be of the basic ladder network form with the component labeling
as shown in Figure 2. The input voltage is V0 and the output voltage is V2N+1.
The circuit currents and voltages are related by Kirchhoffs equations that are
given by a system of linear equations defined in (3). This system of linear
equations is pre-solved utilizing Kramers rule from which the continuants are
calculated. The multilayered structures overall z-parameters and gain are then
calculated directly from the continuants.

v0
0
0
0

0
0

x 1i 1
i1
0
0

0
0

v2
x 2v2

0
i3
x 3i 3
i3

0
0
v4
x 4v4

x 2 N 1i 2 N 1

i 2 N 1

v 2N
0

0
0
0
i5

0
0
0
0

(3)

v 2 N2
x 2 N2 v 2 N2

Three determinants extracted from (3), by use of Kramers rule, are used to
define the required continuants:

x1 1
0 0 0
0
0
1 x2 1
0 0
0
0
0 1 x3 1
0
0
0
0
0 1 x4 1
0
0

0 1 1
0
0

0
0 0 0 1 xn 1
0

0 1
1
0
0 0 0 0 1 x 2 N 1
0
0 0 0 0
0
1
0
0 0 0 0
0
0

0
0
0
0
0
0
0
1
x 2N
1

0
0
0
0
0
0
0
0
1

(4)

v 2N2

Two other determinants are defined as follows:

v0
0
1 0

1 0
x2 1
1 x3

0 0

x 2 N2

x1 1 0
1 x 2 1
0 1 x3
2

0 0 0
0 0 0

x 2 N 1
1

v0
0
0

0
0

(5)

The determinant completely characterizes the network because it contains all

of its elements along the diagonal; this is the basis of Kramers rule.
Equations (3), (4), and (5) have been derived through Kramers rule. The value
of the determinant is expressed as (x1,x2,x3,...x2N+2) or (x1x2N+2) called a
continuant. The continuant is computed using the algorithm shown in (6).

D (d1 d K )
q1 1
q 2 d1
q 3 q 2 d 2 q1

(6)

q 4 q 3d 3 d 2

q k q k 1d k 1 q k 2
qK

Evaluation of (4) and (5) yields:

q K 1d K d K 2

1
2

=
=
=

(x1x2N+2)
v0(x2x2N+2)
v0

(7)

Some of the most common network parameters are the network gain and the zparameters. Expressions for these are provided in (8).

network gain = g 12

z 11

1
1

(x 1 x 2 N2 )

(8)

2 (x 1 x 2 N2 )

and z 12 2
1
(x 2 x 2 N2 )
1 (x 2 x 2 N2 )

The values of x contain the sections intrinsic impedance that must be modified
corresponding to the angle and the mode of the incidence wave. The wave
mode is discussed in the next section.

TRANSVERSE-ELECTRIC (TE) AND MAGNETIC (TM) WAVES

The characteristic impedance of each section is equal to the intrinsic impedance
modified by the angle of incidence within each section. The wave front may be
either transverse electric or transverse magnetic as shown in Figure 3. The
modification of the characteristic impedance is derived from Snells law.

TM
wave

TE
wave
Figure 3. TM and TE Wavefronts

The field components E and H constitute a plane of incidence. If the electric field
lies within the plane of incidence the wave is said to be transverse electric (TE)
or parallel polarized. If the electric field is perpendicular to the plane of incidence
the wave is said to be transverse magnetic (TM) or perpendicular polarized.
The characteristic impedance of each section is a function of the angle of
incidence and mode (TE or TM). From Snells law we have:

n 1 sin 1 n 1 sin n , and

n

(9)

Z nT M n cos n , and Z nT E n sec n

where n

1
nn

speed of wave in medium n

The transverse wavenumbers are the same in all sections of the multilayered
structure and leads to the law of reflection. The expressions in (9) are valid for
lossy or lossless media. The multilayered problem is one of satisfying all
interface boundary conditions. If the angle of incidence is normal to the
multilayered structure the characteristic impedance of the sections is equal to the
section intrinsic impedance; this is called the TEM mode because both the
electric field and the magnetic field are transverse to the direction of propagation.
An example of reflection and transmission for TEM propagation through a sixlayer medium is shown in Figure 4 [3]; the analysis is performed in the microwave
range of frequencies.
The reflection coefficient is determined from the network (Figure 2) input
impedance and the source intrinsic impedance; this is given as:

and

z in i z in i i z in 2 i
2
1

1 i 1 2 i
z in i z in i i
z in
z in
2

T2

Et
Ei

i
2

(10)

i
t

The coefficients in (10) correspond to the electric fields. The power reflection
and transmission coefficients are given by |R|2 and |T|2 respectively.

i 0 , i 0

d0 0

1 0 , 1 (3.2 j.32) 0

d 1 15 cm

2 0 , 2 (3.5 j.7) 0

d 2 25 cm

3 0 , 3 (3.8 j1.14) 0

d 3 35 cm

4 0 , 4 (4 j1.6) 0

d 4 45 cm

5 0 , 5 (4.2 j2.1) 0

d 5 50 cm

t 0 , t (5 j3) 0

0.1

Power Transmission (-) and Reflection (...)

0.09
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0

1.2

1.4

1.6

1.8
2
2.2
Frequency (GHz)

2.4

2.6

2.8

Figure 4. Example of TEM Reflection and Transmission [4]

EXAMPLE OF APPLICATIONS
Fiber Stretching Analysis
An interesting case of optical fiber stretching has been addressed. The fiber may
comprise a fiber Bragg grating (FBG) that consists of several sections of different
refractive indices and lengths. The analysis is performed for each stretch
percentage; 100% means that the multilayered structure thickness is doubled.
The results are shown in Figure 5.
0.095
STRETCH = 0%

STRETCH = 50%

0.09

Power Reflection

STRETCH = 100%

0.085

0.08

0.075

0.07

1.2

1.4

1.6

1.8
2
2.2
Frequency (GHz)

2.4

2.6

2.8

Figure 5. TEM Reflection and Transmission Results Due to Stretching

It can be seen that the reflection curve compresses toward the left as stretching
is increased. The Broadband Multilayer CAD computer code, written in
MatLab, was used for this analysis. The code may be also be used to
investigate the effects of nonlinear stretching. The stretching and changes in the
index of refraction may occur due to environmental conditions such as
temperature or physical stretching.

Many other situations may be analyzed; however, this example was chosen for
sake of simplicity and demonstration of capabilities.
3-Dimensional Computations
Often it is desired to determine the reflection and transmission properties of
multilayered structures as a function of the wavelength and the angle of
incidence. Examples of the results for TM propagation are displayed in Figure 6.

i 0, i 0
1 0 , 1 1.7 0

2 0 , 2 2.3 j0.3 0
3 0 , 3 3.2 j.32 0

t 0 , t 80 0

Figure 6. 3D Results of a TM Wave

0.5 m
0.5 m
0.5 m

SUMMARY
Electromagnetic wave multilayer technology has been an important in the
development of radomes, anti-reflective optical coatings, and may other
applications. Multilayered structures consist of a stratified cascade of dielectric
media. A method for determining the reflection and transmission coefficients has
been presented. The results obtained are comparable to those in published
literature.
The author is currently working on a direct synthesis technique in which the
multilayered design is synthesized directly from the desired reflection and
transmission characteristics. This technique will utilize the mathematics of
continuants.

MATLAB CODE
% TITLE:
Multilayer EM Transmission Program
% AUTHOR:
Douglas Colvin
% PURPOSE:
This program computes the transmission and reflection coefficients
%
for multilayered structures such as slabs, transmission lines, AR
%
coatings, radomes, optical fibers and many more applications.
%
Angles of incidence and wavelength (or frequency) are the independent
%
variables and the reflection and transmission coefficients are the
%
resulting dependent variables
% ================================================================================
clear
% CONSTANTS
mu_0
=
4*pi*1e-7;
ep_0
=
8.85418781762039e-12;
c
=
1/sqrt(mu_0*ep_0);

% permeability of free space (Henry/meter)

% permittivity of free space (Farads/meter)
% speed of light (m/s)

% SOURCE &
mu_i
=
ep_i
=
mu_t
=
ep_t
=

%
%
%
%

v_i =
v_t =

SINK
mu_0*1;
ep_0*1;
mu_0*1;
ep_0*80;

1/sqrt(mu_i*ep_i);
1/sqrt(mu_t*ep_t);

permeability
permittivity
permeability
permittivity

of
of
of
of

source medium
source medium
sink medium
sink medium

% speed in source medium

% speed in sink medium

% OPERATIONAL STATS (INPUTS)

mode
=
1;
% TM = 1, and TE = 2
len =
[.5 .5 .5]*1e-6;
% thickness of sections in meters
mu_r
=
[1 1 1];
% relative permeability
ep_r
=
[1.7 2.3+j*.3 3.2-j*.32];
% relative permittivity
freq
=
c./linspace(1400,1600,128)*1e9;% frequencies (Hz)
angi
=
linspace(1e-12,90,128);
% angle of incidence (degrees)
% COMPUTATIONS
N
=
length(len);
mu
=
mu_r*mu_0;
ep
=
ep_r*ep_0;
eta_i =
sqrt(mu_i/ep_i);
eta_t =
sqrt(mu_t/ep_t);
z_l =
sqrt(mu./ep);
v_l =
1./sqrt(mu.*ep);
lamb
=
c./freq;
L
=
length(lamb);
ang =
angi*pi/180;
K
=
length(ang);

%
%
%
%
%
%
%
%
%
%
%

# of layers or sections
absolute permeability
absolute permittivity
source impedance
sink impedance
characteristic impedances for layers
wave velocity in layers
wavelength (m) corresponding to frequencies
# of wavelengths to analyze
angle of incidence (radians)
# of angles to evaluate

for k = 1:K
if mode == 1; qi = cos(ang(k)); end
if mode == 2; qi = 1/cos(ang(k)); end
for l = 1:L
for n = 1:N
if n ==1;
a(1) = asin(v_l(1)/v_i * sin(ang(k)));
if mode == 1; q(1)=cos(a(1)); end
if mode == 2; q(1)=1/cos(a(1)); end
end
if n >= 2;
a(n) = asin(v_l(n)/v_l(n-1) * sin(a(n-1)));
if mode == 1; q(n)=cos(a(n)); end
if mode == 2; q(n)=1/cos(a(n)); end
end
z(n,1) = j*tan(pi*freq(l)*len(n)/v_l(n))*(q(n)*z_l(n));
z(n,3) = z(n,1);
nn=2*n;

z(n,2) = -j*csc(2*pi*freq(l)*len(n)/v_l(n))*(q(n)*z_l(n));
X(nn) = 1/z(n,2);
end
if mode == 1; qt = cos(asin(v_t/v_l(N) * sin(a(N)))); end
if mode == 2; qt = 1/cos(asin(sqrt((3.1-j*.32)/80) * sin(a(N)))); end
for m = 1:(N-1)
mm = 2*m+1;
X(mm) = z((mm-1)/2,3)+z((mm+1)/2,1);
end
X(1) = qi*eta_i + z(1,1);
X(2*N+1) = z(N,3);
X(2*N+2) = 1/(qt*eta_t);
C(1) = 1;
C(2) = X(1);

for n = 3:(2*N+3)
C(n) = C(n-1)*X(n-1) + C(n-2);
end
CC = C(2*N+3);
D(1) = 1;
D(2) = X(2);
for n = 3:(2*N+2)
D(n) = D(n-1)*X(n) + D(n-2);
end
DD = D(2*N+2);
g(l,k)
=
1/CC;
z11(l,k)
=
CC/DD;
z12(l) =
1/DD;
T(l,k) = (2*g(l))*sqrt(qi*eta_i/(qt*eta_t));
R(l,k) = 1-2*qi*eta_i*DD/CC;
wl(l,k) = lamb(l);
angel(l,k) = angi(k);
end
end
if L>=2; figure(1);
mesh(wl*1e9,angel,abs(R).^2);
zlim([0 1]);
xlabel('Wavelength (nanometers)')
ylabel('Angle of Incidence (Degrees)')
zlabel('|R|^2')
end
figure(2)
hold on
stem(angi,abs(T(1,:)).^2,'b')
stem(angi,abs(R(1,:)).^2,'r')
hold off
grid on
ylim([0 1])

REFERENCES
[1] Kuh, Ernest S. and Pederson, Donald O., Principles of Circuit Synthesis,
McGraw-Hill, New York, NY, 1959, pages 229-234 (Out of Print).
[2] Colvin, D.H., Computationally Efficient Method of Calculations Involving
Lumped-Parameter Transmission-Line Models, IEEE-EMC-27, No.1, Feb. 1985.
[3] Leung Tsang, Jin Au Kong, and Robert T. Shin, Theory of Microwave
Remote Sensing, Wiley-Interscience Pub., New York, NY, 1985, page 31.