Addressing requirements for crosstalk-free quantum-gate operation in many-body nanofiber cavity QED systems

TL;DR

This paper addresses scalable quantum information processing with a distributed, all-fiber network of nanofiber cavity QED systems hosting multiple Cs atoms. It develops a photon-mediated phase-flip gate framework for both local (within a cavity) and remote (across cavities) CZ gates, quantified through average gate fidelities, success probabilities, and Pauli error rates, while accounting for cavity reflectivity, cooperativity, and qubit splitting. Analytic and numeric results reveal near-crosstalk-free operation is possible under targeted addressing using local AC Stark shifts and atom-fiber distance control, with maximum fidelities limited by finite qubit splitting and cavity losses. In many-body settings, achieving baseline performance requires addressing non-targeted atoms either by detuning or by spatial separation; combining both targeting mechanisms can substantially reduce requirements. The findings yield practical guidelines for designing scalable, fiber-based quantum networks and highlight a tradeoff between local and remote gate strategies depending on system parameters and addressing capabilities.

Abstract

A distributed network architecture in which flying photons connect individual modules containing stationary atomic qubits is a promising approach for scaling up neutral-atom based quantum-computing platforms. We consider an all-fiber based platform consisting of nanofiber cavity QED systems interconnected via conventional optical fibers. Each nanofiber cavity is strongly coupled to multiple atoms through its evanescent field, and atom pairs within one cavity (local) or two distant cavities (remote) are addressed for performing photon-mediated quantum logic gates on them by controlling the effective light-matter coupling via local AC Stark shifts and atom-fiber distance. We numerically evaluate the required parameters for achieving nearly crosstalk-free gate operation using these targeting methods by calculating average gate fidelities, success probabilities, and Pauli error rates for both local and remote controlled-Z gates. For the case of perfect addressing, we also analytically determine the theoretical optimum gate performance as limited by cavity reflectivity, cooperativity, and qubit level-splitting.

Figures (16)

• Figure 1: Distributed quantum computing architecture. Individual computing nodes consisting of atomic qubits coupled to cavities are interlinked via conventional optical fibers to form a network in which photon-mediated controlled phase-flip gates can be performed either locally on two atomic qubits targeted within the same cavity or remotely on one atom targeted from two distinct cavities each.
• Figure 2: Circuit diagram (a) and sketch (b) of the physical setup for performing a photon-mediated controlled-$Z$ gate between two atomic qubits located in the same cavity and targeted for example via local light shifts from focused laser beams.
• Figure 3: Circuit diagram (a) and sketch (b) of the physical setup for performing an atom-photon controlled-$Z$ gate. A photon initially prepared in $\ket{H}$ is brought into a superposition using a half-wave plate (HWP) acting as a Hadamard gate. A polarizing beam-splitter (PBS) leads to only the $\ket{H}$ component being reflected from the cavity before recombining with the $\ket{V}$ component. Inside the cavity a single atom is targeted, for example via a local AC Stark shift from a focused laser beam. The optical circulator ensures correct order of operation.
• Figure 4: Circuit diagram (a) and sketch (b) of the physical setup for performing a photon-mediated controlled-$Z$ gate between two atomic qubits located in remote computing modules. The photonic qubit is initialized with $\ket{H}$ polarization. Half-wave plates (HWP) act as photonic Hadamard gates and the computing modules consisting of an optical cavity and polarizing beam splitter (PBS) perform atom-photon CZ gates for individually targeted qubits. Optical circulators ensure the correct order of operation. The final $Z$-gate on Qubit 1 is only performed conditional on a $\ket{V}$ polarization measurement outcome.
• Figure 5: Comparison of average gate fidelity for a post-selected local atom-atom gate with $N=2$ calculated for identical parameters both via the entanglement fidelity in Eq. \ref{['eq:average_fidelity']} and via the superposition input in Eq. \ref{['eq:superposition_input']}. See Sec. \ref{['sec:model']} for calculation details.