Gate teleportation for device-scale quantum computing

A useful quantum computer will need to perform logic operations on qubits that are physically distant; gate teleportation is one method to do that. It uses an entangled state, or a pair of messengers, to teleport a logic gate, which executes the operation in much the same way a qubit is teleported. The messengers are entangled beforehand, and each one travels to a qubit and ropes it into the entangled state. The Ion Storage Group at NIST in Boulder, Colorado, experimentally demonstrated gate teleportation in trapped ions. The demonstration served as a test case for a quantum computing architecture using trapped ions.

The group built a device that steered ions into and out of traps, including one where qubits interacted under laser illumination. Using the device, they teleported a controlled-NOT (CNOT) gate, similar to the classical XOR logic gate, in which a target qubit’s spin flips only if a control qubit’s spin is down. A CNOT paired with arbitrary single-qubit rotations can build any possible operation in quantum computing.

In the experiment, researchers perform a CNOT operation on two beryllium ions (B1 and B2 in the figure) kept over 300 µm apart. First, they prepare B1 and B2 in superpositions of spin states and entangle two magnesium ions (M1 and M2) to use as messengers. M1 is shuttled off to B1, and a CNOT operation on them leaves B1, M1, and M2 entangled if B1 is in a superposition. A measurement of M1’s spin, with a possible rotation on M2 depending on the outcome, leaves B1 and M2 entangled. Afterward, M2 is entangled with B1 and B2 through another CNOT, and a measurement of M2’s spin, with a corrective rotation on B1 conditional on the result, yields the desired CNOT operation on B1 and B2. The final gate performs as expected for an ideal CNOT 85–87% of the time.

Teleporting a logic gate tests the pros and cons of different quantum computing platforms. In the team’s trapped-ion platform, the error rates for all the steps performed in the same device were higher than the best rate for any one operation in isolation. The largest error contribution, about 4%, came from entangling M1 and M2, whereas preparing the initial B1 and B2 spin states contributed only about 1%. The current error rate is too high for fault-tolerant quantum computing or quantum computing that forgoes error correction, but researchers in the field now know which steps need the most improvement. (Y. Wan et al., Science 364, 875, 2019 .)