A four-photon measurement could link distant qubits in one step
A proposed optical circuit recognizes a useful family of outcomes with probability 9 in 32, up from 4 in 32 for earlier designs. Both figures assume an ideal experiment.
Reading tools for text size, copying, printing and your reading list
How light can connect four stationary qubits
Imagine four distant stations, each holding a spin that can store quantum information. Each station emits a photon with its own spin and sends the photon to a shared detector setup. A suitable measurement can leave the four distant spins entangled with one another, even though the photons that connected them have been absorbed.[2]
Stefan van den Hoven and colleagues propose routing those four photons through an optical circuit that lets their paths interfere. Several patterns of detector clicks correspond to successful joint measurements. Each tells the experimenter which member of a family of entangled states was selected. The family is called GHZ after Greenberger, Horne and Zeilinger.[1][2]
Where 9 in 32 comes from
Among 16 equally likely four- GHZ basis states, six send two photons to each half of the apparatus. For those six, three quarters of the outcomes reveal enough information to identify the state. Multiplying 6/16 by 3/4 gives 9/32. Earlier designs for the same ideal task achieve 1/8, or 4/32.[2]
How often the four-photon measurement succeeds
12.5% ideal success probability
28.125% ideal success probability
The remaining outcomes are not all equivalent. Some project onto a different kind of entangled state, which can be useful in another protocol. For the distant-spin example, a lab would also have to deliver all four photons, overlap them closely enough for their paths to interfere and detect them reliably. Loss during transmission cuts the number of usable links per hour.[2]
What a higher success rate buys
The team also calculated the cost of assembling a particular six-node ring used in photonic quantum computing. Their design needed an estimated 1,140 three-qubit GHZ states on average, down from 1,520. That is about one quarter fewer starting states for the same encoded ring.[2]
These calculations give an experiment two different benchmarks. Counting successful patterns among trials where all four photons arrive would test the proposed optical measurement. Counting every attempt from emission onward would show how quickly the whole apparatus can produce useful connections. The latter includes photons that never reach a detector.