The Night Quantum Mechanics Beat Einstein

The Mechanism

In 1935, Albert Einstein, Boris Podolsky, and Nathan Rosen published a paper in *Physical Review* titled *Can Quantum-Mechanical Description of Physical Reality Be Considered Complete?* (Einstein, Podolsky & Rosen, "EPR," *Phys. Rev.* 47: 777-780, 1935). The paper argued that quantum mechanics, as a description of nature, must be *incomplete*: the theory predicts that if you measure one of two correlated particles, you instantaneously determine the corresponding property of the second particle, *no matter how far apart they are*. Either, Einstein argued, this is real — in which case there must be a *faster-than-light influence* between the two particles, which violates special relativity — or it is a mathematical illusion, an artifact of the theory's incompleteness, and the apparent instantaneous correlation is just a reflection of some additional *hidden variable* attached to the two particles when they were created, fixing their later measurement outcomes from the start. Einstein preferred the second interpretation: there must be hidden variables, and quantum mechanics was an averaged statistical theory over them, the way thermodynamics is an averaged statistical theory over atomic motions. He called the alternative — instantaneous influence between far-apart particles — *spukhafte Fernwirkung*, "spooky action at a distance," and he refused to believe in it. Niels Bohr's reply, the same year, argued that EPR's notion of *element of reality* was itself ill-defined: at the quantum scale, the very property being measured does not exist until the measurement is performed; there is no fact-of-the-matter about the second particle's polarization before someone measures it. The disagreement was philosophical. By 1935, with no experimental way to distinguish the two interpretations, the question entered a thirty-year dormancy. The mathematics was changed in 1964 by an Irish-born theoretical physicist named John Stewart Bell, then at CERN in Geneva. Bell, working alone on a small paper in his spare time, produced *Bell's theorem*: a mathematical inequality that any *local hidden-variable* theory must satisfy, but that *quantum mechanics' predictions violate*. The inequality is the long-sought experimental test of EPR's central claim. The basic structure of the theorem: take two correlated particles (in a typical setup, two photons produced together with correlated polarizations); send them in opposite directions; let each photon pass through a polarizer set at some angle; record whether or not it passes; for each pair, record both outcomes. Now do this for many such pairs, with the *angles of the two polarizers varied independently and randomly*. The *correlation function* between the two measurement outcomes — how strongly the "yes/no" answer on one side predicts the "yes/no" answer on the other, as a function of the angles — is the quantity to test. *Bell showed:* if there is any *local hidden-variable* explanation (each photon has a pre-existing program, set when the pair was created, telling it how to respond to every possible polarizer angle, with no influence between the two photons after they have separated), then the correlation function obeys an inequality — *Bell's inequality*. *Quantum mechanics, by contrast, predicts a correlation function that violates the inequality*. The two theories make *different, experimentally distinguishable predictions*. Bell published the paper in the obscure new journal *Physics* (vol. 1, no. 3, 1964); the journal folded the following year; the paper was widely overlooked for nearly a decade. The first experimental tests began in the early 1970s. John Clauser and Stuart Freedman, at Berkeley, did the first test in 1972 (using a calcium-cascade photon source and *static* polarizers) and reported a clear violation of Bell's inequality. Clauser's experiment was the first measurement against EPR, but it had a critical loophole: the two polarizers were set at fixed angles for hours at a time; in principle, the hidden variable carried by each photon could have "known" what angle each side was set to. The test that closed this loophole — the test that made the violation experimentally airtight, the test Bell himself had specifically proposed — required that the *angles of the polarizers be set, randomly, at moments after the photons had been emitted but before they reached the polarizers*, so that no signal traveling at the speed of light could pass from one polarizer's setting to the other side and influence the result. This is called the *locality loophole closure*. It is the version of the Bell experiment that, if it violated the inequality, would settle the question definitively in quantum mechanics' favor. The locality-loophole experiment was attempted, for the first time, in December 1982, by a team of three physicists at the Institut d'Optique in Orsay, twenty kilometers south of Paris. The team: Alain Aspect, then a 35-year-old assistant professor and the experiment's lead; Jean Dalibard, a young assistant; and Gérard Roger, a research engineer responsible for the optical apparatus. The setup: a calcium atomic cascade, excited by a krypton laser, produced pairs of polarization-correlated photons at a rate of about 100 pairs per second. The two photons of each pair traveled in opposite directions along the laboratory's optical axes. At a distance of 6 meters from the source on each side, the photons entered a *fast acousto-optic switch* — a piezoelectric crystal driven by an ultrasound transducer at roughly 25 MHz that would deflect each incoming photon, every 10 nanoseconds, randomly into one of two paths, each ending in a different polarizer set at a different angle. The two switches, on opposite sides of the apparatus, operated *independently and asynchronously*. Because the switches were 12 meters apart and the switching time was 10 nanoseconds — and 10 nanoseconds at the speed of light is 3 meters, less than 12 — the choice of which polarizer each photon would meet was made *too quickly* for any signal traveling at the speed of light to carry information from one side to the other within the relevant interval. The locality loophole was closed. The result, recorded over several weeks of running, was a violation of Bell's inequality at the *5σ* level. The paper, Aspect, Dalibard & Roger, "Experimental Test of Bell's Inequalities Using Time-Varying Analyzers," was published in *Physical Review Letters* vol. 49, pages 1804-1807, on December 20, 1982. The result was widely received as the experimental settlement of the EPR question. Quantum mechanics is *not* a statistical theory over hidden local variables. The world is *fundamentally non-local* in the precise sense Bell defined: the outcomes of measurements on entangled particles cannot be explained by any pre-existing programs the particles carry. The instantaneous correlations between far-apart entangled particles — what Einstein called spooky action — are *real*. Quantum mechanics, as written down in 1925-1928, is the correct description of the physical world; Einstein's preferred alternative is excluded. Subsequent experiments have closed every remaining loophole. The 2015 *Hensen et al.* experiment in Delft used spacelike-separated nitrogen-vacancy centers in diamonds, closing the detection and locality loopholes simultaneously; the 2018 *BIG Bell Test* used 100,000 internet volunteers to provide the random measurement settings, closing the freedom-of-choice loophole. The violation, at every loophole closure, is consistent with quantum mechanics. The 2022 Nobel Prize in Physics was awarded to John Clauser, Alain Aspect, and Anton Zeilinger "for experiments with entangled photons, establishing the violation of Bell inequalities and pioneering quantum information science." Aspect was 75. The night quantum mechanics beat Einstein was December 20, 1982. The result was published in a four-page *Physical Review Letters* paper. The universe, as the paper made clear for the first time experimentally, does not have local hidden variables. Nothing in physics has been the same since.