Photons Are Here to Stay, Deal With It

I spend a lot of time promoting Rhett Allain’s Dot Physics blog, enough that some people probably wonder if I get a cut of his royalties (I don’t). I’m going to take issue with his latest, though, because he’s decided to revive his quixotic campaign against photons, or at least teaching about photons early in the physics curriculum. We went through this back in 2008 and 2009 (though Rhett’s old posts are linkrotted away, so you only get my side of the story…). I’m no more convinced this time around, even though he drags in Willis Lamb and David Norwood for support.

There are basically two pieces to the anti-photon argument, neither of which I find remotely convincing. In fact, I will happily stipulate that both of the central points are correct, and even after that, I don’t find this to be a problem that requires fixing.

The first claim is that photons are redundant, given that you can explain all of the phenomena they’re usually invoked to explain without ever referring to particle-like characteristics of light. Which is true– you can construct a semi-classical model of the photoelectric effect in which electrons inside metals occupy quantized states and are excited out of them by classical electromagnetic waves of the appropriate frequency. That reproduces essentially all of the features of the quantum model proposed by Einstein in 1905 and confirmed by Millikan’s experiments in 1916. Using that instead of the photon model would be a little ahistorical– the semiclassical model was first published by Mandel and Wolf in the 1960’s, and relies on a Fermi Golden Rule calculation assuming the Schrödinger equation which wasn’t invented until the late 1920’s– but you could do it.

But even allowing that it’s true, I don’t see what the point is. To bring it back to classical physics, in the same way that it’s perfectly true that you can describe the photoelectric effect with a semiclassical model, it’s perfectly true that you can describe an elastic collision between two objects by direct integration of Newton’s second law (or the Momentum Principle, in the language of the Matter and Interactions textbook that Rhett and I both use). But there’s absolutely no reason to do that, other than to make a philosophical point– sane people considering elastic collision problems make use of energy in addition to momentum, because it makes the problem simpler.

And that’s the case for photons in talking about the photoelectric effect: it’s much, much easier. Actually working out the details of the semi-classical model of the photoelectric effect is really complicated: you need to know about the Schrödinger equation, make a few approximations, and do an integral involving complex numbers. The photon version requires subtraction.

We invoke photons for the photoelectric effect because it’s much, much simpler. And since a quantized model of light is known to be necessary to explain photon anti-bunching (a point even Rhett concedes), there’s no good reason not to employ it there.

(I’ll note in passing that Norwood makes repeated references to some sort of experiment that supposedly shows a delay in electron emission for low light intensity. I have absolutely no idea what he’s talking about, and he doesn’t provide a citation. The only “delay in photoemission” measurements I’m aware of are attosecond scale delays after excitation with an ultrafast alser pulse, which is not remotely the same thing.)

The second argument against photons basically amounts to the language used to describe light in terms of particles being imprecise in a way that offends some people’s aesthetic sense. And, again, strictly speaking this is perfectly true. Photons are not perfectly described as particles with all the properties of classical particles. A proper description is that photons are quantized excitations of particular modes of the electromagnetic field.

But you know what else isn’t perfectly described as a particle with all the properties of classical particles? An electron. In fact, strictly speaking, electrons also ought to be described as quantized excitations of an “electron field.” There are some differences between the mathematical descriptions of photons and electrons as field excitations, but we’ve known since Dirac’s day that electrons are best described as field quanta.

And yet, you don’t find many physicists willing to argue that we shouldn’t teach electrons as particles. But all the same linguistic ambiguities are present with electrons that are present with photons. Neither is truly a particle or a wave in the classical sense– rather, they’re both a third kind of object for which we lack a convenient single word. There’s necessarily a lot of imprecision in the language used to talk about this; Lamb’s article includes a kind of snotty remark blaming Bohr for this, but I don’t see any great alternatives.

And even after stipulating that talking about photons as particles is imprecise, I fail to see what the problem is. Or, more specifically, I don’t see where this particular bit of imprecision creates a real problem for anything. If this is a pernicious misconception, what is the physical problem that thinking of light in terms of photons keeps you from solving? Where does it lead students astray in a way that gets a clearly wrong answer, as opposed to getting the right answer by aesthetically unappealing means? I haven’t seen a good example yet, though this is at least the third time I’ve read a bunch of anti-photon rants.

So, like the post title says, I see no reason to drop photons. A fully quantized model of light is unquestionably necessary for more advanced experiments, invoking a simplified version of it makes certain classes of intro problems vastly easier to deal with, and the imprecision that the simplified model introduces doesn’t seem to cause any significant problems (particularly for the vast majority of students who will only ever take introductory-level classes). There’s just nothing there that rises to the level of a problem requiring a change in pedagogy.

(I suspect the closest analogue in classical physics is the “work done by friction” business, where a lot of physics education folks vehemently object to the notion of frictional forces doing work, for reasons I have never quite understood. I’ve had it explained to me several times, and all I remember about it is that it turns on a poor choice of what you call the system. This has basically eliminated a whole class of problems from the intro classes– you won’t find problems where students find the stopping distance for a sliding object using energy methods any more– in order to avoid a “misconception” that seems to me to be almost entirely an aesthetic issue.)

16 Replies to “Photons Are Here to Stay, Deal With It”

  1. I am not a physicist, but can Rhett Allain explain Compton scattering without photons?

  2. As a reference, Art Hobson at U. Arkansas makes an argument for fields over particles:

    I continue to think that the particle model has useful relevance, and I think that switching to an all-fields model for introductory physics would make physics all that much more unreachable (and irrelevant, and unfundable) to the “the common people”.

    Disclaimer–I haven’t ready Rhett’s post yet, so this comment is general with respect to the topic, not specific to Rhett’s article.

  3. Beginners intuitively connect photons in photoelectric effect with billiard balls: “photon in, electron out, we’re done here”. Discussions of wave properties of photons and electrons rarely connect back to discussions of photons and particles, beyond generic wave-particle duality. To me this summarizes the case against photons: they create a intuitive but incorrect model overpowering the much more important wave model.

  4. A photon doesn’t just have energy. It has linear momentum and angular momentum as well. You could presumably construct a field with those properties, but that involves math sufficiently esoteric that it would have to be relegated to a graduate level course. So while it is technically possible to explain all of the relevant physics without resorting to photons, the attempt would likely introduce more problems than it solves.

    Conversely, despite never having taken a formal solid state physics course I am aware of the existence of phonons, which are basically quantized sound waves, as a tool for describing certain phenomena. As with photons, this is done because some phenomena are conceptually easier to describe in terms of phonons rather than a field of sound waves. I’ll defer to others more knowledgeable than I about the feasibility of eliminating discussion of phonons from solid state courses, but if you think phonons are useful, it is hard to argue that photons are not.

    I’m also boggling about the objection to friction forces doing work. What’s the alternative description that would be intelligible at the freshman level? I get that invoking friction is a highly simplified description of what’s going on, and as with photons vs. fields there is a level where that simplification is inappropriate. But it’s rare for such situations to arise at the undergraduate physics level.

  5. I don’t like the notion of photons because they muddle more then they illuminate. The only sensible definition of a photon is as a click in a photodetector.

  6. The notion of a photon as a particle has some problems, but you need to think pretty deeply about physics, and study it at a pretty high level, to get there. OTOH, you can get started in the lab from day 1 if photons are particles. Indeed, I suspect that even the people who object most loudly to this probably think in terms of photons as particles with (more-or-less-well-defined) energy, linear momentum, and angular momentum when they are trying to understand experiments. I’m a theorist, but experiments are what science is ultimately all about.

    As to friction, yeah, I’ve heard some of those complaints, but the issue never made much sense to me. I get the impression (perhaps mistakenly) that it isn’t the newer wave of education researchers who complain about the “work done by friction” thing, so much as an older wave of people who want to “get it right, damnit!” I guess my answer is that if it depends on how we define the system, what part of the surface we call the system, etc., then how can we talk about “work done by gravity”? What part of the earth is pulling on me when I fall? In fact, different parts of the earth are exerting different forces on me, and with different directions even! If we’re doing to worry about the fact that different parts of the road are exerting frictional forces on me at different times, then we should worry about the fact that the gravitational force exerted on me by mountains in Canada is mostly northward, not downward. (I’m in the US.)

  7. You can also use a geocentric system to do all of physics if you wish, ideally with the origin of the universal coordinate system right THERE at the corner of a lab bench or your telescope mount. The latter is actually convenient, but not simple, the same issue here.

    If Rhett acknowledges that you need photons to explain and/or predict experimental results like photon anti-bunching, then that is the more fundamental theory. What is wrong with deriving Maxwell’s Equations from QED? More fundamentally, what is wrong with telling future engineers and future physicists that you can derive Maxwell’s Equations from QED but a first year course is not the time or place to do that, anymore than it makes sense to teach Hamiltonian or Lagrangian mechanics to freshmen.

    I also really don’t get the problem with work by friction. Do they also dispute work done by a gas because of some misconception held by someone? By a hand operated by muscles? By the muscle cells when the hand does not move? If you worry about that, shouldn’t you also worry about what “force” means? (The only place where it is clearly defined is in a field theory.) Why does Rhett call gravity a force when the most likely theoretical explanation, necessary for GPS to work, says particles simply follow a geodesic in a model that entirely excludes quantum mechanics? Can you even justify using continuous functions to describe the motion of objects? What experiment tells us that space is continuous?

    Yeah, you can go nuts worrying about the foundations of classical mechanics even before you start worrying about the semantic misconceptions inherent in referring to an atom or a proton, let alone a ball, as “a” particle.

    We use what works.

  8. I would find it very cumbersome to explain things such as photon shot noise if we do away with photons.

  9. Physicists have been trying for 400 years to get rid of the concept of there being “particles” of light, but it refuses to go away. There’s probably a good reason for that.

  10. Complaining about photon sounds like complaining about soliton because some people think it’s some kind of billiard-ball. Anyway, I think “electron field” is a bigger problem myself. Take a look at and note that pair production occurs because pair production occurs. That’s junk science. It ignores the hard scientific evidence. You can make an electron (and a positron) out of light in pair production, you can diffract it, it’s got a magnetic dipole moment, and the Einstein-de Haas effect “demonstrates that spin angular momentum is indeed of the same nature as the angular momentum of rotating bodies as conceived in classical mechanics”. Oh, and in atomic orbitals electrons “exist as standing waves”, and after annihilation you’ve got light again. The photon is a singleton electromagnetic field-variation/wave (or four-potential pulse) propagating linearly at c. The electron is “Dirac’s belt” standing-wave configuration where the field variation is now a standing field. Look to TQFT. It’s like an electromagnetic knot.

  11. @Eric Lund:

    1. I think you can simulate most of the photon’s properties with a classical wave for everyday situations. The linear momentum translates to the exp(ikr) part of the plane wave, and the angular momentum translates to the polarizability (I guess the latter part, but since it is an additional thing that you need to rotate, it should behave like an angular momentum).

    2. The people that do not like photons despise phonons even more with a similar argument. At least Lamb does so, and it makes some sense for consistency. So you would not convince a hardcore Antiphotoner with this argument.

    Having said this, I do agree with Chad. This discussion is about as arcane as hidden parameter theories.

  12. The Hanbury Brown Twiss experiment is a good example of an effect which is easier to understand using classical electromagnetism than using quantum field theory. Roy Glauber got a Nobel prize for straightening everybody out about how to analyse HBT correctly in QED.

    Another slippery point about QED is that photon are massless and often not conserved. Indeed, the most commonly encountered states of the electromagnetic field, coherent states and thermal states, do not have a well defined photon number. If you just think of a light beam as a bunch of particles it’s a little bit hard to get your head around this idea, and can lead to some stupid statements.

    Fock states, which do have definite number of photons, are hard to produce and very fragile. And when you start thinking about Fock states and phase measurements, your intuition, if it only goes as far as “photons are particles”, can quickly lead you astray. For example I sometimes hear arguments like: “If I send a single photon into this interferometer, there will be no fringes because single photons don’t have a phase”. Response: Maybe and maybe not, it depends on exactly what you measure. Thus there are many situations in electromagnetism in which classical thinking will serve you much better than superficial “photon theories”. I therefore have some sympathy with the antiphoton curmudgeons.

    Of course, the best thing to do is to THINK CAREFULLY. This often means you should attempt both a classical and a quantum explanation for a given effect and make sure you understand how and why they are different. I’m not saying it’s easy.

  13. “Neither is truly a particle or a wave in the classical sense…”

    Exactly. We talk about photons as particles when it’s convenient (or, more precisely, say that they behave like particles), and we speak of them as waves (or more precisely, say that they behave like waves) when it’s convenient. And appropriate. Arguing that they aren’t particles is more an exercise in semantics than anything else. Of course they aren’t particles. They are, as you say, a third thing. But so what?

Comments are closed.