# QED: the strange theory of light and matter. Light is made of particles ?

QED: the strange theory of light and matter. " Light is made of particles " ?

Over Christmas I ordered the book:

QED: The strange theory of light and matter by Richard P feynamn
2006 edition. Note: This book was written in 1985.

"We know that light is made of particles because we cant take a very sensitive instrument that makes clicks when light shines on it, and if the light gets dimmer, the clicks remain just as loud - just fewer of them"

Page 14 , paragraph 2.

"I want to emphasize to you that light comes in this form- particles. It is very important to know light behaves like particles, especially for those of you who have gone to school , where you were probably told something about light behaving like waves. I'm going to tell you the way it does behave like particles."

Page 15 , paragraph 4.

Richard P. Feynman "QED: The strange theory of light and matter" Princeton University Press , 1985.

I bought this book to learn some facts about light as I was curious to know about light. We have yet to study it. I have been told that light posses properties of both waves and particles. So when i read the above quotes I was concerned. The second quote makes me think he has informed us that he will only be explaining light in terms of its particle like behavior.

The book is old so i am not certain if he really believed light behaved only like a particle. If this is true then a lot of the information i will be learning is incorrect and wish to know if he did mean to only explain it in terms of its particle behavior. If this is true then ill be happy to continue reading the book but for the mean time I have stopped. I have no desire to devote my time to a book that will teach me incorrect things. ( besides the obvious simplifications he makes in the book , such as his reference to "Counting beans". He is talking about the way they calculate lights reflection and says he teaches you a limited way to do it. )

Q1: If I continue to read this book will I learn correct information?

Q2: Do you think he meant to explain it in terms of only particle behavior ?

Q3: He mentions light is composed of photons. He later talks about photons hitting a photon multiplier's "plates" in which electrons bounce off these palates. Is light composed of electrons ?

No, he knows what he's talking about. Nothing has really changed much since Feynman, Schwinger and Tomonaga created QED. Light is particles, they are photons (quanta) which behave as a wave at the same time. This is observed in the double-slit experiment.

mathman

Quantum mechanics is strange. Electrons, which we all consider particles, have wave properties - double slit experiment gives results similar to that for photons.

Thank you for the responses. Thanks to Kevin i will continue to read the book ( which so far seems very interesting ). Thanks to mathman for an explanation of my photon question.

It seems to me the trouble people have with the wave particle duality is largely semantic.
I prefer the term Wave Packet used to describe a thing that is not simply a wave, nor simply a particle, but can exhibit properties of both.

http://en.wikipedia.org/wiki/Wave_packet

ZapperZ
Staff Emeritus

It seems to me the trouble people have with the wave particle duality is largely semantic.
I prefer the term Wave Packet used to describe a thing that is not simply a wave, nor simply a particle, but can exhibit properties of both.

http://en.wikipedia.org/wiki/Wave_packet

But that is "simpler" only if one ignores a whole slew of other issues. For example, try looking at the fourier components that is required to make a "wave packet". It certainly isn't monochromatic. Yet, a "photon" is characterized by only one particular frequency/wavelength. So where are these other frequencies that are needed to form the wave packet?

Zz.

Last edited:
Fredrik
Staff Emeritus
Gold Member

Q1: If I continue to read this book will I learn correct information?

Q2: Do you think he meant to explain it in terms of only particle behavior ?

Q3: He mentions light is composed of photons. He later talks about photons hitting a photon multiplier's "plates" in which electrons bounce off these palates. Is light composed of electrons ?
1. Yes. It's a great book, so keep reading.
2. Yes, but these are quantum particles, not classical particles.
3. No, but the signal produced by a photomultiplier is.
I prefer the term Wave Packet used to describe a thing that is not simply a wave, nor simply a particle, but can exhibit properties of both.
I prefer the term "particle". Terms like "particle", "energy", "electromagnetic field", etc., are defined by the theories, so if a term is used in two different theories, it doesn't mean exactly the same thing in both. If physicists and science teachers could get used to mentioning that little detail, other people wouldn't get so confused by the terminology.

But that is "simpler" only if one ignores a whole slew of other issues. For example, try looking at the fourier components that is required to make a "wave packet". It certainly isn't monochromatic. Yet, a "photon" is characterized by only one particular frequency/wavelength. So where are these other frequencies that are needed to form the wave packet?

Zz.

You're right ZapperZ, the term "wave packet" fails to encapsulate the discrete nature of the photon.
Confusion does arises from the reliance on classical words like "particle" and "wave".
I like what Fredrik said, that the photon is a quantum particle! :)

It seems to me the trouble people have with the wave particle duality is largely semantic.
I prefer the term Wave Packet used to describe a thing that is not simply a wave, nor simply a particle, but can exhibit properties of both.

http://en.wikipedia.org/wiki/Wave_packet

The confusing thing (at least to me) is that we have to associate a fixed frequency to the photon: E = h * f
This is the energy quantization of electromagnetic waves found by Einstein in the Photo-Effect: the energy transported by an electromagnetic wave of frequency f is always a multiple of its frequeny times the planck constant. These elementary excitations of the electromagnetic field are called photons.

However, interpretation as a wave packet requires necessarily the superposition of various frequencies. So, can the photon really be regarded as a wave packet ?

Last edited:
Cthugha

This is the energy quantization of electromagnetic waves found by Einstein in the Photo-Effect: the energy transported by an electromagnetic wave of frequency f is always a multiple of its frequeny times the planck constant. These elementary excitations of the electromagnetic field are called photons.

However, interpretation as a wave packet requires necessarily the superposition of various frequencies. So, can the photon really be regarded as a wave packet ?

The main problem is that the term photon is used for two different things. First, a quantized excitation of a single mode of the em field is termed photon. Second, the photon number eigenstates of the electromagnetic field with a total occupation of unity are often also called photon or better: single photon. In this case, you get a superposition of several modes which have to sum up to a fixed photon number. This is what you get in applications from the so-called single-photon emitters.

Thanks, Cthugha - something to think about. :)

But that is "simpler" only if one ignores a whole slew of other issues. For example, try looking at the fourier components that is required to make a "wave packet". It certainly isn't monochromatic. Yet, a "photon" is characterized by only one particular frequency/wavelength. So where are these other frequencies that are needed to form the wave packet?
In the interference pattern you find sending single photons to a diffraction grating or to a simple prism...

A. Neumaier

a "photon" is characterized by only one particular frequency/wavelength. So where are these other frequencies that are needed to form the wave packet?

This is not true. A 1-photon state in vacuum can have the shape of an _arbitrary_ solution of the vacuum Maxwell equations. In particular, it is the rule that so-called ''photons on demands'' are created in wave packets - or rather mixtures of wave packets, described by reasonably localized density matrices.

More or less monochromatic photons are a special case, commonly produced by lasers. However, even these are not strictly monochromatic. ''In the physical sense, no source of electromagnetic radiation is purely monochromatic, since that would require a wave of infinite duration as a consequence of the Fourier transform's localization property (cf. spectral coherence). Even very controlled sources such as lasers operate in a range of frequencies (known as the spectral linewidth). In practice, filtered light, diffraction grating separated light and laser light are all routinely referred to as monochromatic. Often light sources can be compared and one be labeled as “more monochromatic” '' (http://en.wikipedia.org/wiki/Monochromatic_light )

This is not true. A 1-photon state in vacuum can have the shape of an _arbitrary_ solution of the vacuum Maxwell equations. In particular, it is the rule that so-called ''photons on demands'' are created in wave packets - or rather mixtures of wave packets, described by reasonably localized density matrices.

But isn't it the case that not just any wave solution, but one which indicates a self-propagating wave is manifest that applies to the photon? In other words, evanescent waves or wave ripples aren't associated with photons?

If that's the case then there would seem to be a limit to the amount of dispersion or diffraction a photon can tolerate before breaking up and fracturing into what might be termed virtual photons.

ZapperZ
Staff Emeritus

This is not true. A 1-photon state in vacuum can have the shape of an _arbitrary_ solution of the vacuum Maxwell equations. In particular, it is the rule that so-called ''photons on demands'' are created in wave packets - or rather mixtures of wave packets, described by reasonably localized density matrices.

More or less monochromatic photons are a special case, commonly produced by lasers. However, even these are not strictly monochromatic. ''In the physical sense, no source of electromagnetic radiation is purely monochromatic, since that would require a wave of infinite duration as a consequence of the Fourier transform's localization property (cf. spectral coherence). Even very controlled sources such as lasers operate in a range of frequencies (known as the spectral linewidth). In practice, filtered light, diffraction grating separated light and laser light are all routinely referred to as monochromatic. Often light sources can be compared and one be labeled as “more monochromatic” '' (http://en.wikipedia.org/wiki/Monochromatic_light )

It's my turn to bring you back to some reality. Even if the source is not strictly monochromatic, it certainly does NOT have the same fourier components to actually make such a well-defined wavepacket to describe a single photon. Do you think the slight spread in frequency can actually construct an isolated wavepacket that approaches a square pulse?

Furthermore, the frequency spread of the photons themselves is not what we use to construct the wavefunction that represent these photons.

Zz.

physicsworks
A. Neumaier

But isn't it the case that not just any wave solution, but one which indicates a self-propagating wave is manifest that applies to the photon? In other words, evanescent waves or wave ripples aren't associated with photons?
_All_ waves that correspond to a solution of the Maxwell equation in vacuum are associated with a single photon if only a single photon had been around before the spectacle began.
If that's the case then there would seem to be a limit to the amount of dispersion or diffraction a photon can tolerate before breaking up and fracturing into what might be termed virtual photons.
Only the intuitive particle view of a photon cannot tolerate this and breaks down.

The space of 1-photon wave functions in vacuum consists of all superpositions of fixed helicity monochromatic plane wave solutions of the Maxwell equations. Thus for each 4-vector p with p_0>0 and p^2=0, and for each sign h of the helicity, there is an unnormalized basis state |p,h>, whose superpositions
$$\psi = \int dp \Big(\psi_+(p)|p,+\rangle + \psi_-(p)|p,-\rangle\Big)$$
define an arbitrary photon state. It is monochromatic precisely when only the p with
fixed p_0 are nonzero. It forms a beam in the direction of a spatial unit direction vector \n
precisely when (assuming c=1) the spatial part \p of p is approximately equal to p_0 \n.

Diffraction at a grid turns the photon state into a superposition of spherical solutions of the Maxwell equations, and dispersion changes the direction of p in a p_0-dependent way, so that a beam that is far from monochromatic becomes wider (and the colors separate).

All this happens to _each_ single photon passing an optical instrument. There is nothing virtual in this. But when the photon is no longer well localized along a beam it no longer makes sense to think of the photon as a particle, whereas the field view remains valid.

Last edited:
A. Neumaier

Even if the source is not strictly monochromatic, it certainly does NOT have the same fourier components to actually make such a well-defined wavepacket to describe a single photon. Do you think the slight spread in frequency can actually construct an isolated wavepacket that approaches a square pulse?

Furthermore, the frequency spread of the photons themselves is not what we use to construct the wavefunction that represent these photons.

1. Wave packets are not necessarily square pulses but wave functions whose support at a given time is almost wholly in a tiny region of space. http://en.wikipedia.org/wiki/Main_Page

2. I was associating the wave packets with photons on demands, not with a general source that is not strictly monochromatic. I associated the lack of monochromaticity with laser light, which produces random photons, not photons on demand.

3. The 1-photon Hilbert space is a Hilbert space. This means that arbitrary superpositions of 1-photon states are (apart from a normalizing factor) again 1-photon states.
in particular, superimposing monochromatic 1-photon states of very different colors produce photon states that are very far from monochromatic. In principle, they could even be white.

ZapperZ
Staff Emeritus

1. Wave packets are not necessarily square pulses but wave functions whose support at a given time is almost wholly in a tiny region of space. http://en.wikipedia.org/wiki/Main_Page

2. I was associating the wave packets with photons on demands, not with a general source that is not strictly monochromatic. I associated the lack of monochromaticity with laser light, which produces random photons, not photons on demand.

3. The 1-photon Hilbert space is a Hilbert space. This means that arbitrary superpositions of 1-photon states are (apart from a normalizing factor) again 1-photon states.
in particular, superimposing monochromatic 1-photon states of very different colors produce photon states that are very far from monochromatic. In principle, they could even be white.

I still don't see how in these arguments, you could say that the wavefunction representing a single photon is a wavepacket. It is also is confusing because you're bringing up different arguments under different situations that don't quite apply to this case, which is similar to your previous post as well.

Can you show me an experiment that has a single photon that is actually a superposition of several frequencies?

Zz.

A. Neumaier

I still don't see how in these arguments, you could say that the wavefunction representing a single photon is a wavepacket.
The wave function representing a single photon during the time interval [t_1,t_2] is a wave packet if and only if at each time in this interval the amplitude is virtually zero outside a tiny region (which moves with the speed of light, and grows with time). Many such solutions of the free Maxwell equations exist.
It is also is confusing because you're bringing up different arguments under different situations that don't quite apply to this case, which is similar to your previous post as well.
If you can pin down the source of the confusion a bit more precisely, I'll be glad to learn how to express myself more clearly. This is the main purpose I have in being here on PF.
Can you show me an experiment that has a single photon that is actually a superposition of several frequencies?
On slide 14-28 of my lecture at http://www.mat.univie.ac.at/~neum/ms/lightslides.pdf , I discuss one of the experiments from the literature producing single photons on demand. The photons produced are quite localized - they need to be it to be useful for signalling information. They are realistically described (as most real photons) by density matrices rather than wave functions, but if you consider the rank one approximation, you get ordinary wave packets.

Since the Fourier decomposition of every wave packet is composed of a wide range of frequencies, you get the required superposition.

_All_ waves that correspond to a solution of the Maxwell equation in vacuum are associated with a single photon if only a single photon had been around before the spectacle began.

Only the intuitive particle view of a photon cannot tolerate this and breaks down.

I think I see your point and could agree with phrase "associated with a single photon". But I'd very much hesitate to say "comprise a single photon" which you seem to be asserting. I think it's a matter of definition to say what a photon is comprised of and strict definitions are lacking in the field as far as I can gather.

You may be entitled to develop a definition based on your understanding and conception, but others might do so with different qualifications. I'm wondering if it might not necessarily be a step forward to always dispense with a semi-classical perspective. One of the potential problems with your line of thinking as I understand it so far is in dealing with both transverse and longitudinal solutions of the Maxwell equations in dispersive media. Those are well known and studied in the field of Plasma Physics. Naturally the longitudinal solutions are not self-propagating but do have trajectories that diverge significantly from the transverse solutions. I'd tend to call those "effects of a photon's transit" rather than the photon itself as the energy perturbations are no longer localized.

Thanks, by the way, for the lecture notes. I'll try to read them over the next few days.

A. Neumaier