Hyugens principle and laser wave dispersion

In summary: I'm not sure what you are asking. A spatial filter is a device that allows a beam of light to pass through a small opening. A common type of spatial filter is the lens, which allows a diverging beam to be made to converge by a simple transformation.
  • #1
Ulysees
516
0
According to wikipedia below:

> Huygen's principle states that each point of an advancing wave front is in fact the center of a fresh disturbance and the source of a new train of waves; and that the advancing wave as a whole may be regarded as the sum of all the secondary waves arising from points in the medium already traversed.

http://en.wikipedia.org/wiki/Photon_dynamics_in_the_double-slit_experiment

Then how is it possible that laser light moves in a narrow beam with very small dispersion sideways?

Can radiowaves be produced to be like that? Cause that would make an extremely high directivity antenna for satellite communications etc.
 
Last edited by a moderator:
Science news on Phys.org
  • #2
Laser beams disperse just like any other beam. Perhaps you are wondering how any well-collimated beam can exist, given Huygen's principle?
 
  • #3
Andy Resnick said:
Laser beams disperse just like any other beam.
Sure. It was mentioned:
very small dispersion sideways
Perhaps you are wondering how any well-collimated beam can exist, given Huygen's principle?

Exactly. In fact I'd like to simulate it with sound waves if possible. Maybe with a large number of speakers driven differently or something.
 
  • #4
A laser beam that is (say) 1 mm wide is still about ten thousand times wider than a single wavelegth which is on the order of [itex]10^{-7}[/itex] m.
 
  • #5
So what happens at the edge then? If we illuminate a camera chip directly with a laser beam, what does the edge look like?
 
  • #6
  • #7
Alright. It's hard to imagine why with all the smoothness at the edge it still maintains the direction.

Does the Hyugens principle apply to sound waves too? Can the same directionality be achieved with sound waves?
 
  • #8
FWIW, you are in good company, some of the finest scientific minds in history have written explanations of this. I have been trying to think of a simple way to explain this- here goes:

First, consider an abstract representation of the source- a hole in an opaque screen, located at z = 0. You have no idea what happens at z < 0, all you know is that at z = 0, the intensity is 1 in the hole and 0 everywhere else. If that's all there was, then indeed, you would expect the light to wildly diverge at z>0. But it doesn't- why?

Because light is a *vector* field rather than a scalar field. The field in the hole is not simply 1- it's got a direction associated with it. Intuitively, you can see that if the vectors all point along z, then the light should continue to go along z. If the vectors point 'out', then the divergence will be large (or conversely, can converge). If you like, it's possible to work in funny coordinates that take this into account- look up the "Rayeligh length" for a Gaussian beam.

The edge is a problem- the field is discontinuous, which implies that something diverged. Since infinities do not actually exist, what this means is that the field 'spreads out' in order to maintain a smooth function.

In practical terms, to make a perfectly collimated beam, the beam diameter must be infinitely large. Gaussian beams (beams from laser cavities) can be made well-collimated by passing them through a spatial filter (to make a pinhole source), and then to a lens. *But*, the best performance is when the pinhole clips the Gaussian near the zeros, not arbitrarily in the middle. This is to minimize the 'spreading' effect at the edge (which I now call diffraction).

Does this help?
 
  • #9
Gaussian beams (beams from laser cavities)

Is a Gaussian beam one with a Gaussian density function along a line perpendicular to its direction?

can be made well-collimated by passing them through a spatial filter (to make a pinhole source)

An example of a spatial filter that makes a pinhole source? Is a small hole through an opaque sheet an example? Is yes, wouldn't passing the beam through the hole make it diverge a lot?

and then to a lens

That makes sense. If a beam diverges 1 mm per metre, then a lens with a focal distance of 1 metre should make a 1 mm-wide beam perfectly parallel?

the best performance is when the pinhole clips the Gaussian near the zeros, not arbitrarily in the middle. This is to minimize the 'spreading' effect at the edge (which I now call diffraction)

I don't understand this bit. Do you mean that illuminating a pinhole with the very edge of the beam, would be ideal?
 
Last edited:
  • #10
Ulysees,

Let's go through point-by-point.1)Is a Gaussian beam one with a Gaussian density function along a line perpendicular to its direction?

Yes, the beam within an optical cavity behaves as Hermite-Gaussian (or Laguerre-Gaussian) functions for the transverse field density. A TEM00 beam has a Gaussian (~exp^-r^2) dependence.

2)An example of a spatial filter that makes a pinhole source? Is a small hole through an opaque sheet an example? Is yes, wouldn't passing the beam through the hole make it diverge a lot?

A proper spatial filter is a microscope objective focused to a pinhole. And yes, the beam can diverge a lot: the numerical aperture of the microscope objective sets the angle.

3) That makes sense. If a beam diverges 1 mm per metre, then a lens with a focal distance of 1 metre should make a 1 mm-wide beam perfectly parallel?

Well, usually the NA is closer to 0.5, so the divergence is much higher. But the raw beam coming out of a laser has a rough divergence of 1 mm/meter (whatever that is in radians). The 'collimated' beam will not be perfectly parallel in either case, because the source (pinhole or exit pupil) has a size. Using smaller pinholes will in the end give a better collimation.

4) I don't understand this bit. Do you mean that illuminating a pinhole with the very edge of the beam, would be ideal?

No, the pinhole needs to be correctly sized. It's always centered on the beam axis. But the size of the pinhole should be matched to the size of the focused beam- specifically, the size should be equal to the size of the central peak of the Airy disk.
 
  • #11
If we imagine a long, thin and straight tube made of light-absorbent material with a light source at one end, will the light coming out the other end be as collimated as the ratio diametre/length suggests?

I think not, but do you know the maths for it, any analytical solution how divergent the output beam would be?

If this works, it could be done with sound too, and make a sound laser.
 
Last edited:
  • #12
If this works, it could be done with sound too, and make a sound laser.

And by the way, no need to say laser is monochromatic or coherent, sound can have these properties easily.
 
  • #13
Ulysees said:
If we imagine a long, thin and straight tube made of light-absorbent material with a light source at one end, will the light coming out the other end be as collimated as the ratio diametre/length suggests?

I think not, but do you know the maths for it, any analytical solution how divergent the output beam would be?

If this works, it could be done with sound too, and make a sound laser.

Definitely not. The tube is not relevant- the only thing that matters for the propagation of light once it leaves the tube are the field values at the end face of the tube.

"light pencils" cannot exist. That is a fundamental conceptual error unfortunately aided by geometrical optics. A related conceptual error is the concept of "combustion from afar" (Archimedes legend). The two errors are:

1) parallel rays to not exist
2) even if they did, they could not carry energy.

If you can find it, an excellent read is "On the possible and impossible in optics", by G. G. Slyusarev (1962). It's a DTIC document, FTD-TT-62-175 (accession number 886286, I think).

I appreciate the idea of manipulating coherent sound- there's been some interesting work (odd overlap between military and advertising interests) with directional broadcast techniques:

http://www.temple.edu/ispr/examples/ex02_06_23.html
 
Last edited by a moderator:
  • #14
Andy Resnick said:
Definitely not. (...) "light pencils" cannot exist. That is a fundamental conceptual error unfortunately aided by geometrical optics.

Which is as was expected, but what are the equations here? I'd like to know the maths by which we derive the intensity profile below.

Image1.gif


http://img262.imageshack.us/img262/4459/image1ls6.gif

The flat part I can do from geometric optics: it has a width given by:

[tex]w = D(L+d)/L[/tex]

How do we derive the part due to diffusion, above and below the flat part? I think it can only be done numerically with a method like TLM. But do you know an analytical solution? Maybe something based on Hyugens principle, so we get a chance to understand how the principle accounts for collimated waves?
 
Last edited by a moderator:
  • #15
Ulysees said:
How do we derive the part due to diffusion, above and below the flat part? I think it can only be done numerically with a method like TLM. But do you know an analytical solution? Maybe something based on Hyugens principle, so we get a chance to understand how the principle accounts for collimated waves?
Firstly, I think you mean diffraction (not diffusion).

It is possible to get a solution analytically by solving the scalar wave equation and using the paraxial approximation [itex]d^2E/dz^2 = 0[/itex]. You do however need to make some assumptions about the solution (that it is beam-like for example). Note too that since the solution is a scalar function, it does not take into account vector effects such as polarisation.

A search on "Gaussian Beams" will yield plenty more information.

Claude.
 
  • #16
Ulysees said:
Which is as was expected, but what are the equations here? I'd like to know the maths by which we derive the intensity profile below.

<snip>

How do we derive the part due to diffusion, above and below the flat part? I think it can only be done numerically with a method like TLM. But do you know an analytical solution? Maybe something based on Hyugens principle, so we get a chance to understand how the principle accounts for collimated waves?

I don't understand the picture- there's a point source within a long tube, the interior walls of which are covered by absorbent material?

Usually, when an incoherent source is used, people work in terms of the intensity instead of the field. In this case, sometimes people speak of the 'diffusion' of energy (intensity) rather than diffraction (field). They are similar concepts, but lead to very different and surprising results: for example, using incoherent light when imaging can provide twice the resolution over using coherent illumination.
 
  • #17
Andy Resnick said:
I don't understand the picture- there's a point source within a long tube, the interior walls of which are covered by absorbent material?

Exactly. The curved edges of the beam that are labeled "diffusion" are not rays of light, they are just the edges beyond which the intensity is below a threshold.

Claude, I definitely didn't mean diffraction, diffraction requires different materials, eg glass and air, or a varying diffraction index causing the rays to bend. "Diffusion" is not accurate either, because it is normally used for the scattering due to air molecules. I want to understand the spreading due to the wave nature of light, in a perfect vacuum. What is that spreading called and how is it calculated?

Claude Bile said:
It is possible to get a solution analytically by solving the scalar wave equation and using the paraxial approximation [itex]d^2E/dz^2 = 0[/itex]. You do however need to make some assumptions about the solution (that it is beam-like for example).

Thanks. But what is that paraxial approximation? Looks like assuming the E field varies linearly along the axis of the beam:

[itex]E = az + b[/itex]

at x=0, y=0

What equation would you be solving analytically under this assumption? Maxwell's?
 
Last edited:
  • #18
I think you do mean diffraction (it sounds like you are confusing diffraction with refraction?).

With regard to the paraxial wave approximation (I made a slight error in my previous post by the way...);

First, if you assume a monochromatic wave, the wave equation (which can be derived from Maxwell's equations) reduces to the Helmholtz equation;

[tex]\nabla^2E+k^2E=0[/tex]

Now define an envelope function, U such that;

[tex]E(r) = U(r)e^{ikz}[/tex]

Substituting this into the Helmholtz equation yields;

[tex]\frac{d^2U}{dx^2}e^{ikz} + \frac{d^2U}{dy^2}e^{ikz} + \frac{d^2U}{dz^2}e^{ikz} + 2ik\frac{dU}{dz}e^{ikz}=0[/tex]

The paraxial approximation is that the [itex]d^2U/dz^2[/itex] term reduces to 0 (not the [itex]d^2E/dz^2[/itex] term as I mistakenly said in my previous post). This approximation is, in essence, saying that since the solution is beam-like, the envelope is slowly varying with z.

This yields the Paraxial Wave Equation;

[tex]\frac{d^2U}{dx^2}e^{ikz} + \frac{d^2U}{dy^2}e^{ikz} + 2ik\frac{dU}{dz}e^{ikz}=0[/tex]

The Gaussian Beam is a solution to this equation.

EDIT: The derivatives should be partial derivatives.

Claude.
 
  • #19
That's what I was looking for, thank you.

So if instead you do not assume the envelope U(r) varies linearly along z, you get something close to a Gaussian U(r) which can only be derived numerically.

It seems strange to me that the paraxial approximation is close to the real thing. Isn't the approximation saying that the envelope U(r) does NOT approach zero at z=infinity asymptotically, but linearly, like below?

U(r)=(az+b)G(r),

G(r)=Gaussian

Do actual light beams look more like an inverse square function of z as we approach infinity?
 
Last edited:
  • #20
Well, the paraxial approximation is essentially saying that a beam diverges at a fixed angle for large z, which is a fairly accurate approximation.

If you assume a solution of the form (noting that r = distance from the beam axis - [itex]x^2 + y^2[/itex] not the position vector);

[tex]U(x,y,z) = E_0e^{iQ(z)r^2}e^{iP(z)}[/tex]

The Gaussian solution can be derived analytically. The actual solution is of the form;

[tex]U(x,y,z) = E_0\frac{w_0}{w(z)}e^{-itan(z/z_0)}e^{-r^2/w^2(z)}e^{ikr^2/2R(z)}[/tex]

Where;

[itex]E_0[/itex] is the E-field amplitude.
[itex]w_0[/itex] is the spot radius at the beam waist.
[itex]w(z)[/itex] is the spot radius as a function of distance.
[itex]z_0[/itex] is the Rayleigh Range (or the Confocal parameter) of the beam.
[itex]R(z)[/itex] is the Radius of Curvature of the beam as a function of distance, where;

[tex] R(z) = \frac{1}{z}(z^2+z_0^2)[/tex]

The 1st exponent is the Guoy (longitudinal) Phase term, the 2nd exponent is the transverse beam amplitude and the 3rd exponent is the transverse phase term. Also, note that there is some curvature in the beam profile, however it is very small, which is why the [itex]d^2U/dz^2[/itex] term is assumed to be negligible.

Claude.
 
  • #21
Thank you.

Is there a theoretical maximum on the directivity of an e/m wave? Perhaps measured as the angle for a 3 dB fall in power relative to the centre of the beam, at 1 metre?
 
  • #22
Not sure what you mean by directivity. Do you mean the angle of divergence?

For Gaussian Beams, the angle of divergence increases as the beam waist decreases. The beam waist radius cannot be smaller than [itex]\lambda/4[/itex] due to the diffraction limit, which will, in turn, limit the angle of divergence.

For EM waves in general though, there is no theoretical limit.

Claude.
 
  • #23
Claude Bile said:
Not sure what you mean by directivity. Do you mean the angle of divergence?

Yes, but only at a given distance from the source (because, as you said too, the angle increases with distance - the edge is curvy).

And because the profile U(r) is typically similar to Gaussian (ie asymptotically extends to infinity sideways), we need a threshold for the intensity to define an edge. Eg 3 dB below the maximum value of U(r) as I said.

The beam waist radius cannot be smaller than [itex]\lambda/4[/itex] due to the diffraction limit

How do you define the beam waist? There must be a threshold implied since U(r) extends to infinity sideways, but what is the chosen threshold?

For EM waves in general though, there is no theoretical limit.

Don't the same rules apply as with light? Don't EM waves come from sources of finite size, just like light?
 
  • #24
Ulysees said:
Yes, but only at a given distance from the source (because, as you said too, the angle increases with distance - the edge is curvy).
I see, however the divergence angle does approach a fixed value for large z - which is why the far-field divergence is typically used to characterise the spread of a beam.

Ulysees said:
And because the profile U(r) is typically similar to Gaussian (ie asymptotically extends to infinity sideways), we need a threshold for the intensity to define an edge. Eg 3 dB below the maximum value of U(r) as I said.
Yes, the 3 dB point is typically defined as the "edge" of a Gaussian beam. If you know the beam width at a certain distance from the beam waist, calculating the angle of divergence is trivial.

Ulysees said:
How do you define the beam waist? There must be a threshold implied since U(r) extends to infinity sideways, but what is the chosen threshold?
I'm confused by this, the "r" in U(r) is the radial distance from the beam axis. Hence the beam is confined in both x and y. The beam waist is defined as the point where the beam width is a minimum.

Ulysees said:
Don't the same rules apply as with light? Don't EM waves come from sources of finite size, just like light?
Not all light travels as a Gaussian Beam, isotropic sources (such as fluorescent sources for example) emit spherical waves that spread equally in all directions.

Claude.
 
  • #25
Thank you.

I think you forgot something I was asking.

I'd like to know if radiowave EM beams can be made as directional as a light beam.

Does light make more directional beams than microwaves?

Also, is laser the most directional beam we can send?
 
  • #26
Ulysees said:
I'd like to know if radiowave EM beams can be made as directional as a light beam.
In principle, yes.
Ulysees said:
Does light make more directional beams than microwaves?
When a wave hits an obstacle or aperture, the degree to which the beam spreads depends on [itex]\lambda/a[/itex] where [itex]\lambda[/itex] is the wavelength and [itex]a[/itex] is the size of the obstacle or aperture.

In other words, a 100 m wavelength beam through a 100 m aperture will spread the same amount as a 500 nm wavelength beam through a 500 nm aperture.

So for this reason, it is far easier to keep an optical wave directional than a radio wave in most circumstances.

Ulysees said:
Also, is laser the most directional beam we can send?
No, there are several types of waves that are more directional than a standard laser beam.

Bessel beams (whose transverse cross section is a Bessel function - just as Gaussian beams have a Gaussian transverse cross section) are somewhat unique in that they do not spread out as they propagate.

Spatial solitons are another form of waves that do not spread out as they propagate, due to some form of nonlinear response in the medium they propagate through. You can't get a spatial soliton in a vacuum like you can with Bessel beams.

Finally, there are laser filaments. Laser filaments arise when you have very high intensity pulsed lasers - the intensity is so great that you get a combination of nonlinear Kerr self-focusing (which is the phenomenon whereby a medium possesses an intensity dependent refractive index, which causes regions of high intensity to act like a converging lens, hence the term self-focusing) and defocusing caused by plasma generation through ionisation of the surrounding medium. The result is that the laser beam remains tightly confined over long distances. Laser filaments are particularly useful for LIDAR (light radar) applications.

Claude.
 
  • #27
Thank you very much, this post is very interesting.
 
  • #28
jtbell said:
This Wikipedia article shows a couple of examples of laser beam intensity profiles.

http://sensoria.in/TU/knowledge-pool/34-electronics-and-communication/78-laser-principles.html"
 
Last edited by a moderator:

1. What is Hyugens principle?

Hyugens principle is a concept in physics that states that every point on a wavefront can be considered as a new source of spherical wavelets. These wavelets combine to form a new wavefront, and this process continues as the wave spreads through a medium.

2. How is Hyugens principle related to laser wave dispersion?

Hyugens principle is crucial in understanding laser wave dispersion. As a laser beam travels through a medium, it encounters different materials with varying refractive indices. According to Hyugens principle, each point on the wavefront can be considered as a new source, and these sources experience different speeds of propagation due to the varying refractive indices. This results in the dispersion of the laser beam.

3. What factors affect the degree of laser wave dispersion?

The degree of laser wave dispersion is affected by several factors, including the wavelength of the laser beam, the refractive index of the medium it is passing through, and the angle at which the beam enters the medium. The higher the refractive index and the larger the angle of entry, the greater the degree of dispersion.

4. How does laser wave dispersion impact the quality of a laser beam?

Laser wave dispersion can significantly impact the quality of a laser beam. As the beam spreads out due to dispersion, it can lose its coherence and become less focused, resulting in a decrease in its power and precision. This can be a significant concern in applications such as laser cutting and medical procedures.

5. Can laser wave dispersion be reduced or eliminated?

Yes, laser wave dispersion can be reduced by using materials with lower refractive indices or by using techniques such as beam shaping and adaptive optics. In some cases, it may be possible to eliminate dispersion entirely, but this depends on the specific setup and conditions of the laser beam.

Similar threads

Replies
2
Views
4K
Replies
6
Views
9K
  • Quantum Interpretations and Foundations
Replies
1
Views
2K
Replies
1
Views
3K
  • Other Physics Topics
Replies
2
Views
3K
Back
Top