What the interferometer actually measured
Ron Folman's group at Ben-Gurion University of the Negev cooled about 20,000 rubidium atoms into a Bose-Einstein condensate and split each atom into a superposition of two magnetic states, about 113 micrometres below an atom chip patterned with gold wires 2 micrometres thick. One half was made insensitive to magnetic fields and left to gravity; the other was held in place by a force tuned to cancel gravity exactly. At the top of the arc the two halves of a single atom stood about 7.5 micrometres apart, roughly seven times the width of the atomic wave itself, before the sequence was reversed and they were brought back together. The team calls the apparatus the quantum Galileo interferometer.[1]
What survives the recombination is a phase, and theory says it grows with the cube of the falling time: double the fall and the effect grows eightfold. Stretching the free fall to about 2.4 milliseconds, the team logged 633 cycles over 5.3 hours and counted 13 complete oscillations, some 80 radians of accumulated phase. The measured curve departs from the cubic law by about 2 radians across that span, a deviation of about 2.5 per cent. Charles Galton Darwin and Earle Kennard wrote the prediction down in 1927, and until now nobody had read it off an instrument.[1]
The reach of the measurement and its limits
The number matters because the same phase can be derived twice over. Treat gravity as a force acting on a quantum wave and one answer comes out; step into the falling frame, where gravity disappears, invoke the equivalence principle, and the same answer comes out again. Vlatko Vedral, a co-author at the University of Oxford, describes the outcome as showing no conflict between quantum physics and gravity at this level of accuracy. That agreement is a consistency check on the seam between the two theories, at the energies this apparatus reaches — which is a narrower claim than it sounds, because a theory that violates the equivalence principle can still land on the same phase.[1]
The honesty of the measurement is in what it does not reach. The authors caution that the phase they measured leaves standing every theory of equivalence-principle violation that predicts the same phase. Contrast is the other ceiling: interference fringes started at 80 per cent on short runs and faded to 20 per cent on the longest, which is what limits how far the free-fall time can be stretched, and the phase only rises with the cube of that time. The masses and timescales needed to test Roger Penrose's conjecture — that gravity is what destroys superposition — sit far above anything an atom cloud reaches. Penrose is a co-author of the paper that cannot yet test him.[1]
The next measurement
Two directions are already named. Vedral points to running the same experiment in a rotating frame, and to superposing two such systems so that they can pull on each other gravitationally. The other direction is mass: carrying the technique from atoms to something far heavier, a nanodiamond say, is the route by which the question of whether gravity follows quantum rules turns into an experiment. Both inherit the contrast problem, and both have to work against the same trade, in which every extra millisecond of free fall costs interference and pays in phase.[1]
For now the useful thing is the number itself. A first measurement that lands within 2.5 per cent settles nothing, and it was not built to settle anything; it fixes one point on a curve that had only ever been written down, and it tells the next apparatus how much better it has to be. That is the ordinary way a frontier moves: a bar that used to be nowhere in particular now sits somewhere in particular, 13 oscillations and 5.3 hours from the 1927 page it came off.[1]