Quantum memory may be closer to reality thanks to this new router

Quantum computing is full of ideas that sound great until you try to build them. A new experiment now shows a way to move quantum data to exactly where it is needed. It reports 95.3 % success on that routing task in the lab, which is strong enough to take notice.

Instead of treating memory as an afterthought, the team built a router that can direct a quantum signal based on a quantum address.

That ability is central to designs for quantum random access memory (QRAM). It is also critical in other architectures that need flexible data movement between qubits, not just fast logic.

Why a quantum router matters

Classical computers stay nimble because they can stash and fetch data quickly from memory. QRAM is the general idea for doing something similar on a quantum processor.

An address qubit can point to one location or a superposition of locations, a feature described in the original QRAM proposal.

If a router can send a quantum state to a superposition of memory cells, then algorithms that search or sample across many records can be implemented with fewer steps.

That promise has been discussed for years. It has been hard to test, however, without a practical device that actually routes according to a quantum address.

The device, built by a team at Stanford University, is a compact module designed to do exactly that. It routes an incoming qubit to one of two outputs or to a coherent mix of both. This is set by the state of a control qubit.

How the device works

The hardware is built with transmon superconducting qubits (TSQ). This is a widely-used circuit design that is less sensitive to charge noise than early Cooper pair boxes. The team used four, fixed frequency transmons arranged as a switch, an input, and two outputs.

Routing relies on two controlled swaps. These are called iSWAP operations and are activated or blocked by a strong ZZ interaction between the switch and the input.

Two iSWAPs applied in sequence move the quantum state from input to output 1 if the switch is in one state, or to output 2 if it is in the other.

“We realize a Q2 router that uses fixed frequency transmon qubits to implement a routing protocol based on two native controlled iSWAP gates,” wrote Connie Miao from Stanford University, lead author of the study.

The control comes from the always-on ZZ coupling. This shifts the transition frequency depending on the switch state, and lets the device select the right path.

Quantum router results

The reported routing fidelity of 95.3 percent is an average over a set of input and switch configurations that include both classical and quantum states. That number reflects errors mainly from decoherence and from preparing and measuring states.

The group analyzed those error sources. They found that most of the loss can be explained by the limited lifetimes and phase stability of the qubits during the gate sequence.

That is a common pattern in superconducting circuits, as noted in a broad review that tracks how energy relaxation and noise reduce performance when operations take longer than desired.

The gate method matters here because routing involves conditional operations that must be selective yet fast. The device uses sideband drives to implement the swaps.

In addition, the control qubit’s ZZ shift acts like a built-in filter that blocks the wrong swap without extra hardware.

What this does not yet solve

The module routes between two outputs, which is the smallest useful test. Turning this into a large QRAM will require stacking many such modules into a tree structure. The router at each node will load an address bit and pass the state down to the next level.

Scaling that tree introduces fresh challenges. Always-on ZZ couplings make cross talk worse as the tree grows. Future versions will likely need tunable couplers that can turn interactions on for gates and off during idle times.

The error budget also needs to tighten. Even if routing itself is accurate, the memory elements at the leaves must store states long enough to be useful, then return them without extra noise.

Where it could go next

One near-term use is as a local interconnect on a chip, where a few memory or sensor elements share an input line.

A selective router can move the signal with fewer detours, which reduces overhead for error correction and frees timing slack for harder parts of the program.

Longer term, the same idea maps onto quantum networks. A router with quantum addresses could point to entanglement links and set up paths without revealing the route. Such a pattern could matter for privacy and for distributed computation.

These possibilities depend on steady progress in component quality, from better qubit materials to cleaner control pulses.

None of that is automatic, but the device shows that routing and addressing in the quantum domain can be executed together on a superconducting platform.

Key definitions

A transmon is a superconducting qubit made by combining a Josephson junction with a large shunt capacitor, which reduces charge noise at the cost of weaker anharmonicity.

An iSWAP is a two-qubit gate that exchanges excitations between qubits with a specific phase. It is often used to move quantum states around a chip.

A ZZ interaction is a static coupling that shifts one qubit’s frequency depending on another qubit’s state. This can be exploited to make gates conditional, with minimal wiring.

Decoherence is the blanket name for processes that randomize a qubit’s phase or cause it to relax to the ground state, limiting how long information remains reliable.

Understanding quantum router data

Routing fidelity is the fraction of experimental runs that match the target output when you reconstruct the three-qubit state after the protocol.

Higher is better, but context matters. The same qubit set also has to run single-qubit and two-qubit gates for the rest of the computation.

The device demonstrated that the two controlled swaps have strong on-off contrast and that the address can be in a superposition, which is the subtle requirement behind QRAM ideas.

The next test will be keeping that contrast while shrinking gate times and increasing the number of addressable outputs.

The study is published in PRX Quantum.

—–

Like what you read? Subscribe to our newsletter for engaging articles, exclusive content, and the latest updates. 

Check us out on EarthSnap, a free app brought to you by Eric Ralls and Earth.com.

—–