Short answer: Neutral atoms in optical tweezers can be entangled using Rydberg states, but gate errors were long the main obstacle. A 2019 experiment reported Bell-state fidelity of 0.86 to 0.89 on a pair of sites in a 121-site array. A 2023 Nature paper reports two-qubit gates with 99.5 percent fidelity on up to 60 atoms in parallel, and a 2023 theory paper designs gates that stay above 0.999 fidelity despite control errors. These are laboratory and simulation results.
The challenge
The 2023 Nature paper says neutral-atom arrays have emerged as a promising platform, with coherent control over hundreds of qubits and any-to-any gate connectivity in a flexible, reconfigurable architecture. The main outstanding challenge, in its words, was reducing errors in entangling operations mediated through Rydberg interactions.
An early array experiment
A 2019 Physical Review Letters paper demonstrates two-qubit Rydberg blockade and entanglement on a pair of sites in a two-dimensional array defined by a grid of blue-detuned light lines with 121 trapping sites. The observed Bell-state fidelity was 0.86 with an uncertainty of 0.02. Accounting for state preparation and measurement errors the authors infer 0.88, and accounting for single-qubit errors they infer 0.89 for a Bell state made by the Rydberg-mediated controlled-Z gate. A detailed error model points to finite atom temperature and laser noise as the dominant error sources.
99.5 percent on many atoms in parallel
The 2023 Nature paper reports two-qubit entangling gates with 99.5 percent fidelity on up to 60 atoms in parallel, which the authors say surpasses the surface-code threshold for error correction. The method uses fast single-pulse gates based on optimal control, atomic dark states to reduce scattering, and improvements to Rydberg excitation and atom cooling. They benchmark fidelity with several methods based on repeated gate applications, characterise the physical error sources and outline improvements. They also generalise the method to entangling gates on more qubits and demonstrate low-error three-qubit gates, saying the advances lay groundwork for large-scale algorithms, error-corrected circuits and digital simulations.
Robust control by design
A 2023 Physical Review Research paper studies how robust two-qubit gates are to deviations in experimental controls on a neutral atom platform using Rydberg states. The authors build robust controlled-Z gates using quantum optimal control that keep Bell-state fidelity above 0.999 in the presence of significant deviations of the coupling strength to the Rydberg state, which can come from laser intensity noise and atomic motion in an inhomogeneous coupling field. They also discuss mitigating laser detuning errors. The pulses are short compared with the decay timescale set by spontaneous emission and blackbody radiation, and the optimisation and fidelity calculations include the finite Rydberg-state lifetime. They calculate properties such as lifetimes, polarizabilities and blockade strengths for strontium-88 atoms.
What this does not tell you
The 0.999 figure is a simulated result for designed pulses, while the 99.5 percent and 0.86 to 0.89 figures are measured, with different methods and conditions. A high two-qubit fidelity is one ingredient and not a full error-corrected computer.
FAQ
What is a Rydberg gate?
A two-qubit entangling gate that uses the strong interaction between atoms excited to Rydberg states.
How good have the gates become?
A 2023 Nature paper reports 99.5 percent fidelity on up to 60 atoms in parallel.
What limited earlier work?
The 2019 experiment points to atom temperature and laser noise.
Can control errors be tolerated?
A 2023 theory paper designs gates that keep fidelity above 0.999 despite significant coupling deviations, in simulation.
Is this a working computer?
No. These are gate-level results from laboratories and simulations.
All field notes