# What happens when light hits light?

• junguo93
In summary, according to wave theory, photons interfere and cancel their amplitudes, double in number, and change their trajectory after the interaction.
junguo93
As titled, let's say Source A is beaming light at Source B and Source B is doing the same to Source A, what is actually happening with the protons?

My guess is that light photon collides and slows down o.o? But that is based on a wild guess, can someone come up with another theory/explanation? (:

Photons do not bounce off each other. However, according to the wave description, they could interfere.The result of this would depend, according to the wave theory, on their difference in phase. However, things aren't as simple as this implies because interference is known to take place even with single photons. If one imagines two photons "colliding" in the way you describe, and further imagines similar "collisions" many times, the outcomes would not be expected to be always the same. However, I think the average behaviour (taken over many such collisions) is the same as is predicted by wave theory.

Note that the ''interference" is the same as if two water waves crossed. At the time of crossing, two "highs" and two "lows" sum while a "high" and a "low" cancel. After they have passed through one another, they will not have changed.

If,say, two automobiles headlights are pointed at each other. what happens?? Each is illuminated, right?

You know light is an electromagnetic wave, right? So its behavior is characterized by Maxwell's equations.

If you mean monochromatic light sources, like that produced in theory via a laser, I'm not sure exactly what happens...
in theory they can destructively interfere...but I do know that all electromagnetic waves travel at c in a vacuum. Photons don't slow down. You might also achieve some unique results in fiber optic cables between a pair of 'identical' light sources.

In general as noted already, wave interference via superposition should describe the interactions...see the diagram here for an illustration of constructive and destructive wave interference:

http://en.wikipedia.org/wiki/Interference_(wave_propagation )

and check the additional diagrams for different interference patterns.

Last edited by a moderator:
HallsofIvy said:
Note that the ''interference" is the same as if two water waves crossed. At the time of crossing, two "highs" and two "lows" sum while a "high" and a "low" cancel. After they have passed through one another, they will not have changed.

What exactly does it mean for two photons to interfere and cancel their amplitudes, or to double it? For example, if we place a detector right where they interfere, would they leave more bright or less bright dots, or something?

I always thought they actually do get changed after the interaction as well, at least their trajectory seem to change in double-slit experiments.

Br.S said:
What exactly does it mean for two photons to interfere and cancel their amplitudes, or to double it? For example, if we place a detector right where they interfere, would they leave more bright or less bright dots, or something?

I always thought they actually do get changed after the interaction as well, at least their trajectory seem to change in double-slit experiments.

Well for starters, the property of the interference is going to be part of the overall description of the detection of the photons, not with the individual photons themselves. I just saw a discussion the other day where a person was asking how the phase of one photon interferes with the phase of another photon. This is incorrect in thinking that we ascribe a phase to each photon. What we do is we solve for a wave function that has a phase associated with it and it is the wave function that experiences interference. Physically, this wave function can be thought of as the probability of detecting a photon in a given volume. The thing is, this interference is independent of the number of photons present at a given point in time. We can emit one photon at a time and still get the exact same interference pattern over time as if we emitted trillions at a time. So there is no difference in trying to solve the problem for one, two, or more photons (though this isn't true for every kind of particle). There's more to it than this but I want to try and keep this brief.

To answer your question though, since the wave function represents the probability density, then we find that the interference is associated with regions of greater or lesser intensity in the rate of photons being detected depending upon the kind of interference. In other words, the interference describes regions of darkness or brightness.

I think it's just easier to think of this in the classical model, where light is simply a wave. Now the electromagnetic field that describes light has a phase associated with it. So when two waves combine, the relative phase difference between the waves in time and space gives rise to the interference.

I always thought they actually do get changed after the interaction as well, at least their trajectory seem to change in double-slit experiments.

There IS a change...individual particles interfere with themselves because the method of detection [ slit] affects the outcome. Particles have a wave nature and even specific frequencies. If there is no image at a point, the quantum mechanical view is that the probability of a photon's appearing at that point is very low. We can't predict where the next particle image will appear...

But in open space, say when you turn on a light bulb, each emergent photon [wave packets of energy] follows a path [a geodesic] and none 'disappear'.

Br.S said:
What exactly does it mean for two photons to interfere and cancel their amplitudes, or to double it? For example, if we place a detector right where they interfere, would they leave more bright or less bright dots, or something?

I always thought they actually do get changed after the interaction as well, at least their trajectory seem to change in double-slit experiments.

Two photons cannot interfere. Photons in two light beams are Bosons and, as such, can all occupy exactly the same space location and Energy. If they are not actually interacting with matter in some way (passing through a transparent non-linear medium) then they have no way of affecting what the other does.
The effect of interference is essentially how the statistics of a single photon passing through a system give a probability distribution of where it will be found.

Born2bwire said:
Well for starters, the property of the interference is going to be part of the overall description of the detection of the photons, not with the individual photons themselves. I just saw a discussion the other day where a person was asking how the phase of one photon interferes with the phase of another photon. This is incorrect in thinking that we ascribe a phase to each photon. What we do is we solve for a wave function that has a phase associated with it and it is the wave function that experiences interference. Physically, this wave function can be thought of as the probability of detecting a photon in a given volume. The thing is, this interference is independent of the number of photons present at a given point in time. We can emit one photon at a time and still get the exact same interference pattern over time as if we emitted trillions at a time. So there is no difference in trying to solve the problem for one, two, or more photons (though this isn't true for every kind of particle). There's more to it than this but I want to try and keep this brief.

To answer your question though, since the wave function represents the probability density, then we find that the interference is associated with regions of greater or lesser intensity in the rate of photons being detected depending upon the kind of interference. In other words, the interference describes regions of darkness or brightness.

I think it's just easier to think of this in the classical model, where light is simply a wave. Now the electromagnetic field that describes light has a phase associated with it. So when two waves combine, the relative phase difference between the waves in time and space gives rise to the interference.

The wave function can't be just an abstract description, those waves ought to actually be there, in whatever physical form, but one way or another they must be describing some real physical property. Wave-like interference is actual and real after all, so the question is then more along the lines whether this "wave" property is contained individually in each photon, or it emerges from some combination of their group existence, motion, or whatever else.

Take vertically polarized light for example, double slit wave-like interference is then supposed to impact photons only over their vertical oscillation frequency, their amplitude in the vertical plane should either increase or decrease, yet the effect we measure is that their trajectory would be bent horizontally, forming spacing between the pattern fringes. I think, maybe not, maybe vertically polarized light does not actually produce the pattern with horizontally spaced double slit?

sophiecentaur said:
Two photons cannot interfere. Photons in two light beams are Bosons and, as such, can all occupy exactly the same space location and Energy. If they are not actually interacting with matter in some way (passing through a transparent non-linear medium) then they have no way of affecting what the other does.
The effect of interference is essentially how the statistics of a single photon passing through a system give a probability distribution of where it will be found.

I'm pretty sure there are quite a few experiments demonstrating two beams from independent light sources interfere and can produce the pattern. Do you think then it's not two photons that interfere with each other, but something else is going on there, or perhaps you doubt in such experiments?

Br.S said:
Take vertically polarized light for example, double slit wave-like interference is then supposed to impact photons only over their vertical oscillation frequency, their amplitude in the vertical plane should either increase or decrease, yet the effect we measure is that their trajectory would be bent horizontally, forming spacing between the pattern fringes. I think, maybe not, maybe vertically polarized light does not actually produce the pattern with horizontally spaced double slit?

I don't think polarization affects the interference pattern, only the shape/size of the slits and the wavelength of the light.

Drakkith said:
I don't think polarization affects the interference pattern, only the shape/size of the slits and the wavelength of the light.

I think polarization must have some impact.

These guys say it influences visibility and some angles in the pattern. They also say this: "importance of the polarization characteristics is that only parallel oriented polarization modes interfere...", whatever is that supposed to mean.

Br.S said:
I think polarization must have some impact.

These guys say it influences visibility and some angles in the pattern. They also say this: "importance of the polarization characteristics is that only parallel oriented polarization modes interfere...", whatever is that supposed to mean.

Perhaps. I was thinking only about the double slit experiment itself, not about phase masks, which may be different. Maybe someone else here will know.

Br.S said:
I'm pretty sure there are quite a few experiments demonstrating two beams from independent light sources interfere and can produce the pattern. Do you think then it's not two photons that interfere with each other, but something else is going on there, or perhaps you doubt in such experiments?

This is the problem when people try to think about in terms of photons. Quantum theory states that photons do not interfere with each other because they are bosons. This is why the interference pattern is independent of the number of photons. It has been shown experimentally that you still get the same interference pattern if you only use a single photon source.

Br.S said:
The wave function can't be just an abstract description, those waves ought to actually be there, in whatever physical form, but one way or another they must be describing some real physical property. Wave-like interference is actual and real after all, so the question is then more along the lines whether this "wave" property is contained individually in each photon, or it emerges from some combination of their group existence, motion, or whatever else.

Take vertically polarized light for example, double slit wave-like interference is then supposed to impact photons only over their vertical oscillation frequency, their amplitude in the vertical plane should either increase or decrease, yet the effect we measure is that their trajectory would be bent horizontally, forming spacing between the pattern fringes. I think, maybe not, maybe vertically polarized light does not actually produce the pattern with horizontally spaced double slit?

The wave function is not physical in that it is not observable. The wave function does however give the physical results when we use it with an appropriate operator. By performing the desired operation on the wave function, we can find the physical properties of the system like energy or the probability density of where the particle would be detected. The phase is not assigned to the photons, it is assigned to the wave function that describes the specific state of the photon system that we are looking at. We could have multiple states combine and interfere with each other in that way, but we do not have photons coming together to interfere with each other.

Br.S said:
I think polarization must have some impact.

These guys say it influences visibility and some angles in the pattern. They also say this: "importance of the polarization characteristics is that only parallel oriented polarization modes interfere...", whatever is that supposed to mean.

I also agree with Drakkith. Double slit interference is not affected by polarization. What you have linked to in your post is about gratings, not double slit. These gratings are made using dielectric fibers, not a perfect conductor as we assume in the double slit experiment.

Br.S said:
I'm pretty sure there are quite a few experiments demonstrating two beams from independent light sources interfere and can produce the pattern. Do you think then it's not two photons that interfere with each other, but something else is going on there, or perhaps you doubt in such experiments?

One must be a bit careful when discussing that. Two-photon interference is real, but it is a bit different than your typical single-photon interference which is tested by double slits or Mach-Zehnder interferometers. The simplest version of two-photon interference is the Hong-Ou-Mandel effect or maybe photon bunching/bosonic final state stimulation.

If you use different sources to create two-photon-interference, it is necessary that you make these two modes indistinguishable, so they have exactly the same spectral, spatial and temporal properties (to be more exact, it is enough to make emission and detection events caused by these two light fields indistinguishable, but that is nitpicking). In other words, you rather have more complicated interference properties for modes which are occupied by more than one photon, rather than two individual photons interfering.

Br.S said:
I think polarization must have some impact.

Well, if you put polarizers at the positions of the slits, it can of course have an impact as you can create distinguishable fields which do not interfere. But the polarization you start with, does not matter unless you use slits which are larger than the wavelength of the light used.

Cthugha said:
One must be a bit careful when discussing that. Two-photon interference is real, but it is a bit different than your typical single-photon interference which is tested by double slits or Mach-Zehnder interferometers. The simplest version of two-photon interference is the Hong-Ou-Mandel effect or maybe photon bunching/bosonic final state stimulation.

If you use different sources to create two-photon-interference, it is necessary that you make these two modes indistinguishable, so they have exactly the same spectral, spatial and temporal properties (to be more exact, it is enough to make emission and detection events caused by these two light fields indistinguishable, but that is nitpicking). In other words, you rather have more complicated interference properties for modes which are occupied by more than one photon, rather than two individual photons interfering.

Well, if you put polarizers at the positions of the slits, it can of course have an impact as you can create distinguishable fields which do not interfere. But the polarization you start with, does not matter unless you use slits which are larger than the wavelength of the light used.

I don't see what to be careful about, two independent light beams can either interfere or not. Sure, no slit is different thing than two slits, but it's supposed to be the same mechanics relating to their wave property and the principle of wave superposition. Right?

So anyway, are you saying two independent light beams can not interfere unless they are both polarized perpendicularly the same plane, for example both vertically polarized? Or can two independent light beams interfere where one is vertically and the other horizontally polarized?

Interference does not affect light propagation. It's just superposition in action. You literally add the electromagnetic field of one beam with electromagnetic field of another.

You don't need to drag up any quantum mechanics here. So long as we are not looking at entanglement, electromagnetic field IS the wave-function of photons. This, by the way, points out the fact that while wave-function need not be observable, it does not mean it has to be non-observable. For a single photon, its wave function is a measurable field. The electromagnetic field.

Br.S said:
I don't see what to be careful about, two independent light beams can either interfere or not. Sure, no slit is different thing than two slits, but it's supposed to be the same mechanics relating to their wave property and the principle of wave superposition. Right?

No, two-photon interference is exactly not that. It is very different. This is why you need to be careful. An interesting discussion on TPI can be founs in "Can Two-Photon Interference be Considered the Interference of Two Photons?", Phys. Rev. Lett. 77, 1917–1920 (1996). You can also find free copies of it all over the web.

In a nutshell, TPI arises not due to superposition of waves, but due to superposition of indistinguishable probability amplitudes associated with the simultaneous detection of two photons. I do not know your background, so it is somewhat hard to tell, whether it is sensible to go into details or not. If you are not interested in very special stuff, that does not occur in eveyday life, "different light sources do not interfere" is rather correct.

Br.S said:
So anyway, are you saying two independent light beams can not interfere unless they are both polarized perpendicularly the same plane, for example both vertically polarized? Or can two independent light beams interfere where one is vertically and the other horizontally polarized?

Typically you need indistinguishable photons for two-photon interference. That also means you need the same polarization. Interference terms will also cancel out on average, if you do not have a fixed phase relationship between the two fields of interest. To maintain a fixed phase relationship over a longer timescale, you need the fields to be as similar as possible. Just taking two arbitrary light beams will therefore not create interference. Two-photon interference is therefore a rather rare thing happening only under lab conditions.

it depends on when one has read that two photon sources can not interfere.It is mentioned in older books that two distinct photon sources can not interfere because of random phase relationship.But this statement is quite wrong now because it is possible to make sources which have constant phase relationship over a long time.

andrien said:
it depends on when one has read that two photon sources can not interfere.It is mentioned in older books that two distinct photon sources can not interfere because of random phase relationship.But this statement is quite wrong now because it is possible to make sources which have constant phase relationship over a long time.

I think you are making assumptions about the nature of the photon. Just because two sources are in phase doesn't imply that photons have to come from both sources 'at once' to produce interference. We are quite happy to bandy about the notion of virtual photons when talking about charges interacting so why not regard the interference effect as being virtual too? Whenever I come across discussions which involve 'shoehorning' particles into explanations of phenomena that are essentially wavelike, it seems that the explanations get more and more convoluted in order to accommodate them. I know there is overwhelming evidence of the Quantisation of EM but, beyond that, the actual nature of photons is pretty unclear. Calling them particles doesn't actually seem to help - I know that Feynman insisted that they are but I never heard of him justifying that statement where RF is concerned. His famous diagrams were never intended as pictures of an actual event, I'm sure; rather, they were two dimensional functional diagrams. I thought that the whole point of QM is that it's not actually 'like' anything else. I'll bet that it is only in the minds of a very few, really well informed Physicists that, when the word 'photon' comes up, the little bullet model doesn't lurk somewhere in there.

It has been possible since the invention of the radio valve! Like I said before, people seem to regard the Laser phenomenon as being something special and needing special terms to explain it with. The equivalent has been around at lower frequencies for ages and needs, just as much, to be explained in common terms, surely.

The fact that photons with high enough energies can interact to produce massive particles is an interesting facet but a bit difficult to reconcile with a 'linear space'. It seems that there are no particles with low enough mass for visible light photons to create. Does this, somehow imply a limit to the linearity of space as a medium? A sort of gear change, which occurs when the energies correspond to a lepton (or possiby a quark?) mass.

well,the statement I have just written comes from the book of feynman lectures itself.I don't know what you mean by linear space here.Some vector space,minkowski space,affine space?

andrien said:
well,the statement I have just written comes from the book of feynman lectures itself.I don't know what you mean by linear space here.Some vector space,minkowski space,affine space?

I guess I mean minkowski space, in as far as SR doesn't imply quantum effects. I am talking in terms of Maxwellian (?) space, in which E = E1 + E2 for all values of the Es.

The point I am making is that two optical quanta will not (afaik) mutually annihilate to produce a particle but they will have no effect on each other in the same way that two AC signals on a linear transmission line will not.

Of course, people quote Feynman all over the place because his lectures, interviews and books are highly respected. I am not arguing against what I know of what he's said but did he ever discuss my point publicly? I do know that his 'diagrams' were not intended to be taken literally, any more than Equations are intended to be a physical representation. I have heard him assert that photons 'are' particles but do we all mean the same thing as he did when he used the word particle? Is there any record of what he had to say about low frequency photons, for instance?

Has nothing, since his time, ever successfully challenged his ideas? I am not challenging them (far too humble) but I do wonder, sometimes, whether people follow more what they thought he said rather than what he meant. There is so much 'interpretation' involved in this subject (like the Koran and Bible??) and I don't think my question actually goes against the Copenhagen Interpretation.

Is there any record of what he had to say about low frequency photons, for instance?
yes,I know what he has said.He said to give photon a mass in low energy limit to avoid infrared divergence and then combining it with cut-off provided by bethe in non-relativistic limit gives a fully divergenceless result i.e. no infrared divergence!and take mass zero in the end.mass is invariant.

andrien said:
yes,I know what he has said.He said to give photon a mass in low energy limit to avoid infrared divergence and then combining it with cut-off provided by bethe in non-relativistic limit gives a fully divergenceless result i.e. no infrared divergence!and take mass zero in the end.mass is invariant.

I don't really understand that except that it confirms that photons do not exhibit mass. (I'll take "infra red" to include LF radio too.) When I asked what he said about low frequency photons, I was wondering about the situation with interference from two independent synchronous, RF sources. This is such an easily produced phenomenon at RF and not that difficult with two phase locked lasers, apparently.
If, indeed, it were really necessary for actual photons from each RF source to interfere with each other in order to produce an interference pattern then surely we would expect an awful lot of photons from each source not to interfere ('cancel' in one direction) - producing a very diluted interference pattern. In fact, very deep, stable nulls (0.1% amplitude) can be formed from two independent RF sources. If the sources are actually not phase locked at all but just very stable and slowly drifting past each other, a very clean interference pattern can result. (All that is necessary is that the amplitudes are equal, of course)
So the choice is between saying that, magically, pairs of matched photons are always present in the two 'beams' so that they can combine into a perfect pattern or that the interference is nothing to do with photons at all and the photons only turn up when they're actually detected (which is in line with the Copenhagen interpretation, I think).

I wish someone could comment on this and find the flaw - if there is one.

Cthugha said:
No, two-photon interference is exactly not that. It is very different. This is why you need to be careful. An interesting discussion on TPI can be founs in "Can Two-Photon Interference be Considered the Interference of Two Photons?", Phys. Rev. Lett. 77, 1917–1920 (1996). You can also find free copies of it all over the web.

In a nutshell, TPI arises not due to superposition of waves, but due to superposition of indistinguishable probability amplitudes associated with the simultaneous detection of two photons. I do not know your background, so it is somewhat hard to tell, whether it is sensible to go into details or not. If you are not interested in very special stuff, that does not occur in eveyday life, "different light sources do not interfere" is rather correct.

Typically you need indistinguishable photons for two-photon interference. That also means you need the same polarization. Interference terms will also cancel out on average, if you do not have a fixed phase relationship between the two fields of interest. To maintain a fixed phase relationship over a longer timescale, you need the fields to be as similar as possible. Just taking two arbitrary light beams will therefore not create interference. Two-photon interference is therefore a rather rare thing happening only under lab conditions.

Is there a real difference between these two things? Is it not just two ways of saying the same thing?
Using RF sources tends to take care of the polarisation issue.
I looked for that article but could only find sources that charge for it.

you can not count about every photon,in most precise definition a photon interferes with itself.All those classical ideas fails,there i no intuition.

andrien said:
you can not count about every photon,in most precise definition a photon interferes with itself.All those classical ideas fails,there i no intuition.

You can count individual gamma photons by the clicks they give on a GM tube.
You are right enough about intuition failing. I think this is the general problem that people have when talking of photons. When Feynman asserted that they were particles, I think he did no one any favours because the particle word that he used was not the particle that comes to most people's minds. There are so many apparent paradoxes involved when we compare what happens with photons of different energies and those paradoxes should be taken as a strong message that photons are nothing like most people think.

I don't think it helps that most specialists on photons seem to be concerned with optical photons which are only a small sub-set of the beasts.

Interesting thoughts there. Thanks i learned quite a bit

From my experiments as a teenager installing aircraft landing lights in our automobiles-
superposition makes the best sense.
When you place two of them face to face and energize both , one of the filaments soon melts.

Ohh, nostalgia!

sophiecentaur said:
You can count individual gamma photons by the clicks they give on a GM tube.
I meant there that you can not follow a single photon's path.Photon is a result of quantization of electromagnetic field.It is just a quantum of EM field.In large occupation number limit,you can treat photons as light.Photons and light are same thing.

From a practical standpoint, I don't get the obsession with throwing the photon concept into situations which are perfectly adequately described classically. Unless the light the OP is beaming is such high frequency that you get photon-photon interaction or such low amplitude that you get single quanta, then just use Maxwell's equations and superposition.

andrien said:
I meant there that you can not follow a single photon's path.Photon is a result of quantization of electromagnetic field.It is just a quantum of EM field.In large occupation number limit,you can treat photons as light.Photons and light are same thing.

That's right, more than that starts to become a strreeeettch in thinking. The only actual evidence for photons is when they are are formed or detected. What goes on in between is a total mystery. To describe the nature of a photon whist energy is being transferred (in the wave) is, to my mind, a bit glib. And I think this applies however low the flux happens to be.

At the high frequency end, where photons interact to produce matter, the situation can still obtain. I would like to know just what is the minimum frequency for this to happen, though, and what particle is involved. It seems here must be a major change in the Physics of EM at that point. Is there some kind of breakdown in the way 'space works' then or could it be looked upon as some sort of minimum quantum EM energy for a change of 'mass state'?

sophiecentaur said:
At the high frequency end, where photons interact to produce matter, the situation can still obtain. I would like to know just what is the minimum frequency for this to happen, though, and what particle is involved. It seems here must be a major change in the Physics of EM at that point. Is there some kind of breakdown in the way 'space works' then or could it be looked upon as some sort of minimum quantum EM energy for a change of 'mass state'?
Two oppositely-travelling 511 keV photons could interact to produce an electron-positron pair. If the photons were just barely 511 keV then the resulting electron and positron would have very little KE and so they would attract each other, anhilate, and produce two 511 keV photons. The net result would be scattering of the photons.

DaleSpam said:
Two oppositely-travelling 511 keV photons could interact to produce an electron-positron pair. If the photons were just barely 511 keV then the resulting electron and positron would have very little KE and so they would attract each other, anhilate, and produce two 511 keV photons. The net result would be scattering of the photons.

So, is 511keV the minimum? This would make 511keV a very significant energy quantity, wouldn't it? It would seem to be some sort of threshold value for the production of 'free mass', rather than just 'mass defect'.

Replies
4
Views
4K
Replies
1
Views
1K
Replies
26
Views
4K
Replies
19
Views
19K
Replies
1
Views
1K
Replies
38
Views
3K
Replies
13
Views
1K
Replies
5
Views
2K
Replies
4
Views
9K
Replies
7
Views
5K