# Does Engineered Noise Actually Entangle Superconducting Qubits?

A concurrence of 0.10 — modest by any measure — marks the first experimental realisation of a theoretical proposal more than 20 years in the making. Researchers at the Institute of Science and Technology Austria (ISTA), working with collaborators in Munich and Madrid, have entangled two superconducting [transmon](https://quantumintel.tech/glossary/physical-qubit) qubits separated by a metre of coaxial cable by deliberately exposing them to the same quantum-correlated microwave field, published in *Physical Review X*. The [entanglement](https://quantumintel.tech/glossary/entanglement) is not generated through synchronized pulse sequences, heralding, post-selection, or feedback. It persists autonomously for as long as the correlated field is applied — a qualitative departure from how virtually every other entanglement-distribution scheme in superconducting quantum computing works today.

Lead author Alejandro Andrés-Juanes, a PhD student at ISTA, and co-author Johannes Fink lead the experimental effort. The core mechanism — using dissipation rather than fighting it — belongs to a class of techniques called **dissipation engineering**, first proposed theoretically by Barbara Kraus and Ignacio Cirac over two decades ago. This paper is its first experimental demonstration with transmon qubits.

For distributed quantum computing and quantum networking, an always-on entangled link eliminates significant classical control overhead. The tradeoff: you're confined to the interior of a [dilution refrigerator](https://quantumintel.tech/glossary/dilution-refrigerator), and the concurrence numbers are still far from what fault-tolerant architectures will require.

---

## How the Experiment Works

The correlated noise source is a Josephson parametric converter — a superconducting device that splits each pump photon into a pair of entangled photons at gigahertz frequencies. Each photon of a pair travels down its own coaxial cable to one of two transmon qubits, each half a metre away in opposite directions, for a total qubit separation of one metre.

Because the noise reaching both qubits originates from the same entangled photon pairs, relaxing into that environment does not randomise the phase relationship between the qubits. Instead, they settle into a specific superposition state where photon emission and absorption processes interfere destructively — what the authors describe as "the destructive interference between a photon emission event in the first waveguide and a photon absorption process at the location of the second qubit." Once in this dark state, the qubits stop evolving, and the field passes through unchanged.

The result is a conversion of continuous-variable (CV) entanglement carried by the two microwave beams into discrete qubit entanglement — with roughly a tenth of the original entanglement inherited, yielding the reported concurrence of 0.10.

This is not a post-selected snapshot. The entangled state is maintained continuously for as long as the driving field is active.

---

## Why "Always-On" Matters Architecturally

Conventional entanglement generation in superconducting systems is event-driven: a pulse sequence fires, a detector registers a herald, an entangled pair is announced, and then the clock starts ticking as [decoherence](https://quantumintel.tech/glossary/decoherence) erodes the state until the sequence runs again. At scale — across many qubit pairs across many nodes — this pulse-sequence overhead compounds.

Andrés-Juanes makes the scaling argument directly: "One correlated photon source can drive many pairs at once." The autonomous protocol means the entangled state is available on demand rather than re-initialised on every use cycle. For distributed quantum computing architectures that must stitch together multiple processors via entanglement, this is a meaningful engineering simplification, even if the entanglement quality must ultimately improve substantially.

The practical ceiling, however, is explicit: "We use microwave frequencies, which are only quantum at very low temperatures," Andrés-Juanes notes. "The real limit in the separation is how big your dilution refrigerator is." The longest microwave entanglement link reported so far is 30 metres, at ETH Zurich. The ISTA experiment operates at one metre.

---

## Sceptical Read: What 0.10 Concurrence Actually Means

A concurrence of 0.10 on a scale where 1.0 represents a perfect Bell state is, to be direct, a weakly entangled pair. Fault-tolerant quantum error correction operates well above this regime — surface codes and other [QEC](/glossary/fault-tolerant-quantum-computing) schemes demand high-fidelity entangled links approaching the error threshold, not low-concurrence steady states.

The authors acknowledge this: "For now, the numbers are modest." The group is working on a module with more than one qubit per node, which is the natural next step toward improving concurrence and demonstrating multi-node operation.

There is also the temperature constraint. Microwave-frequency superconducting hardware is fundamentally bound to dilution refrigerator operation. Scaling this approach to distributed quantum computing across a building — let alone a datacenter — requires either vastly larger cryogenic infrastructure or a frequency conversion pathway to optical domains, where the coupling physics become their own unsolved problem.

What this work does not yet demonstrate: high concurrence, multi-qubit node operation, integration with active QEC cycles, or operation beyond the metre scale in the superconducting regime. Positioning it as a near-term networking primitive would be premature.

---

## What It Means for the Industry Trajectory

Dissipation engineering as a field has been theoretically attractive for decades precisely because it offloads entanglement maintenance from classical control hardware onto the physics of the system itself. This first transmon realisation — published in *Physical Review X*, not a preprint — establishes that the physics works as predicted, even if the engineering parameters need significant improvement.

For the quantum networking and distributed quantum computing space, this represents a distinct path from the heralded entanglement approaches being pursued by most superconducting and trapped-ion vendors. The no-overhead, always-on character of the entangled link is genuinely differentiated. If concurrence can be improved by an order of magnitude or more through device optimisation — better Josephson parametric converters, reduced cable loss, improved qubit-waveguide coupling — the architecture becomes worth serious engineering investment.

The applicability extends beyond computing. The authors cite quantum repeaters and networked sensing as target applications. Quantum repeaters, in particular, are constrained by the no-cloning theorem: quantum states cannot be amplified classically, so relay stations must distribute entanglement. An autonomous, always-on entanglement source at each node could simplify repeater design considerably, provided fidelity can reach operationally useful levels.

This is a proof-of-concept result that validates a 20-year-old theoretical prediction. The gap between 0.10 concurrence in a one-metre, two-qubit system and a practical quantum network node is large. But the physics is now confirmed experimental fact.

---

## Key Takeaways

- ISTA researchers entangled two transmon qubits one metre apart using quantum-correlated microwave noise — the first experimental realisation of a dissipation engineering proposal by Kraus and Cirac dating back more than 20 years
- The entanglement source is a Josephson parametric converter that produces entangled photon pairs; each photon drives one qubit, yielding a concurrence of 0.10
- No pulse sequences, heralding, post-selection, or feedback were required — the entangled state is autonomous and persists as long as the driving field is active
- The practical separation limit is set by dilution refrigerator size; the longest microwave entanglement link reported to date is 30 metres, at ETH Zurich
- Concurrence of 0.10 is far below what fault-tolerant QEC architectures require — the group is working toward multi-qubit node modules as a next step
- Target applications include distributed quantum computing, quantum repeaters, and networked quantum sensing

---

## Frequently Asked Questions

**What is dissipation engineering in quantum computing?**
Dissipation engineering refers to deliberately designing a qubit's interaction with its environment so that relaxing into that environment drives the system toward a useful quantum state — such as an entangled state — rather than a classical, decoherent one. Normally, environmental coupling is minimised because it destroys quantum information; dissipation engineering exploits it constructively.

**What does a concurrence of 0.10 mean?**
Concurrence is a measure of entanglement between two qubits, ranging from 0 (no entanglement) to 1 (maximally entangled Bell state). A concurrence of 0.10, as reported in this ISTA experiment, indicates weak but genuine entanglement — sufficient to confirm the physical principle, but well below the fidelity levels needed for practical quantum error correction or networking protocols.

**Why is an always-on entangled state useful?**
Conventional entanglement generation is event-based: a pulse sequence creates an entangled pair, which then decoheres and must be recreated. An always-on entangled link eliminates this reinitialisation overhead and makes the entanglement immediately available whenever a computation or sensing task needs it, which matters most as systems scale to many qubit pairs.

**What limits the distance of this approach?**
The experiment uses microwave-frequency photons, which only behave quantum mechanically at very low temperatures — requiring dilution refrigerator operation. The physical separation between qubits is therefore bounded by the size of the cryogenic system. The authors note the longest demonstrated microwave entanglement link is 30 metres, achieved at ETH Zurich.

**How does this relate to quantum repeaters?**
Quantum repeaters are relay stations needed to extend quantum communication links over long distances, since quantum states cannot be copied or classically amplified (the no-cloning theorem). An autonomous, always-on entanglement source at each repeater node could simplify the classical control logic and timing requirements of repeater networks, if entanglement fidelity can be raised to operationally useful levels.