noc18-ee28-Lecture 14-Systematic procedure to obtain modes of a waveguide

IIT Kanpur July 201
20 followers
Dec 3, 2023
Welcome to NPTEL MOOC on Fibre Optic Communication Systems and Techniques.
We were discussing in the previous module, the mode, the concept of a mode.
And then I indicated at the end of the module that we are going to look at a systematic
procedure to understand these waveguide modes, okay.
There are a couple of assumptions that one talks about or one makes when we start understanding
the modes of a waveguide.
The waveguide mode or the, sorry, the waveguide structure is assumed to be uniform along the
z-axis, okay.
And the cross-section could be any irregular cross-section, but the cross-section would
remain the same along the z-axis.
So, that is what we mean by when we say that the waveguide is uniform along the cross-section.
For simpler treatments, we will also neglect any material losses.
So, all of the materials that our waveguides are made out of are considered to be lossless.
These two assumptions are of course very, very unrealistic, because you have in many
situations a requirement that your waveguide needs to be bent around, okay.
For example, in integrated optics, you may want to connect your input to multiple outputs,
so in which case it is necessary that you use couplers and those couplers have to be
spatially separated by bending appropriate, at appropriate angles, bending of the waveguide
at appropriate angles.
So, clearly this assumption is unrealistic, but it is sufficient for us to actually, I
mean that assumption what we use in understanding the modes of a waveguide is because you know
it is mathematical simplicity.
So, we do not want to complicate our analysis by allowing the waveguide to be non-uniform.
Moreover, we do not want to allow the cross-sectional variations that naturally occur when you manufacture
a fiber or fabricate a waveguide to actually cloud our understanding of the fundamental
transmission problems.
So, purely for mathematical simplicity, we are going to neglect any of the non-uniformities
that might appear and for the same reason, we are also going to neglect the losses, okay.
If you take advanced courses in waveguides, you would know how to incorporate the non-uniform
waveguide structure as well as losses and many other things that you may not really
like to consider in the practical waveguide.
Anyway, so the idea here is that I have a waveguide of some arbitrary cross-section,
and this cross-section remains the same as I go along the z-axis, which is the direction
in which the waveguide modes are assumed to be propagating, okay.
And this cross-section, although I have shown it to be very irregular, remains consistent
at all points.
So, I mean, hopefully this cross-section is the same as this cross-section, although the
figure may not really show it, but the cross-section remaining uniform is a key which allows us
to forget about the z part, because I can then express all of the z variations simply
by writing this as e power j kz z, okay, kz z as the propagating factor along the z and
then focus my attention to obtain f of x comma y for the triangular waveguides or f of r
comma phi for a circular waveguide or in general f of u comma v for a general cross-section
or general coordinates u, v and w.
So, these f of u, v or f of x, y, appropriately f of r, phi are called as the modes of the
waveguide and this is the mode that I am, that we are trying to find out, okay.
So, modes of the waveguide, this is called as modal analysis.
So, the procedure that we are going to follow to obtain this f of u comma v or f of r comma
phi or f of x comma y is called as modal analysis of the waveguide.
So, did you understand why we removed any z axis from our discussion or rather we assumed
z axis discussion to be in the form of e power minus j kz z, because we have assumed the
waveform, I mean the waveguide to be uniform along z axis and more propagating along z
axis.
So, more propagates along z axis, okay.
So, what is the systematic procedure?
Well, the procedure seems to be quite straight or you know is actually straight, not really
complicated, but of course the actual calculations may become tedious and complicated because
if you allow for very general cross-sections to consider, okay.
The idea here is that I know that electric field actually has two components, right.
So, I have what is called as a transverse component and then I have what is called as
a parallel component.
The parallel component is the component that would be you know along the z axis for the
case that we are considering or in general it would be the component that would be along
the direction in which the waveguide itself is extended, okay.
And this Et corresponds to the transverse components, that is components which are in
the plane that is transverse to the direction of the waveguide.
So, in the general case, u, v are the transverse coordinates, w is the parallel or the longitudinal
coordinates.
So, Et will be a vector in the u and v plane, whereas I mean having components u and v and
E parallel will be a component that would be along w direction.
Of course, it is important to realize that the transverse component could be function
of all three coordinates, right.
So, it is not necessary that the transverse coordinate be a function only of u and v.
Of course, this coordinate dependence is on the w axis is what we have captured by writing
as e power minus j Kw times w.
So, we do allow for dependence on w, but we impose a specific dependence on w in the form
of travelling wave.
So, all the components that are there as the electric field and the magnetic fields can
have transverse components or can be separated into transverse components and the longitudinal
component.
And each of those transverse components and the longitudinal component by themselves can
be a function of all three coordinates u, v and w, but along w, which is the direction
of the waveguide, we impose a very specific functional form, which is e power minus j
Kw times w.
If you are confused by u, v, w, you can replace u, v, w by x, y, z and work with planar cross
sections to simplify your, you know, understanding, but it is not really necessary to think always
of the Cartesian coordinate system, right.
The electric field we have decomposed into transverse and parallel components.
We will similarly decompose H fields also as transverse and parallel components with
the same remarks that go with the electric field.
So, the first step is to understand that you have E and H and then choose a coordinate
system, ok, in such a way that your life becomes simple.
For example, if the waveguide happens to be an optical fibre, which, you know, as such
as a circular cross section, then choosing a coordinate system x, y, z will be a very
big problem for you, because the analysis becomes very difficult as x and y components
couple to give rise to r and phi components or rather the components E x and E y will
be in a very complicated manner depend on x, y itself if you were to analyze them in
the circular cross section.
So, it is always better to choose a coordinate system in which the analysis actually becomes
quite simple, ok.
That itself is an art, but luckily we have had a literature of about 100 years to guide
us.
So, we do have standard coordinate systems.
And if your problem happens to be not falling into any of the standard coordinate systems,
then good luck to you.
Well, there are ways to, you know, get around that problem, but that is not our concern
here.
So, the first step would be to choose a coordinate system, which is simpler for you to work.
In this course, we will either be looking at x, y, z coordinate system, which is the
Cartesian rectangular coordinate system or we will be looking at r, phi and z, which
is the circular cylindrical coordinate system.
The second step is that, remember, we have decomposed our electric field and magnetic
fields into those E t and H t, E parallel and H parallel components, correct.
So, now what we do is, we express those transverse fields in terms of parallel components.
It is in fact possible to do so by applying Maxwell's equation.
So, express E t and H t as functions of E parallel and H parallel, that is longitudinal
components.
The reason for this is again mathematical simplicity.
There is no specific reason as to pick only the horizontal components or the longitudinal
components.
You could have done the other way around as well, but it is a simplified, I mean, this
analysis or this method simplifies the problem.
So, that is the second step.
So, I express E t and H t as functions of E parallel and H parallel using Maxwell's
equations.
So, whatever the coordinate system that I have chosen, in that coordinate system, I
have to express Maxwell's equations and then write E t and H t as functions of both E parallel
and H parallel.
In general, they will be functions of both E parallel and H parallel.
The third step is an interesting step, which essentially means or which essentially calls
for you to obtain and solve a wave equation.
Remember, these are all waves, except that the wave is not uniform, it is a non-uniform
wave, having certain x and y dependence or certain u and v dependence.
But it is your responsibility to obtain and solve wave equation for E parallel and H parallel.
Obviously, the functional form of the wave equation will be same for both longitudinal
components.
Therefore, if you solve one, you can easily guess the solution of the other, there will
be some constants of solution, which you will have to fix by applying the boundary condition.
So, apply boundary condition to determine the propagation constant.
Remember, one of the goals of your analysis would be to actually determine the propagation
constant kz or in general the propagation constant kw.
Of course, you will also have the perpendicular or the transverse wave number kt, kw will
be function of the transverse wave number kt and whatever the medium k vector that is
there.
Anyway, but you need to apply a boundary condition to actually determine these quantities, the
longitudinal propagation constant.
So, I will just write this as kw and this corresponds to the propagation constant or
I could have written it as k parallel.
The fifth step is to obtain the transverse and the transverse electric and magnetic field
components Et and Ht in terms of E parallel and H parallel.
We have already written Et and Ht in terms of this, but once I know the solution for
E parallel and H parallel by following this four step procedure, the fifth step procedure
is to go back to this step and then obtain the solutions.
These solutions of course have to satisfy boundary conditions as well and this will
allow us to obtain that transverse mode form f of u and v.
So, this five step procedure is what we are going to follow in the next few modules to
understand different waveguides.
Of course, our understanding of waveguides is for slab waveguides, which we have already
seen not really in this using the systematic procedure, but using a big picture of the
ray approach.
We will see that this approach also gives the same results and of course, this approach
gives you much more information.
In the ray approach, you would not know what is the way in which the fields are distributed,
whereas this one will allow you to know what the fields are distributed, ray theory does
not tell you the evanescent fields exist, but this theory will easily give you the evanescent
fields that exist and the functional form of that as well.
So, there is lot of merit in following this procedure and for optical fiber analysis,
the simple ray approach is not satisfactory.
So, you need to understand this systematic procedure to derive those mode fields of a
mode fields of the waveguide modes of an optical fiber.
So, we are going to look at this systematic procedure and before I look at it for the
slab waveguides and the optical fibers in the next modules, what I am going to do is
to reconsider the parallel plate waveguide, which consisted of two perfect electric fields.
And then see how far I can go with these steps to solve and obtain the corresponding electric
fields and the magnetic fields, the total electric and the magnetic fields or essentially
do a modal analysis of this particular problem.
So, here I can be little careful in choosing the coordinate system, I do not want to work
with negative values of x or rather x equal to minus h, I do not want to put in, I will
put x equal to h and x equal to 0.
So, this of course, will be the z axis, luckily for me, u is x, v do not need to be considered
and w is z.
So, the first step of choosing the coordinate system, I have already done it, this is the
z direction, this is the x direction.
So, this is how the x direction is and this is how the z direction is.
I have also assumed the uniform lossless property, so that the cross section at any time is actually
nice plane or a line here and this height would not change.
So, this height of x equal to h, I mean the height of h will not change as you move along
the z direction.
So, the first step is done choosing the coordinate system.
The second step of course, while choosing the coordinate system, I have also assumed
e power minus j kz times z dependence with kz being the unknown quantity that I need
to determine.
Remember, that was the idea of step number 4, so we have not reached step 4.
I am not going to fully solve this case, I am going to leave most of it as an exercise
to you.
I will just indicate the procedure.
I want to use this procedure to solve for the case of a slab waveguide in the next module.
So, these are basically indicative of what we are going to do, but the full use of the
systematic procedure will not be for this one, but it could be for the slab waveguide.
So, first step is to choose the coordinate system which we have done and along with that
we have also put down the way in which these fields depend on z which is in the form of
e power minus j kz times z.
Now, the second step would be to express the transverse components which are Ex, Ey, Hx
and Hy as functions of Ez and Hz.
Remember, these are the components that I have, the longitudinal components Ez and Hz.
To do that one, I need to apply Maxwell's equation.
The equation that I am going to use are the curl equations.
These are the ones that will allow me to interconnect electric and magnetic field.
In the case of a free space medium that I am considering between the two parallel plate,
between the parallel perfect electric conducting surfaces, this would be the curl expression
and this would be the expression for H, no current exists in between.
Therefore, this is simply j omega epsilon naught.
I am also assuming that the medium in between the parallel plate waveguide is a free space
medium.
So, what I get is this expression.
This is in the frequency domain or the time harmonic expressions.
Now, what is curl of E?
Curl of E is basically obtained by obtaining the determinant of this matrix, which will
be Ex, Ey and Ez and obtain the determinant of this one, that would be x hat del Ez by
del y minus del Ey by del z minus y hat del Ez by del x.
Since I am going to anyway take the minus sign later on, so I can actually rewrite this
as del Ex by del z minus del Ez by del x.
The z component of this field will be del Ey by del x minus del Ex by del y.
This should be equal to minus j omega mu naught x hat Hx minus j omega mu naught y hat Hy
and then minus j omega mu naught z hat Hz.
I do not really have time to carry on this analysis in this module.
So, what I am going to do is to take this step in the next module.
It is not, the problem is not finished because we are just at the point of, you know, solving
the step number 2, where we are trying to express the transverse field components.
But these expressions will also be useful for us to discuss the slab waveguide, which
we are going to do it in the next module.
So, we are going to use the systematic procedure to complete the problem for the parallel plate
waveguide and then use the same expressions to see how we can obtain the modal solutions
for the slab waveguide as well.
Thank you very much.
