Top Eleven: Alain Aspect

The final and most recent of the Top Eleven is an experiment that goes right to the heart of the weirdness inherent in quantum mechanics.

Who: Alain Aspect (1947-present), a French physicist. (Again, Wikipedia is a let-down, but CNRS has useful information.)

When: Around 1982 (there are several experiments involved, but the 1982 one is cited by most people).

What: His group performed the first experimental tests of Bell's Inequality, which shows that the predictions of quantum mechanics cannot be explained by a "local hidden variable" theory. Explaining that will take some space, so I'll move it below the fold...

The whole thing comes back to the Einstein, Podolsky, and Rosen (EPR) paper that was one of the nominees for the Cosmic Variance contest. In that paper, Einstein and co-workers showed that there are certain quantum states that exhibit correlations over great distances, and apparently instantaneously. This was deeply troubling to Einstein, who, after all, is famous largely because he showed that instantaneous transmission of information is impossible.

The argument goes like this: You set up a quantum system so that there are some correlations between two particles, say, that their spins are always in opposite directions, but don't measure the exact spin of either particle. Then you shoot those particles out in different directions, and wait until they're well separated before measuring them. When you measure them, you'll discover that whenever one of the two has its spin pointing up, the other spin is pointing down. Or, if you measure one to be pointing to the left, the other will be pointing to the right.

"Well," you say, "that's no big deal." After all, you can do something similar with a classical system-- write something on a piece of paper, rip it in half, and put the two halves in sealed envelopes that you mail to different locations. When the recipients open the envelopes, one will have the left half, and the other the right. You can repeat this as many times as you like, and the correlation will always hold.

The key difference here is that in the classical system, the two envelopes each had a definite state. You didn't know which half was in which envelope, but you knew that one envelope contained the left half, and the other the right, and that was true throughout. In the quantum case, the state of the particle is indeterminate. Not only do you, the experimenter, not know which particle was in which state, neither do the particles. They're in a superposition of "up" and "down" states right up until the moment that the state is measured.

Here's where the problem comes in according to Einstein, Podolsky, and Rosen: suppose your two particle are sent to observers on opposite sides of the Earth, and the observers agree that one will measure the state of her particle at a predetermined time, and the other will measure the state of the other particle a nanosecond later. According to quantum mechanics, the correlation will still hold absolutely, even though there's no time for light to travel from one to the other carrying the information about the state change.

Einstein and his co-authors used the memorable description "spooky action at a distance" to describe this instantaneous transmission, and objected to the whole idea. They felt that there had to be something else going on, some hidden variable that determined the state of the two particles in advance, so that there didn't need to be instantaneous transmission to make things work. Many people thought they were wrong, but nobody could see a way to prove it, until John Bell came up with his famous theorem.

What Bell figured out is that there's a set of measurements you can make on an EPR-type particle pair that will give different results under a "local hidden variable" theory than in regular quantum theory. Basically, you want to do two different measurements on the two particles-- in one case looking for "spin up" versus "spin down," and in the other, looking for "spin left" and "spin right," or some other combination of angles. By compiling the results of a bunch of such measurements, you can determine the probabilities of getting each of the possible outcomes, and those probabilities are different if the states are pre-determined than if they're indeterminate until you make the measurement.

Bell published his result in 1964, and it took almost twenty years before Aspect and his group figured out a way to make it work. They used a "cascade" transition in calcium atoms, in which the atoms decay from an excited state to the ground state by emitting two photons in rapid succession. In order to conserve energy, momentum, and angular momentum, these photons have to be emitted in opposite directions, and with opposite polarizations. You don't know which direction will have which polarization, but you know that there will be two photons, and the polarizations will be correlated, and you can do the measurements to test Bell's theorem by measuring the photons with polarizers at different angles. Aspect and his colleagues demonstrated this in their first experiment.

There's another loophole here, though, if you set the polarizers in advance: some information could be transmitted from one detector to another, or from the detectors to the photon source, before the photons are emitted, and the states could be determined on that basis. To really nail things down, you need to change the polarizer settings while the photons are in flight, and Aspect and his colleagues worked out a way to do that, too. The results were in accordance with Bell's prediction for quantum theory, and in disagreement with the predictions for local hidden variables.

(The original papers are collected at this site, along with some more recent results and commentary. There's a nice article on Bell here.)

Why It's Important: Aspect's Bell inequality experiments demonstrated that the correlations Einstein, Podolsky, and Rosen objected to are really there, and that they can't be explained by a hidden variable theory. Ths is a hugely important step in our understanding of the world, because it shows that there's no way out of the weirdness of quantum theory-- the theory is bizarre, but it works, and any attempt to construct an alternative is going to have to account for some seriously odd stuff.

These EPR correlations are also the basis for a number of quantum information technologies, most notably the idea of quantum cryptography, in which you use measurements on EPR states to generate and distribute random numbers that you can use to encode and decode messages. Because the quantum correlations between these states depend on the states remaining indeterminate, this quantum key distribution is secure against eavesdropping-- somebody measuring the state of one of the two particles in an attempt to steal the key will inevitably disturb the process in a way that can be detected. QKD has been demonstrated over many kilometers of optical fiber, and through kilometer-scale distances in open air (they're working toward satellite distribution).

The entanglement between particles that causes the interesting effects in the EPR case is the same thing that provides the basis for quantum computing, which is a hot topic these days, again because of a relationship between quantum mechanics and code-breaking. Quantum information research doesn't really depend on having Bell inequality tests, but it doesn't exactly hurt.

Reasons to Vote for Him:: Did the first experiments testing Bell's inequality, and ruling out hidden variable theories. Has a fabulous mustache.

Reasons to Vote Against Him: There are still loopholes in the experiments. Deep attachment to local realism. Francophobia.