Photon particle independent of wave question

[kurt101]

A photon acts like a wave and a particle. In the double slit experiment the photon seemingly interferes with itself which is troublesome to me. To help better understand this, I would rather think of the photon as a particle and the wave as something that is independent of the photon where the wave is generated at the photon source and travels in all directions. What are the problems with thinking that a photon is a particle that travels in a specific direction and the wave is something separate that travels from the source in all directions (spherically) and that the photon happens to be synchronized with the wave when it is propagated from its source? Are there things we would expect to observe that we do not if this was the case?

I was also thinking this conceptual model would lead to the Feynman path integral approach because when determining the photon's trajectory you would have to consider the wave going in all directions and the strength of each contribution would diminish by distance. I only understand the path integral approach at a high level as explained by Dick Feynman in his book "QED: The Strange Theory of Light and Matter".

 

[.Scott]

Perhaps the situation that makes it easiest to see how the "particle" model doesn't work is with an interferometer.
You can use a beam splitter to send a beam of light in two directions. Then mirrors to bring the two beams back together.
The result is an interference pattern with bright and dark bands.

Now position your photo detector exactly along one of those dark bands. It detects nothing.
Now block one of the two beams created by the beam splitter. The detector begins registering photons.
This will happen even if the beam of light is of such low intensity that only one photon is emitted at a time.

So the situation you have is that when you block one of the paths, the photons that are detected are there because they could have hit the block, but didn't. This is called a counter-factual.

 

[DrChinese]

A photon acts like a wave and a particle. In the double slit experiment the photon seemingly interferes with itself which is troublesome to me. To help better understand this, I would rather think of the photon as a particle and the wave as something that is independent of the photon where the wave is generated at the photon source and travels in all directions. What are the problems with thinking that a photon is a particle that travels in a specific direction and the wave is something separate that travels from the source in all directions (spherically) and that the photon happens to be synchronized with the wave when it is propagated from its source? Are there things we would expect to observe that we do not if this was the case?

I was also thinking this conceptual model would lead to the Feynman path integral approach because when determining the photon's trajectory you would have to consider the wave going in all directions and the strength of each contribution would diminish by distance. I only understand the path integral approach at a high level as explained by Dick Feynman in his book "QED: The Strange Theory of Light and Matter".

Yes, some photons spherically emanate in all directions, and some photons are later found in specific well localized spots. But many are not. The devil is in the details.

It is not hard to model light if you only intend to model a few cases. However, there are thousands of experiments around light/photons, and simple models just don't cut it past a certain point. Here is a paper that explores some of the differences between quantum and semi-classical models of light. Specifically, it details how photons detector clicks magically match the arrival and existence of light waves. The point being, if there were a wave, that matching would not be so exact. The article explains this, and goes on to show that models (much like yours) are ruled out by 377 standard deviations.

http://people.whitman.edu/~beckmk/QM/grangier/Thorn_ajp.pdf

Now of course you can always wave your hands to explain away any single experiment. But you end up with yet another new model with progressively less and less predictive power (because you are hand adding elements). Instead, there is another theory - the standard QM model - that works day in and day out without needing any adjustment or ad hoc elements. It is more mathematical in nature, and defies a single simple representation in (geometrical) model form. It is very good, but even it doesn't explain why a photon randomly appears in a specific spot (rather than some other, equally likely spot).

 

[kurt101]

Perhaps the situation that makes it easiest to see how the "particle" model doesn't work is with an interferometer.

Thank you for the response.

In the experiment you describe, in the case where there is no blocking, I would expect the wave to be split by the beam splitter, reflect off of the mirror, and then interfere when the waves are recombined. In the case where you block one of the paths, I would expect no interference. In both cases, the photon would travel on one of the waves and you would get the expected results.

 

[kurt101]

Now of course you can always wave your hands to explain away any single experiment. But you end up with yet another new model with progressively less and less predictive power (because you are hand adding elements). Instead, there is another theory - the standard QM model - that works day in and day out without needing any adjustment or ad hoc elements. It is more mathematical in nature, and defies a single simple representation in (geometrical) model form. It is very good, but even it doesn't explain why a photon randomly appears in a specific spot (rather than some other, equally likely spot).

Thanks for the reference to the paper. I will read it soon.

It is my understanding that the Feynman Path Integral approach is the fundamental basis for Quantum Field Theory and QFT explains more than the QM. I was thinking that what I was describing was similar to the Feynman Path Integral approach that Feynman explains in his book. At this point, I still don't understand how they would give a different result. The only thing that I think I am saying different then what is described in physics books is that the wave that guides the photon is independent of the photon and propagates with or without a photon riding it. At least from my naive perspective it explains why there is interference when only 1 photon is sent at a time through the double slit. I was wondering why this perspective is not valid. Hopefully the link you gave helps explain why.

 

[.Scott]

Thank you for the response.

In the experiment you describe, in the case where there is no blocking, I would expect the wave to be split by the beam splitter, reflect off of the mirror, and then interfere when the waves are recombined. In the case where you block one of the paths, I would expect no interference. In both cases, the photon would travel on one of the waves and you would get the expected results.

Okay. But without the wave, the particle has no way of knowing whether it is allowed to hit the detector. The wave is not just incidental - it is what defines the "path" of the particle.

Whenever the particle hits the detector, it is doing do because of something that was not in its path.

 

[kurt101]

It is not hard to model light if you only intend to model a few cases. However, there are thousands of experiments around light/photons, and simple models just don't cut it past a certain point. Here is a paper that explores some of the differences between quantum and semi-classical models of light. Specifically, it details how photons detector clicks magically match the arrival and existence of light waves. The point being, if there were a wave, that matching would not be so exact. The article explains this, and goes on to show that models (much like yours) are ruled out by 377 standard deviations.

http://people.whitman.edu/~beckmk/QM/grangier/Thorn_ajp.pdf

I read the paper and it is my understanding that the paper discusses an experiment that is trying to prove that photons are detected as distinct events. As far as I can tell you misunderstood my question. I think the other person who replied did as well, but thanks for taking the time to respond.

I fully believe photons are particles and are detected as distinct events (like the experiment in the paper is proving).

I might just be describing de Broglie–Bohm pilot wave theory. I understand in this theory the photon is guided by the quantum potential or the wave equation. At this point, I probably need to learn more about pilot wave theory. I don't understand how the QM wave used to drive the photon in pilot wave theory differs from some kind of periodic spherical wave created at the point a photon is emitted. I also don't understand how the pilot wave theory differs from the Feynman path integral approach and why pilot theory can't be made to give the same end result, but in a deterministic fashion, where probability is the end result, but not the driving mechanism.

 

[kurt101]

Okay. But without the wave, the particle has no way of knowing whether it is allowed to hit the detector. The wave is not just incidental - it is what defines the "path" of the particle.

Whenever the particle hits the detector, it is doing do because of something that was not in its path.

I may have chosen a poor choice of words to say that the photon is "independent" of the wave. I meant to ask why the photon and the wave are not considered distinct physical entities with different properties?

 

[PeterDonis]

A photon acts like a wave and a particle

No, it doesn't. Whenever we measure sufficiently faint light, we always detect individual, discrete particles; we never detect waves.

The "wave" behavior comes in when we look at the pattern formed by a large number of particle detections; under appropriate conditions, it is an interference pattern similar to those formed by waves. But that pattern is not a single observation; as I said, it's a combination of many, many particle observations.

What are the problems with thinking that a photon is a particle that travels in a specific direction and the wave is something separate that travels from the source in all directions (spherically) and that the photon happens to be synchronized with the wave when it is propagated from its source? Are there things we would expect to observe that we do not if this was the case?

I can't tell because you haven't given an actual model that can make predictions. You would have to construct such a model and then calculate what it predicts. Just describing your idea in words isn't enough. (And constructing such a model and calculating its predictions would be original research, which is not what PF is for.)

I was also thinking this conceptual model would lead to the Feynman path integral approach because when determining the photon's trajectory you would have to consider the wave going in all directions and the strength of each contribution would diminish by distance.

The Feynman path integral does not tell you the photon's trajectory. It tells you the amplitude for a photon detection event at a particular point in spacetime, given that there was a photon emission event at some other point in spacetime. And knowing those two things, which is all we ever know in an actual experiment, does not tell you that the photon took any particular trajectory between those events; that's the whole point of the path integral, that it doesn't pick out any particular trajectory as "privileged", but just adds up the amplitudes for all trajectories, no matter how "unphysical" they seem.

 

[kurt101]

No, it doesn't. Whenever we measure sufficiently faint light, we always detect individual, discrete particles; we never detect waves.

The "wave" behavior comes in when we look at the pattern formed by a large number of particle detections; under appropriate conditions, it is an interference pattern similar to those formed by waves. But that pattern is not a single observation; as I said, it's a combination of many, many particle observations.

This is all consistent with my understanding, but thanks for describing it.

I can't tell because you haven't given an actual model that can make predictions. You would have to construct such a model and then calculate what it predicts. Just describing your idea in words isn't enough. (And constructing such a model and calculating its predictions would be original research, which is not what PF is for.)

I am looking for insight regarding models like what I described and why they don't work. Maybe there are not any, but that is why I ask on this forum. I am not trying to propose a new model.

The Feynman path integral does not tell you the photon's trajectory. It tells you the amplitude for a photon detection event at a particular point in spacetime, given that there was a photon emission event at some other point in spacetime. And knowing those two things, which is all we ever know in an actual experiment, does not tell you that the photon took any particular trajectory between those events; that's the whole point of the path integral, that it doesn't pick out any particular trajectory as "privileged", but just adds up the amplitudes for all trajectories, no matter how "unphysical" they seem.

I understand that the Feynman path integral approach is a mathematical construct used to determine the probability that a photon from one point in spacetime will arrive at another point in spacetime. It is also my understanding that Feynman based this idea from the double slit experiment and what would happen if he made it with 3 slits and then 4 slits and then with a infinite number of slits. So there is a close tie to the double slit experiment and this approach. If you were to come up with a deterministic model that recreates the result of the Feynman path integral approach, it seems like the double slit experiment with photons is a good place to start. I am trying to understand what deterministic models there are for explaining this experiment.

I have read a little about the de Broglie–Bohm pilot wave model. Other than the premise of determinism, I don't like it. It seems to just hand wave the trajectory of particles with quantum potential and does not seem to add any understanding.

 

[PeterDonis]

I am looking for insight regarding models like what I described and why they don't work.

And I'm saying that unless you can give an actual model--a mathematical model that makes quantitative predictions--such a question can't be answered. Just a hand-waving description of something that looks to you like a model isn't enough.

I am not trying to propose a new model.

Since what you describe doesn't look like any model that exists in the literature, it looks to me like you are, whether you realize it or not.'

 

[PeterDonis]

It is also my understanding that Feynman based this idea from the double slit experiment and what would happen if he made it with 3 slits and then 4 slits and then with a infinite number of slits.

That is a heuristic argument for why something like the path integral is plausible, yes. It's not a rigorous calculation of what the path integral predicts.

So there is a close tie to the double slit experiment and this approach.

Not really. The fact that the double slit experiment happened to provide Feynman with a heuristic plausibility argument does not mean the actual path integral approach, the rigorous approach that is actually used for calculations, is more closely connected to the double slit than to any other quantum experiment.

I am trying to understand what deterministic models there are for explaining this experiment.

I don't think there are any in a practical sense. Yes, technically the de Broglie-Bohm model is deterministic, in the sense that given a particular set of initial positions and velocities for all of the quantum particles, the time evolution is deterministic. But it's also a fundamental premise of the model that the actual positions and velocities of the quantum particles are in principle unknowable and unobservable, so the "determinism" of the model does no work as far as actually making predictions is concerned.

 

[Demystifier]

A photon acts like a wave and a particle. In the double slit experiment the photon seemingly interferes with itself which is troublesome to me. To help better understand this, I would rather think of the photon as a particle and the wave as something that is independent of the photon where the wave is generated at the photon source and travels in all directions. What are the problems with thinking that a photon is a particle that travels in a specific direction and the wave is something separate that travels from the source in all directions (spherically) and that the photon happens to be synchronized with the wave when it is propagated from its source? Are there things we would expect to observe that we do not if this was the case?

There is an interpretation of quantum mechanics that proposes something very similar to that. It is called Bohmian mechanics. See e.g. https://arxiv.org/abs/quant-ph/0611032

 

[kurt101]

And I'm saying that unless you can give an actual model--a mathematical model that makes quantitative predictions--such a question can't be answered. Just a hand-waving description of something that looks to you like a model isn't enough.



Since what you describe doesn't look like any model that exists in the literature, it looks to me like you are, whether you realize it or not.'

Fair enough. I described it all wrong, but what I am thinking of is the picture you see with waves going through the double slit experiment in many general physics books. The kind described by Huygens–Fresnel principle.

I know using Wikipedia is frowned up, but I am just using it to help with my question.

From https://en.wikipedia.org/wiki/Huygens–Fresnel_principle it says:

Huygens' theory served as a fundamental explanation of the wave nature of light interference and was further developed by Fresnel and Young but did not fully resolve all observations such as the low-intensity double-slit experiment that was first performed by G. I. Taylor in 1909, see double-slit experiment. It was not until the early and mid 1900s that quantum theory discussions, particularly the early discussions at the 1927 Brussels Solvay Conference, where Louis de Broglie proposed his de Broglie hypothesis that the photon is guided by a wavefunction.[9] The wavefunction presents a much different explanation of the observed light and dark bands in a double slit experiment. Feynman partially explains that a photon will follow a predetermined path which is a choice of one of many possible paths. These chosen paths form the pattern; in dark areas no photons are landing and in bright areas many photons are landing. The path of the photon or its chosen wavefunction is determined by the surroundings: the photon's originating point (atom), the slit, and the screen. The wavefunction is a solution to this geometry. The wavefunction approach was further proven by additional double-slit experiments in Italy and Japan in the 1970s and 1980s with electrons.

It says that Fresnel and Young did not fully resolve all observations such as the low-intensity double-slit experiment. However if you consider the photon as distinct from the wave, I assume this problem goes away. That is the heart of my question. I assume this was considered by others, but why did it fail? Why did instead the wave function approach take hold with those following this type of model?

 

[DrChinese]

I read the paper and it is my understanding that the paper discusses an experiment that is trying to prove that photons are detected as distinct events. As far as I can tell you misunderstood my question. I think the other person who replied did as well, but thanks for taking the time to respond.

I fully believe photons are particles and are detected as distinct events (like the experiment in the paper is proving).

...

A photon with a reasonably definite position (where the photon number is well defined as 1) can be modeled as a "particle". When its position is vague (photon number still well defined as 1), you could consider it a "wave". On the other hand, a photon with a reasonably definite position (where the photon number is well defined as 1) can be predicted to arrive at a variety of places according to the evolution of a "probability wave". It's evolution is affected by self-interference of the possible paths that could have been taken. Yet a "photon as particle" is only detected at a single spot (since its photon number is well defined as 1).

You can label it whatever you like. And yes, models of photons as "wave" or "particle" can be useful in many situations. But don't confuse the model with the underlying object. That is what PeterDonis is saying: it is the mathematical model which is maximally useful... while these heuristic models (i.e. what I describe above) have limitations that render them of marginal value. You can easily poke holes in them if you try. So don't forget the caveat.

Note that in my description above, I had to specify that the photon number is 1. That is a Fock state. In the experiment I cited, it was arranged so the photon number was 1. There are also situations where the average number is 1, but you do not have a Fock state (the photon number might sometimes be 0 or 2, for example, but averages to 1). In that case, the photon description (behavior) might be considerably different.

 

[PeterDonis]

what I am thinking of is the picture you see with waves going through the double slit experiment in many general physics books. The kind described by Huygens–Fresnel principle.

This picture is based on the classical model of light, not the quantum model of light.

I know using Wikipedia is frowned up, but I am just using it to help with my question.

You shouldn't. Wikipedia in general is not a good source if you don't have an independent way of validating what it says. In this particular case I don't think Wikipedia is a good source. For one thing, describing our current quantum model of photons as using a "wave function" is not correct; we model photons using quantum field theory, which does not use wave functions. "Wave function" is really a description of the mathematical object that satisfies the non-relativistic Schrodinger equation, but there is no such thing as a non-relativistic photon.

It says that Fresnel and Young did not fully resolve all observations such as the low-intensity double-slit experiment.

That's because the double slit experiment wasn't even done until a century or so after Young, and half a century or so after Fresnel.

if you consider the photon as distinct from the wave, I assume this problem goes away.

I have no idea why you would make such an assumption. In any case, it's wrong.

I assume this was considered by others, but why did it fail? Why did instead the wave function approach take hold with those following this type of model?

These questions can't even be answered because the premise on which they are based--that photons are modeled in quantum theory using wave functions--is wrong. See above.

I think you need to stop trying to learn about photons from Wikipedia and crack open an textbook on quantum field theory and QED.

 

[PeterDonis]

There is an interpretation of quantum mechanics that proposes something very similar to that. It is called Bohmian mechanics.

And, as you are aware, the original Bohmian mechanics is non-relativistic, and relativistic versions (such as your own) cannot be said to be non-controversial or mainstream (which is not to say they're necessarily wrong, just that their status is not settled). So I'm not sure it can be applied to photons in a "B" level discussion here.

 

[DrChinese]

I think you need to stop trying to learn about photons from Wikipedia and crack open an textbook on quantum field theory and QED.

Always sage advice...

 

[DrChinese]

Fair enough. I described it all wrong, but what I am thinking of is the picture you see with waves going through the double slit experiment in many general physics books. ... However if you consider the photon as distinct from the wave, I assume this problem goes away. That is the heart of my question.

For all practical purposes, there is no model of a photon consisting of particle component and a distinct wave component. Even in the Bohmian world, the hypothetical guide wave really doesn't belong to the photon exclusively (as I understand it anyway).

You ought to be able to see that you have stretched your model, and are being forced to keep stretching it more every time a new objection comes along. Not a sign of a good model. And making assumptions about problems magically disappearing due to a model is pretty much what hand-waving is all about. As I said earlier, the devil is in the details.

On the other hand: the generally accepted mathematical model doesn't have these problems. Instead, it makes new predictions that can be tested experimentally. Entire careers are spent working through the details. There are many many novel effects that are discovered as a result of a great model. Ad hoc models - ones that ignore prior research especially - rarely predict anything new. Hopefully you can see why PeterDonis is recommending you review existing material before attempting to flesh out ideas in your head.

What you are trying to understand has much research available already. You obviously have the interest, why not take it to the next level?

 

[kurt101]

Wikipedia in general is not a good source if you don't have an independent way of validating what it says.

Is there a good online QM reference that you would recommend?

That's because the double slit experiment wasn't even done until a century or so after Young, and half a century or so after Fresnel.

My comment and the Wikipedia article were referring to the Huygens-Fresnel model and around 1927 after the double slit experiment was performed.

I think you need to stop trying to learn about photons from Wikipedia and crack open an textbook on quantum field theory and QED.

I am currently working through https://www.amazon.com/dp/0470026790/?tag=pfamazon01-20

 

[PeterDonis]

I am currently working through https://www.amazon.com/dp/0470026790/?tag=pfamazon01-20

This looks like it's mainly non-relativistic QM. That's certainly useful as a base to build on, but you'll need to follow up with something more focused on quantum field theory. I find Zee's Quantum Field Theory In a Nutshell to be pretty good, but preferences in QFT textbooks seem to be somewhat idiosyncratic, so it's hard to tell what will work best for you. Weinberg's classic 3-volume Quantum Theory of Fields is certainly a comprehensive reference, but it's probably not a good introduction if you have no previous background in QFT (although Chapter 1 of volume 1 gives a very good historical overview of the subject).

Unfortunately I don't know of any good online QFT texts.

 

[Demystifier]

And, as you are aware, the original Bohmian mechanics is non-relativistic, and relativistic versions (such as your own) cannot be said to be non-controversial or mainstream (which is not to say they're necessarily wrong, just that their status is not settled). So I'm not sure it can be applied to photons in a "B" level discussion here.

The OP asked about photons, which indeed are relativistic particles. But it seems to me that he is not really interested in photons only, but also in other particles, such as electrons, which can be treated from a non-relativistic point of view. Hence I believe that he might find Bohmian ideas in non-relativistic QM very insightful for what he really wants to understand. Note also that I did not recommend to him any advanced work. I recommended a basic pedagogic paper on the subject written on a level which might be suitable to him.