Trapped Ion Photonic Interconnect Modular

Short answer: One proposed way to scale ion-trap quantum computers is to build many small modules and link them with light. Local gates act inside each register, and photons create entanglement between separate registers. A 2014 Physical Review A analysis argues such a design can be made fault tolerant, and later work studies optical interconnects and photonic modules. These are architecture studies and proofs of concept, not finished large machines.

The modular idea

The 2014 Physical Review A paper starts from the point that scalable hardware able to run nontrivial algorithms will need different types of quantum systems working together. It analyses a modular ion-trap architecture with a hierarchy of interactions. Local entangling gates between qubit memories within one register use natural interactions between the qubits. Entanglement between separate registers is completed through a probabilistic photonic interface, even over large distances. The authors show the architecture can be made fault tolerant and demonstrate its viability for fault-tolerant execution of modest-size quantum circuits. The journal's accompanying feature described it as a design that shows fault tolerance and scalability to large numbers of qubits.

Why the interconnect is the hard part

A paper titled Q-Fly proposes an optical interconnect for modular quantum computers. It argues that, as in classical supercomputers, scaling up needs an optical interconnect, but that signal attenuation leads to irreversible qubit loss, so design guidelines and metrics differ from conventional computing. The design is inspired by the classical Dragonfly topology. Groups of computational end nodes send photons to a shared pool of Bell state analyzers, which perform the entanglement swapping that creates end-to-end entanglement. The authors present a full-stack performance analysis, protocols and a resource scheduler, and build a prototype three-node switched interconnect. They create two-hop entanglement with fidelities of 0.6 to 0.76. Based on loss and infidelity budgets, they find that moderate-radix switches can meet expected near-term needs and that large systems are feasible, and they expect the design to suit ion traps and neutral atoms.

A contrasting photonic approach

A 2025 Nature paper describes a different route, where the computing itself is photonic rather than ion-based. The authors build what they call a sub-performant scale model of a quantum computer from 35 photonic chips, networked over fibre-optic interconnects, with 84 squeezers and 36 photon-number-resolving detectors giving 12 physical qubit modes per clock cycle. They synthesise a cluster state entangled across separate chips with 86.4 billion modes and implement a foliated distance-2 repetition code with real-time decoding. They also analyse tolerance to optical loss, which they call the dominant hurdle to crossing the fault-tolerant threshold, and say the work lays out a path to that threshold.

What this does not tell you

The ion-trap architecture is an analysis of modest-size circuits, and the Q-Fly fidelities of 0.6 to 0.76 come from a three-node prototype. The photonic machine is described by its authors as sub-performant. None of this shows a large fault-tolerant modular ion-trap computer in operation.

FAQ

What is a modular ion-trap architecture?

Several registers of ion qubits with local gates inside each and photonic links entangling different registers.

Is the photonic link deterministic?

The 2014 paper describes it as a probabilistic photonic interface.

What does Q-Fly add?

A switched optical interconnect design with a three-node prototype and two-hop entanglement fidelities of 0.6 to 0.76.

Is optical loss a problem?

Yes. Q-Fly notes attenuation causes irreversible qubit loss, and the Nature paper calls loss the dominant hurdle for its photonic approach.

Has a large modular system been built?

Not in these papers. They report analyses and small prototypes.