noc19-ee21 Lecture 39-Distribution of Laser Beam

IIT Kanpur July 201
20 followers
Dec 3, 2023
Laser beam, Arago spot, aperture plane, TEM beam, back focal plane, beam waist, Gaussian integral, Gouy phase.
Hello and welcome to NPTEL MOOC on electromagnetic waves in guided and wireless media.
In the previous module, we started our discussions on Fraunhofer diffraction and we stopped at
the point where we were looking at the fields of a circular aperture, that is you have a
source far away from the aperture and then you have observing the aperture from a faraway distance
so that the aperture essentially looks like a circle of certain radius a. In that case,
we found that there is a bright central spot which sometimes is called as Arago spot or sometimes
called as Poisson's spot and around that central bright band you will see additional bright and
dark bands. So you will see that concentric circles except for the central bright spot
and the central bright spot does not have the same aperture area or the same radius as does
the aperture. In fact, it will be slightly bigger and this is what we actually call
as point spread function. This can be thought of as an impulse response
for those who are familiar with signals and systems. So this can be thought of as an impulse
response of an aperture or of a point source, the point source being equivalent to that of a
delta function. So we will not dwell on most of these concepts but these ideas are widely used
in the context of Fourier optics which essentially concerns itself with understanding the spatial
frequencies and how to filter them and how to modify them. This area is also quite important
for digital image processing where many of these concepts of images and image formation
require an understanding of how diffraction affects and how this point spread function
affects the resolution of the instruments which such as cameras and other medical instruments
such as microscopes and optical instruments such as astronomy, I mean astronomical instruments
such as telescopes and so on. While we could say lot of things about diffraction, let us not,
you know, because the scope of the course is not to deal with diffraction itself but to see
what kind of electromagnetic waves are, you know, are possible, the modes that are there
in free space and guided media. With that in mind, let us look at what happens
to a very special type of a beam that is emitted by a device called as laser. Laser as you already
know is, you know, used for emitting light at frequencies which are not just at the visible
range but also at different ranges. Of course, it is not just like any other light source such as a
lamp or a, you know, wood burning which emits wavelengths and other things. Light is very
special in the sense that it has a very high degree of coherence which I will not be able
to talk about it but for uninitiated you can think of coherent as correlated meaning that
the beam essentially remains in a very stable phase relationship with itself over long distances
and over long times. So, these lasers actually emit a beam which is not like a spherical beam
or it is not like a beam which is say a uniform plane wave. And, you know, what is important when
you study lasers is that the beam that is emitted by the laser will be used for further processing.
So, you for example want to cut a metal using laser or you cut some other material using laser,
then it is important for you to know if diffraction affects the beam. If it affects,
then how does it affect. So, you have laser output that is light output coming out from a very small
aperture and over the distance it kind of widens out because of the diffraction. So,
because this is an aperture and then it kind of widens out. The beam that is emitted by the laser
in its, you know, most stable configuration is what is called as a transverse electromagnetic
wave, TEMP00 mode and it has a very special distribution in the aperture plane. So, if you
examine the laser output in the aperture plane, you will actually see something like this.
So, if you cut along the x-axis, then you will see a Gaussian. If you cut along the y-axis,
you will see another Gaussian. So, it is basically going to be a two-dimensional Gaussian function
whose main, you know, or whose intensity will be peaking at the center and then gradually will
be decreasing as you move away from the center. Of course, it never kind of, I mean, it will not
go all the way to minus infinity to plus infinity, but these Gaussian functions are very good
approximations for the TEMP beam that is emitted by the laser. Now, please understand that what
we are talking about is the beam that is coming out of the laser and we are talking about its
distribution in the transverse plane. Please note that the wave is still propagating along
this black direction, for example, that could be thought of as the z-axis. However, in the xy plane
or in the transverse plane, if you look at it, that would be the Gaussian, two-dimensional
Gaussian function that I talked about. So, that plane is called as the aperture plane.
Now, for a beam that is emitted at say z equal to 0 plane and there is some interesting fact
about the z equal to 0 plane that I will talk to you about after I have introduced this diffraction
thing. We can represent the field that is coming out of this one. Remember our x0, y0 from the
last module, these are on the aperture plane coordinates x0 and y0. Here, we have certain
amplitude, let us say A and then the distribution is supposed to be a two-dimensional distribution.
So, you have x0 square plus y0 square divided by w0 square. Now, sometimes instead of w0 square here
in the denominator for this exponential function, you will also see people using 2, this will not
change anything in terms of the understanding or the concept, it only changes the numerical value
slightly. We will not worry about that factor of 2 here, we will simply write this as w0 square
and what is this w0 square is a very interesting thing. If for example, I consider the,
if I plot this exponential function or the Gaussian function as a function of x let us say
or x0 let us say while keeping y0 is equal to 0. So, I am basically taking a cut along the x-axis.
So, then this w0 square is the value of x0, when x0 is equal to w0, then at that point,
the amplitude of the wave would have dropped from A to A e power minus 1.
So, if you again locate at x0 equal to w minus w0, you will again see that that amplitude has
dropped to about A divided by e minus and this width is sometimes called as full width at half
maxima. So, it is some sort of measure as to what is the spread of this light beam in the x0
direction. Because of symmetry, the same spread you can observe in the y direction as well.
So, you can imagine that there is a circle whose central value is kind of bright or center is
bright, but then as the edges go around, then the amplitude kind of decreases and you can kind of
put a circle in that at the points x0 equal to w0 minus w0, y equals, y0 equals w0 and minus w0
and that circle would essentially tell you the size or the spot size of the beam.
So, accordingly, we call this w0 as the spot size of the beam. If you want to understand this one,
what you can do is to just take a simple laser pointer and darken the room a little bit so that
you can see the light correctly and then you take a white sheet of paper, paste it on some cardboard
and keep it at different distances or you can actually keep the laser pointer fixed,
maybe you can pull up a chair and then put the laser pointer fixed.
And then in the same, in the direction of that laser light going around, you place a beam or
rather sorry, you place a screen and then you can move the screen away and then you can observe
that the spot size would be bright, that is there will be a central bright spot and then there will
be small, you know, like or kind of the, you know, light is kind of smearing out at the edges,
but then as you take the screen far and far away, depending on how good or bad your laser is,
at some point, right, you will start to see that the central bright, the brightness of the
central spot starts to diminish and then everything starts to kind of expand out, okay, which means
that the size of the beam in this aperture plane or in this transverse plane is kind of increasing,
right and it is very important for us to know this increase, okay.
Why? Because in most cases, you would have not just going to use lasers, there will be some
optical system, for example, a simple optical system such as a lens, okay. The idea of using
a lens and then maybe you have some kind of a tissue here, which you are trying to illuminate,
I am just giving an example, you could illuminate it with anything else. So, if the laser, if the
position of the lens is very critical, because if you start off with a spot size that looks
something like this, okay, this is the laser beam and if this distance happens to be in the focal
plane of this lens, the so-called back focal plane of this lens, then it is possible for,
by choosing this distance appropriately, not exactly equal to f, but choosing this distance
appropriately, it is possible to make sure that the beam essentially retains or goes back to its
original spot size here at the image or the object point, right.
So, these kind of calculations require you to know what is W0 and what is the rate at which
this is spreading. So, for example, if you write down two horizontal straight lines,
then this would be an approximation for the beam divergence as we would call it and it is
important for you to know what the slope of these lines are, so that you can predict or you can
place this lens at the appropriate position in order to bring the light beam back into its original
spot size, okay. So, with that in mind and with the additional assumption that we are not really
in the near field zone for any of these waves, we are far away from the near field, what we want to
understand is what will be the spot size as a function of z, right. So, z axis being measured
from this minimum waste point, okay. So, this is spot size of the beam and sometimes also
called as the beam waste, okay. For obvious reasons, this is called as a beam waste and
then from there it kind of spreads, okay. The solution is not very difficult. It is
little tedious because you are going to work with Gaussian integrals, but if you look at a good
handbook, then what you are essentially, then you can and you can use those, you know,
relationships of that are given or tables of integrals that are known and Fourier transforms
that are known, what you are essentially trying to do is to take the two-dimensional Fourier
transform. Remember, the field at any point would be something like J by lambda z, right,
and you will have say e power minus x0 square plus y0 square divided by, so I will also have
an amplitude A that I am going to pull this out divided by w0 square, correct.
So, we have this x0 square plus y0 square by w0 square and then usually what you would have had
is e power J 2 pi fx x0 plus 2 pi fy y0, okay. So, this is what you had and then of course,
this is an integral over dx0 and dy0. What I am going to do because this can also be done,
you know the expressions for the Gaussian, so and you also know the idea that Gaussian
input will have a same Gaussian kind of a Fourier transform. So, you can use the tables of Fourier
transform and then get the output which would also be kind of a Fourier transform, right.
So, this is basically and you can also split this Fourier transform over x0 separately and
a Fourier transform over y0 separately, right. So, what would be the Fourier transform over x0?
This would be the Fourier transform of the one-sided Gaussian or single variable Gaussian
e power minus x0 square by w0 square, e power J 2 pi fx x0 dx0 is what you have and this would
turn out to be something like e power minus fx0 square w0 square with some constant which I will
call as say kappa x, okay. This constant needs to be multiplied in the or put into the numerator
because you know of the Fourier transform properties. I do not exactly remember the
Fourier transform, but these are available in a lot of textbooks and online and you can figure
out what would be this kx out there, okay. Similarly, the Fourier transform over y0
would give you something like minus e power kappa y. This is just a constant, right,
that needs to be multiplied here, the factor of 2 pi and something like that, okay.
So, other than that the concept is very simple. It is still the same thing as you know it is a
Gaussian here or rather what I would say is that the Fourier transform is also Gaussian
and then this would be the overall distribution and you can of course remember that fx is basically
x by lambda z and fy is equal to y by lambda z, okay. So, if you fix the screen at z equal to
constant then, oh yeah, we also have a phase factor e power minus jkz which I forgot to write,
but now I am going to write that one. So, yeah, this is the final expression
that you are looking for and as before one may try to understand this analytically, but I would say
the best way to understand these equations is to after you have looked at the tables and then found
out that this is essentially the same form as what you are going to get, then I would say that you
take, you know, you write a MATLAB script and plot this expression as a function of x and y.
Remember, you can write this as a function of x and y simply because for a given z,
fx and fy correspond to the spatial frequencies which are related in terms of x and y observation
points, right. So, I will leave this as a MATLAB exercise for you to figure out what would be the
beam size as you move away from this one, okay. However, in many cases, you are not just
interested in the far field, you are also interested in the near field, okay. And what should
we do about the near field thing? Well, we did some approximations to get to the far field or the
Fraunhofer diffraction. What we have to do is to now not do those approximations.
Remember what approximations we did? Yes, we neglected this x0 square plus y0 square term
in the numerator. So, there was something like divided by 2 s0, s0 we wrote it as z itself. So,
this was the term that you had and you also, so this is the term that we actually neglected,
right. So, you had this e power minus jk of course. So, this term we neglected, we said
that we could neglect this because x0 square plus y0 square maximum would be very, very small
compared to s0 and we replaced s0 by z because s0 and z were approximately of the same order,
right. So, this is the term that we actually neglected in the previous equation and therefore,
we had a Fraunhofer diffraction. Now, what we do is not neglect it, so which
means that we will go back to the initial distribution or the aperture distribution
which is given by this expression and then we write, so e power minus jkz would still be present,
some j by lambda z is also present and amplitude A is also present, but inside here what we will
have is the, you know the other term that we had, you know forgotten, right. So, that would be this
e power minus jk, we will have x minus x0 square plus y minus y0 square divided by 2z, okay.
So, this integration is what would give us the fields or you know the near field expression and
of course, you can get to the far field expression by neglecting x0 and y0 in relationship to z,
okay. But if you do not or if you are interested in knowing what happens to the beam,
you know growth initially, then you simply put this, you do not make this approximation,
retain this part, the quadratic phase factor part and then proceed to solve this integral,
okay. This integral is also not very difficult to solve, this can be done by completing the square,
okay. I will not go to the details here and as before you can split this integral in terms of
the integral over x and integral over y and after you do all that, what you would essentially end up
is something like this, okay, jA pi by lambda, do not worry about this part, you know this is
something that is going to come from the constant pi and this one will come from the Fourier
transform or the integral values of this one but what you would actually get which is interesting
is this quantity, okay. For a given z, this fellow will be constant
but as z increases, what you can see is that the amplitude here is decreasing, right.
So, you can of course take this amplitude itself and then take the magnitude of this one as well
as the phase of this. So, if you take the magnitude, you can see that it would be something
like say 4z square plus k square w0 square in the denominator under root, right. So,
the magnitude of this fellow would be this one and there would be a certain phase, this phase
is called as Gouy phase, okay and this will be important when you start doing this manipulating
the near fields of this Gaussian beam using lenses and other elements but we will not worry about
this Gouy phase, okay. We will simply look at this amplitude and then clearly you know that
psi magnitude square is the one that is going to give you intensity or the power.
So, if I take the magnitude square here, the rest of it will simply be phase factors which will go
away when you take the magnitude square but this amplitude part if you look at it, it is actually
like 1 over 4z square plus k square w0 square, correct. So, when z is very small, then the
beam base would be almost constant or the amplitude of this one would almost be constant
here k square w0 square. However, as z increases, the amplitude kind of decreases because this term
4z square starts to overtake everything, right. So, this quadratic decay, I mean for very large
values of z, the amplitude could be or the intensity essentially you know goes as 1 over z
square which of course is expected out of these type of beams, right. I mean these are the beams
that would have their power fall off as 1 by z square, the inverse relationship and when you
integrate over a certain sphere, then the power would essentially look at this, I mean this would
essentially be a constant. That is what we have seen, right.
But coming back to the phase part here, what you would see is e power minus jkz which of course
was present earlier also but the extra phase that you are going to get because of this quadratic
term x minus x0 square plus y minus y0 square is a very interesting term which is given by 2z 1
plus pi square w0 to the power 4 divided by lambda square z square, okay. So, this one is the
expression that you have for the phase part and of course this is what was actually you know
interesting for us because you see this is still Gaussian but then the effective w has actually
changed, right. Oh, sorry. This is not just the completeness, complete thing. This would be into
e power minus x square plus y square divided by 2 or rather 2 is not there because we did not start
with 2. So, this would be w square of z. So, this expression is not w square of z but this expression
is different. So, what we want to know is this w square of z which you can actually show that
it is given by w0 square 1 plus lambda square z square divided by pi square w0 4, right.
Yes, this is related to this but the basic motivation that the field is still Gaussian
comes from this particular expression. So, the field is Gaussian in the z equal to constant
plane away from every point, every plane away from the beam. The beam is essentially,
the Gaussian beam is essentially, the beam essentially retains its Gaussianity characteristic
but its spot size w will actually depend on z and starts to increase, okay.
So, at z equal to 0, this fellow will be equal to w0 and that is all right but you can also define
this entire thing, right. In this bracket, you can define this one by some other name or you
can denote this by some other name called zr square and then you have a very simple expression
for w square of z which is w0 square into 1 plus z by zr whole square.
So, clearly at z equal to zr, the beam size at that zr plane will be twice of w0 square or the
beam size will be square root 2 times w0 which was the original spot size times square root of 2
and this zr is called as Rayleigh distance or Rayleigh length or sometimes also called as
Rayleigh range, okay. So, this is how beams that are emitted by laser in the transverse plane
behave and this is a very interesting thing. We have seen beams in the slabs, we have seen
beams in the free space because of the antennas radiating the fields. We have seen these beams
or essentially electromagnetic waves or guided modes in waveguides, we have seen them in a free
space in the context of uniform plane waves. So, this one is another type of guided mode.
The guidance is now being the free space material and the initial conditions of this
beam is the laser. So, you can think of the laser as an antenna, okay, which emits a very special
class of beam called as a Gaussian beam and this beam is the one that is propagating in the free
space and because of diffraction, what happens is that the beam size may be small at some z
equal to 0 plane or we take this as z equal to 0 plane. This entire thing maybe we will
put in the black box and call it as a laser and at z equal to zr, what you would find is that
the beam size is square root 2 times the minimum value. The minimum value is called as the beam
waste and it is usually the convenient location where we put this beam up there, okay.
Now, coming back to this situation where we had a lens here and laser beam and then we wanted to
look at what happens to this, you know, light as it goes and hits the lens and then whether it gets
focused or it does not get focused, you can understand this propagation by going to what
is called as ABCD formalism, okay. ABCD formalism is, you know, a two-port type of a formalism
wherein you have certain parameters, which is essentially the slope as well as the distance
and then you have a matrix, which connects the slope and the height at a different plane, okay.
So, you can think of this as one plane, you can think of this as another plane, this is the input
plane and as this input, you know, at this input, you define two parameters for a Gaussian beam,
you define its waste as well as you define its divergence angle and then at another plane,
you have this, you know, relationship of, you know, waste or the spot size as well as the
divergence angle. In the free space, we know how these two are connected, this is w0 and
the divergence angle and this is w of z and the corresponding divergence angle, which you need to
take the derivative of that. However, when you put an optical element in between, what this optical
element does is to modify this w0 or, you know, the waste as well as the slope and how exactly it
gets modified is captured by this ABCD matrix, ABCD are its elements and this matrix will be
different for different elements, it will be different for a simple slab, it will be different
for a mirror, it will be different for a lens and so on. But the point is, if you are given those
matrices, okay, even if you do not understand the theory, if you are given those matrices,
right, and then you know the beam waste as well as the divergence angle to begin with,
you know, at the laser, then you can manipulate these matrices in order to get whatever the
beam size and the divergence that you want. So, let us say there is certain microscope
here, which accepts only a certain beam size and a certain divergence angle, whereas my laser is
incapable of giving that directly out, okay, the laser beam size and this one is not matching,
then you can make a match by putting these optical elements and then manipulating them
or, you know, moving these elements around and ensuring that the overall ABCD matrix is such that
the beam sizes on both sides of this optical system are matched, okay.
So, there is lot of things that one can talk about. Unfortunately, we will not be able to
talk about it anymore, because we want to move on to other interesting aspects in this course,
okay. So, one other interesting aspect in this context of diffraction is the close connection
between interference and diffraction, okay. Interference, as you know, is a phenomenon
wherein two waves, okay, they talk to each other and sometimes if they talk nicely,
then there will be constructive interference and when they talk in a manner that we will soon see,
there will be destructive interference. Constructive interference will always give
rise to an increase in the illumination, whereas destructive interference will destroy the
illumination at that position, okay. The most classic case of
interference or interferometer that you have, that interferometer is a device to observe
interference and the most classic device to observe interference is what is called as the
Young's double slit experiment. This is something that you can do at home and this is something that
has been done countlessly in many, many laboratories and in fact, it is being done every day, okay.
What I want to do is to give a touch which is slightly different, okay. Normally, this,
in the context of optical systems, you first encounter interference and then encounter
diffraction, but you can actually think of diffraction and integral, I mean, interference
in the other way around. In fact, you can use this Fraunhofer diffraction idea to actually talk
about interference itself. For that, let us carve two slits, okay. These slits are of equal size Wx,
you know, width and then we will assume that the width Wy is very, very small, okay.
The reason why we want to assume this Wy to be very small is that we know these rectangular
apertures are going to produce a diffraction pattern which would be sinc Fx Wx times sinc Fy
Wy, right. So, when I make Wy to be very small, then the sinc function would be approximately 1
and I do not need to worry about it, okay. So, these are very narrow slits as in that we are
considering. The only stipulation is that in this plane, these centers are separated by a distance
of 2x0, okay. So, or maybe we will have to use a different, this one.
So, we will separate them by say 2a, meaning that from this point which is x0 equal to 0
to this point will be a and there will be a similar a distance onto this one. So, the center
here is x equal to a or x0 equal to a, the center here is x0 equal to minus a. And then we are
looking at the field in a far away point P, okay, which is x, y and z and what we want to do is to
understand what would be the field at that point, okay. Now, the basic idea you already know,
you know what is the field of a single aperture when it is illuminated by a light beam and then
you also know this aperture. The only thing that is, you know, different in the previous cases to
this case is that you have 2 apertures. Of course, light is super, you know,
superposition principle holds, so you can add the 2 fields, but then you have to simply be
mindful of the fact of the phase difference that comes in because of the separation in the centers,
right. So, the phase shift from this one will be e power minus j 2 pi fx times a and the phase
shift because of this fellow will be e power j 2 pi fx a and the field essentially is the same.
So, when you add the fields outside, this would be sinc of fx wx. Please note that I have already
made this wy almost equal to 0 and this would be the overall field, okay. So, the field will be
cos 2 pi fx a and times sinc of fx wx and what we are going to do is to look at what happens
or what should be the relationship between a and wx in order for us to observe the classic
Young's double set interference in the next module. Thank you very much.
