Wave-particle duality is the concept in quantum mechanics that energy-carrying waves can also behave like particles and that particles can also display a wave aspect. Light, for example, demonstrates the wave phenomena of diffraction and interference but, under other circumstances, appears to be a stream of tiny particles called photons. Electrons, on the other hand, which normally behave like particles, can be made to diffract and interfere as if they consisted of waves.
By the mid-1920s it was obvious, from the photoelectric effect, from the Compton effect, and in other ways, that when light interacts with matter it does so as if it were made of tiny bullets of energy – photons. The rest of the time, it goes about as if it were smeared out in the form of a wave. Apparently, light has an identity crisis, and nowhere was that crisis more evident than in an updated version of an experiment carried out long before quantum theory came on the scene.
Thomas Young's double-slit experiment, dating back to the early 19th century, offers the clearest, most unambiguous proof of the wavelike personality of light. On the screen at the rear of the apparatus appears a series of alternating bright and dark bands. Two waves from a common source, one rippling out from each slit, combine, and the stripes on the screen speak, inarguably, of the adding and canceling of wave crests and troughs.
What happens now if we dim the light source? A standard 60-watt light bulb puts out roughly 150 million trillion photons per second. This vast number underscores why quantum effects, which expose the discreteness of energy, go unnoticed at the everyday level: the individual energy transactions involved are fantastically small. If we want to pursue the question of how light really behaves, on a tiny scale, we have to turn the lamp in Young's experiment down – way down. This was first done in 1909 by the English physicist and engineer Geoffrey Taylor, who was later knighted for his work on aeronautics. Shortly after his graduation from Cambridge, Taylor set up a version of Young's experiment using a light source so feeble that it was equivalent to "a candle burning at a distance slightly exceeding a mile." Even at this level of illumination the interference pattern showed up. The dribble of light passing through the apparatus continued to behave in a wavelike manner.
What if the light source were turned down even further? What if it were dimmed so much that it effectively spat out single photons? There was no way to arrange for this to happen in the early twentieth century; the technology needed just wasn't to hand. Fashioning a light source that emits only one photon at a time isn't as simple as turning on a faucet so that water comes out drip by drip (after all, each water droplet contains many trillions of atoms, each of which is more substantial than a photon.) Consequently, those involved in the formative phase of quantum physics, like those grappling with early relativity theory, had to rely on gedanken – thought experiments – to test their ideas. If Young's experiment could be done using a light source that fired out individual photons, what would be seen on the screen? The only answer that squared both with experiments that had been carried out and with the emerging principles of quantum theory is that the interference pattern would build up, one point at a time. This ought to happen even if there was no more than a single photon passing through the apparatus at any given moment. As the English theoretical physicist Paul Dirac put it: "Each photon then interferes only with itself."
Today double-slit experiments with single photons are routinely set up as demonstrations for undergraduates. An arrangement used at Harvard, for example, employs a helium-neon laser as a light source, two rotatable Polaroid filters to cut the intensity down to barely visible, and a pinhole that is 26 microns (millionths of a meter) in diameter at the front end of a PVC pipe. Further down the pipe is a slide with slits, each 0.04 millimeter wide, set 0.25 millimeter apart. Light from the double-slit then falls onto a sensitive video camera, which produces an image on a screen in which individual flashes, corresponding to single photons, can be seen appearing. With the detection equipment in storage mode, the single flashes of light are captured live, and the characteristic double-slit interference pattern can be watched building up in real time. The familiar bright and dark bands, which cry out for a wave interpretation, emerge like a pointillist painting from the specks that are obviously the marks of individual colliding particles.
Something very, very strange is going on here. In the single-photon, double-slit experiment, each photon starts and ends its journey as a particle. Yet in between it behaves as if it were a wave that had passed through both slits, because that's the only way to account for the interference pattern that forms over time. During its flight from source to detector, the photon acts in a way that defies not only commonsense but all of physics as understood before Planck and Einstein.
You might say, let's keep closer tabs on each photon in the experiment. If it's a particle – a single, pointlike entity – it can't really go through both slits at once, any more than a person can simultaneously walk through two doorways. What happens if we put a detector on one of the slits to tell us whether the photon goes through that slit or the other? This is easy to arrange and, sure enough, the photon is forced to give itself up. By posting a sentry at one of the slits, we learn which slit each photon passes through. But in gaining this knowledge, we lose something else: the interference pattern. Flushed into the open, compelled in mid flight to reveal its whereabouts, the photon abruptly abandons its wavelike behavior and acts purely and simply like a miniature bullet bound on a straight-line trajectory. Somehow the existence of the interference pattern is tied to a lack of knowledge as to which slit the photon actually went through. If we don't ask where the photon is, it behaves like a wave; if we insist upon knowing, it behaves like a particle. In classical physics such a situation would be unthinkable, outrageous. Yet there it is: the act of observing light makes its wave nature instantly collapse and its particle aspect become manifest at a specific point in space and time. It's almost as if a photon knows when it's being watched and alters its behavior accordingly. Evidently, an enigma lies at the heart of the quantum world that, like a Zen koan, resists a solution in familiar, everyday terms.