A reader emailed me with a few questions regarding How to Teach Physics to Your Dog, one of which is too good not to turn into a blog post:
What is a photon from an experimental perspective?… Could you perhaps provide me with a reference that discusses some experiments and these definitional issues?
The short form of the experimental answer is “A photon is the smallest amount of light that will cause a detector to ‘click.'” (For some reason, hypothetical light detector technology has never really advanced past the Geiger counter stage– even though it’s all electrical pulses these days, we still talk about light detectors as if they made audible sounds.)
That seems more like a statement of limited detector efficiency, but in fact you can show that light really does come in discrete “packets” of energy. Even if the energy of a single photon is significantly greater than the energy required to trigger a “click” in your detector, for a given frequency of light, you will only ever find the energy arriving in single-photon amounts.
How do we know this is the case? There are three great historical experiments that mark the key steps on the way to full acceptance of the photon model: the photoelectric effect, the Compton effect, and photon anti-bunching.
The photoelectric effect, as you might guess from the name, involves using light to knock electrons loose from some substance (essentially every photon detector made uses some form of the photoelectric effect). This was discovered in passing by Heinrich Hertz during his experiments to demonstrate that light is an electromagnetic wave, and posed quite a problem for the wave model of light. You can make a number of predictions from the wave model, and only one of them is borne out in the experiments (as described in the pre-ScienceBlogs version of this blog).
One of Einstein’s 1905 papers was “On a Heuristic Viewpoint Concerning the Production and Transformation of Light,” which proposed a simple model to explain the photoelectric effect using quantized light. A single light-quantum comes along, hits an electron inside a metal, and provides all the energy needed to knock it loose. This model is a radical change in the way we look at light– Einstein referred to it as the only truly revolutionary thing he did in his career– but it reproduces all of the observed results. It even stands up to hostile investigation– the American physicist Robert Millikan set out to disprove Einstein’s model, and wound up confirming it in every detail (which didn’t prevent him from getting a little pissy in his paper). Einstein and Millikan each got Nobel Prizes out of the deal.
The photoelectric effect by itself wasn’t enough, though, and it turns out that you can reproduce all the important aspects using a classical model of light with quantized matter (though this wasn’t worked out in detail until the 1960’s). The experiment that really got people to take the photon picture seriously was the Compton effect, an expeiment on the scattering of x-rays by electrons, carried out by the American physicist Arthur Holly Compton in the early 1920’s.
The Compton effect involves what is essentially a collision between two particles: an electron that is more or less at rest, and a photon of light. In the photon model of light, each quantum of light carries a small amount of energy, and thus necessarily a small amount of momentum, both of which depend on the wavelength of the light. When the photon hits the electron, it will transfer some of its energy and momentum to the electron, which means that after the collision, the photon leaving the area has a different wavelength than the photon that entered. The change in the wavelength is related to the angle between the direction of the exiting photon and the entering photon in a very simple way.
Compton fired x-rays of known energy at a metal target (which contains lot and lots of electrons, whose energies are much smaller than the energy of the photon), and looked at the energy of x-rays leaving the target at different angles. Though he went through the usual series of misinterpretations of his data, he eventually found that his observations agreed perfectly with the predictions of the photon model.
This convinced most people that the notion of light as a particle needed to be taken seriously, but it didn’t completely seal the deal. I have seen it asserted that the Compton effect can also be explained using a wave model of light, though I’ve never read a good explanation of how that works, but it is generally agreed that the experiment that absolutely nails the existence of photons is the photon anti-bunching experiment by Kimble, Dagenais, and Mandel in 1977 (more than 70 years after Einstein’s paper explaining the photoelectric effect in terms of photons). Anti-bunching, as you might expect from the name, involves showing that photons are discrete objects that, under the right circumstances, will be spaced out in time. The goal of the experiment is to show that photons of light are emitted one at a time, and that the detector “clicks” we see really represent single photons.
The way they did this was to take a very weak beam of sodium atoms, and illuminate them with light as they passed near a detector. The light would excite some of the atoms to a higher-energy state, and they would drop back down to the ground state a few nanoseconds later, emitting a photon. These photons were then picked up by the detector.
The beam of atoms was chosen to be very weak, so that there was generally only a single atom present in front of the detector at any time. In this case, you expect to see a particular pattern in the arrival time of the photons at the detector– specifically, you expect to see a delay of a few nanoseconds between the detection of one photon and the detection of another, due to the fact that the single atom will need to be excited again, and decay again in order to produce a second photon that can be detected.
This anti-bunching effect is something that cannot be explained using a classical picture of light as a wave. Using a wave model, in which light is emitted as a continuous sinusoidal wave, you would expect some probability of a detector “click” even at very short times. In fact, you can easily show that any wave-like source of light must have a probability of recording a second click immediately after the first one that is at least as big as the probability of recording a second click after a long delay. Most of the time, the probability is actually higher at short times, not lower. A decrease in the probability of a second detection at short times is something that can only be explained by the photon model.
Kimble, Dagenais, and Mandel measured the time required to see a second photon after one photon was detected, and found exactly the behavior they expected: they saw almost no photons within the first few nanoseconds after the detection of one photon, and the few counts they did see could be attributed to the tiny fraction of cases where they had two or more atoms in front of the detector at the same time.
These antibunching experiments have been repeated many times with different systems, all with the same result. A particularly nice version was done by Alain Aspect in the early 1980’s using an atomic cascade source to produce single photons, which went on to also demonstrate the interference of single photons. Later groups have used nonlinear crystals to produce single photons in huge numbers, allowing all sorts of nifty experiments, to the point where Kiko Galvez at Colgate uses single-photon experiments in undergrad labs.
As for references to this stuff, the best collection of material on the particle nature of light that I have seen is The Quantum Challenge by Greenstein and Zajonc, which goes through the whole question of how we know photons exist in an approachable way without skimping on detail. It’s got a bit of math, but nothing too awful. There’s also a centenary review article by Anton Zeilinger in Nature (you probably need a subscription to access it) that gives a good but brief summary of a lot of the cool things that have been done with photons in the hundred years since Einstein suggested they were real.